Flashbulb memories : new challenges and future perspectives [Second edition.] 9781315623481, 131562348X, 9781317226147, 1317226143, 9781317226154, 1317226151

Table of contents :
Cover
Title
Copyright
Contents
Figures
Tables
Contributors
Acknowledgments
Introduction: How research on Flashbulb memories has developed in the last ten years
1 Measurement issues in the study of Flashbulb memory
2 Flashbulb memory methods
3 Using Structural Equation Modelling approaches to better understand the formation of Flashbulb memories
4 Ordinary memory processes shape Flashbulb memories of extraordinary events: A review of 40 years of research
5 The consequences of consequentiality
6 When a flash is caught in a lab: An experimental approach to the investigation of Flashbulb memories
7 Flashbulb, personal, and event memories in clinical populations
8 A comparison of Flashbulb memories for positive and negative events and their biopsychosocial functions
9 Flashbulb memories and social identity
10 Aligning Flashbulb and collective memories
11 Flashbulb memories and collective memories: Psychosocial processes related to rituals, emotions, and memories
12 Culture in Flashbulb memory
13 The study of Flashbulb memories: The good, the bad, and the way forward
Index

Citation preview

FLASHBULB MEMORIES

Are Flashbulb memories special or ordinary memory formations? Are emotional, cognitive, or social factors highly relevant for the formation of Flashbulb memories? How can sociological, historical, and cultural issues help us to understand the process? What is the difference between Flashbulb memories, memories of traumatic experiences, and highly vivid personal memories? How can we provide a valid and reliable measure for Flashbulb memories? This edition of Flashbulb Memories: New Challenges and Future Perspectives revisits these questions, considering significant new evidence and research in the field. It now includes additional chapters focusing on experimental investigations, and review studies on positive vs. negative Flashbulb memories. Bringing together leading international researchers, the book presents significant progress in this area of research, which has remained divisive for the past 40 years.The discussion of Flashbulb memories also contributes to the understanding of the general functioning of autobiographical memory. It will provide essential reading for researchers in Flashbulb memories and will be of great interest to those in related areas such as cognitive psychology, social psychology, cross-cultural psychology, sociology, political sciences, and history, as well as clinicians dealing with those who have strong Flashbulb memories after personal traumatic events. Olivier Luminet is Research Director at the Belgian Fund for Scientific Research and Professor in Psychology at the Université Catholique de Louvain and at the Université Libre de Bruxelles (ULB) in Belgium. Antonietta Curci is Associate Professor at the University of Bari, Italy.

FLASHBULB MEMORIES New Challenges and Future Perspectives Second Edition

Edited by Olivier Luminet Antonietta Curci

First published 2018 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 711 Third Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2018 Olivier Luminet and Antonietta Curci The right of Olivier Luminet and Antonietta Curci be identified as editors of this work has been asserted by them in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Names: Luminet, Olivier, 1969- editor. | Curci, Antonietta, 1969- editor. Title: Flashbulb memories : new challenges and future perspectives / edited   by Olivier Luminet, Antonietta Curci. Description: Second Edition. | New York : Routledge, 2017. | Revised edition   of Flashbulb memories, 2009. Identifiers: LCCN 2017007682| ISBN 9781138653979 (hardback) |   ISBN 9781317226154 (web pdf) | ISBN 9781317226130 (mobi/kindle) |   ISBN 9781315623481 (Master) | ISBN 9781317226147 (ePub) Subjects: LCSH: Flashbulb memory. Classification: LCC BF378.F55 F53 2017 | DDC 153.1/3—dc23 LC record available at https://lccn.loc.gov/2017007682 ISBN: 978-1-138-65397-9 (hbk) ISBN: 978-1-138-65398-6 (pbk) ISBN: 978-1-315-62348-1 (ebk) Typeset in Bembo by Apex CoVantage, LLC

CONTENTS

Figuresvii Tablesix Contributorsxi Acknowledgmentsxiv Introduction: How research on Flashbulb memories has developed in the last ten years Olivier Luminet and Antonietta Curci

1

  1 Measurement issues in the study of Flashbulb memory Antonietta Curci

11

  2 Flashbulb memory methods Daniel B.Wright and Marianna E. Carlucci

31

  3 Using Structural Equation Modelling approaches to better understand the formation of Flashbulb memories Olivier Luminet

48

  4 Ordinary memory processes shape Flashbulb memories of extraordinary events: A review of 40 years of research Jennifer M.Talarico and David C. Rubin

73

  5 The consequences of consequentiality Jasmen Rice, Karina Hamamouche, and John N. Bohannon III

96

vi Contents

  6 When a flash is caught in a lab: An experimental approach to the investigation of Flashbulb memories Tiziana Lanciano, Antonietta Curci, and Giuseppe Sartori   7 Flashbulb, personal, and event memories in clinical populations Michelle J.Tat, Carl A. Gold, and Andrew E. Budson   8 A comparison of Flashbulb memories for positive and negative events and their biopsychosocial functions Charles B. Stone and Alexander C.V. Jay

119 137

161

  9 Flashbulb memories and social identity Dorthe Berntsen

182

10 Aligning Flashbulb and collective memories William Hirst and Robert Meksin

201

11 Flashbulb memories and collective memories: Psychosocial processes related to rituals, emotions, and memories Dario Páez, Anderson Mathias, Stefano Cavalli, Eduardo Guichard, José Joaquín Pizarro, Lander Méndez, and Sofián El-Astal 12 Culture in Flashbulb memory Qi Wang and Cagla Aydin 13 The study of Flashbulb memories: The good, the bad, and the way forward Lucy V. Justice, Shazia Akhtar, and Martin A. Conway

219

240

263

Index272

FIGURES



1.1 1.2 1.3 2.1

2.2 2.3

3.1 3.2

3.3 3.4

3.5 5.1

ICCs for one-parameter IRT models. ICCs for two-parameter IRT models. ICCs for three-parameter IRT models. A simplified FBM model, based on Brown and Kulik (1977) and subsequent models. A distribution, for some population of events, of memorability with events beyond some value being called FBMs. The number of times Brown and Kulik (1977) was cited each year since publication (shown with a +) and the number of times the phrase “Flashbulb memory” or “Flashbulb memories” was used in titles and abstracts (shown with a ●). The data are from the ISI Web of Knowledge collected on 13 March, 2008 and go up to 2007. FBM formation according to Conway et al. (1994). FBM formation according to Finkenauer et al. (1998). Dashed lines are paths that were found in a subsequent test of the model (Luminet & Curci, 2007). FBM formation according to Er (2003) for the victim group. FBM formation according to Tinti et al. (2014). Plain arrows: paths expected to be significant; short-dashed arrow: path expected not to be significant; long-dashed arrow: no prediction. FBM formation according to Day and Ross (2014). Hypothetical taxonomy of FBM-inspiring events and their consequentiality. Horizontal is the nature of the event.Vertical is the target of the event or to whom the event occurs. Diagonal represents the valence of the event. FBMs such as assassinations typically fall in the physical, aversive, and other volume of the FBM space.

16 17 17 32 33

42 52

53 58

62 65

101

viii Figures

7.1 Intensity of emotion reaction on a 1 (low) to 5 (high) scale for patients with Alzheimer’s disease (AD) dementia, mild cognitive impairment due to AD (MCI), and older adults as a function of emotion. Error bars show the standard error of the mean (from Budson et al., 2004). 7.2 Percentage of participants who answered one, two, three, four, or five personal information questions (time, people, place, activity, and source) correctly at the follow-up interview relative to the initial one as a function of group (patients with Alzheimer’s disease (AD) dementia, mild cognitive impairment due to AD (MCI), and older adults), by response type (recall vs. recognition) (from Budson et al., 2004). 7.3 Adjusted recognition data for personal and factual information in patients with Alzheimer’s disease (AD) dementia, patients with mild cognitive impairment due to AD (MCI), and older adults (Older Adult) showing correct, distorted, and failed responses (from Budson et al., 2007). 7.4 A. Memory accuracy for older adults (OACs) and patients with mild cognitive impairment due to Alzheimer’s disease dementia (MCI pts.) in response to questions about details on the outcome of the 2008 Presidential election. B. Breakdown of correct memory for personal information by memory type (recall and recognition). C. Breakdown of correct memory for public information by memory type (recall and recognition). Horizontal black bars indicate significant differences between question types. Horizontal grey bars indicate significant differences between groups. Dashed horizontal black bars indicate significant differences between memory types. All differences are at the p < .05 level. 9.1 The original model of FBM formation introduced by Brown and Kulik (1977). 9.2 An alternative model of FBM formation taking activation of social identity as a causal mechanism. 10.1 Relation between media coverage and memory accuracy for facts about Flashbulb memory events: The attack of September 11th and the Challenger Explosion (From Hirst et al., 2009). 10.2 Proportion of accurately recalled critical and non-critical facts about the attack of September 11th within the first week and after one year, three years, and ten years (Based on Hirst et al., 2015). 12.1 The culturally embedded model of Flashbulb memory. 12.2 Lifespan retrieval curves of a) public event memories and b) FBM details in five cultures.

142

143

144

145 185 188

210

211 241 255

TABLES

4.1 A summary of the differences between FBM and ordinary AM 74 4.2 A summary of the evidence supporting the mechanisms proposed for enhancing FBM relative to ordinary AM 75 5.1 Means across dependent variables for race 112 5.2 Means across dependent variables for delay 113 6.1 Strengths and limitations of naturalistic vs. experimental FBM approaches122 6.2 Contextual details considered in FBM experimental studies 123 6.3 An example of aIAT applied to FBM investigation (adapted by Curci et al., 2015) 130 7.1 Selected studies of national traumatic events in different patient populations152 8.1 A comprehensive list of the studies examining positive FBMs in terms of the authors, the valences compared, the presence of a control group, intensity, type of events compared, temporal 164 distance, operationalization of FBMs, and their general results 8.2 Comparing FBMs for positive events to those for negative and neutral events in terms of vividness, confidence, consistency, and detail169 9.1 Studies examining group differences in relation to the 190 formation of FBMs 11.1 Comparison between the key characteristics of CMs and FBMs 221 11.2 Frequency distribution of types of events considered in FBM 223 and CM studies 11.3 Percentage of remembered events as a function of being within or outside the critical period (experiencing the event between 10 and 30 years old) by country and nature of event 227

x Tables

11.4 Ten most important historical events during participants’ lives: CEVI studies 11.5 Comparison between CM and FBM predictors based on a review of studies

234 235

CONTRIBUTORS

Shazia Akhtar, Department of Psychology, City, University of London, Northamp-

ton Square, London EC1V 0HB, UK. [email protected] Cagla Aydin, Faculty of Arts and Social Sciences, Sabancı University, Orta Mahalle, Universite Cad. No: 27 Tuzla, 34956, Istanbul, Turkey. [email protected] Dorthe Berntsen, Center on Autobiographical Memory Research, Department of

Psychology and Behavioural Sciences, Aarhus University, Bartholins Allé 9, 8000 Aarhus C, Denmark. [email protected] John N. Bohannon III, Wesley A. Dunn Distinguished Professor, Department of Psychology, Butler University, Indianapolis, IN 46208, USA. [email protected] Andrew E. Budson,VA Boston Healthcare System and Boston University ­Alzheimer’s Disease Center, 150 S. Huntington Ave, Boston, MA 02130, USA. [email protected] Marianna E. Carlucci, Department of Psychology, Loyola University Maryland,

4501 N. Charles Street, Baltimore, MD 21210, USA. [email protected] Stefano Cavalli, Centre of Competence on Aging, University of Applied Sciences and Arts of Southern Switzerland (SUPSI), Via Violino, 6928 Manno, Switzerland. [email protected] Martin A. Conway, Department of Psychology, City, University of London, Northampton Square, London, EC1V 0HB, UK. [email protected] Antonietta Curci, Department of Education, Psychology, Communication, University

of Bari “Aldo Moro”,Via Crisanzio, 42, 70121 Bari, Italy. [email protected]

xii Contributors

Sofián El-Astal, Department of Psychology, Al-Azhar University – Gaza, Gaza Strip, Palestine, [email protected] Carl A. Gold, Department of Neurology & Neurological Sciences, Stanford Uni-

versity, 300 Pasteur Drive, Stanford, CA 94305, USA. [email protected] Eduardo Guichard, Centre Interfacultaire de Gérontologie et d’Études des Vul-

nérabilités, Université de Genève, 54, route des Acacias, CH-1227 Carouge (GE), Switzerland. [email protected] Karina Hamamouche, Infant and Child Cognition Lab, Department of Psychol-

ogy, Boston College, 140 Commonwealth Ave, Chestnut Hill, MA 02467, USA. [email protected] William Hirst, Department of Psychology, New School for Social Research, 80

Fifth Avenue, New York, NY 10011, USA. [email protected] Alexander C. V. Jay, Department of Psychology, John Jay College of Criminal Justice & The Graduate Center, The City University of New York, 524 West 59th Street, New York, NY 10019, USA. [email protected] Lucy V. Justice, Department of Psychology, Nottingham Trent University, 50 Shakespeare Street, Nottingham, NG1 4FQ, UK. [email protected] Tiziana Lanciano, Department of Education, Psychology, Communication, University of Bari “Aldo Moro”,Via Crisanzio, 42, 70121 Bari, Italy. tiziana.lanciano@ uniba.it Olivier Luminet, Psychological Sciences Research Institute (IPSY), Univer-

sité Catholique de Louvain (UCL), Place du Cardinal Mercier, 10, bte L3.05.01, B-1348 Louvain-la-Neuve, Belgium. [email protected] Anderson Mathias, Department of Social Psychology and Methodology, University of the Basque Country, Av. de Tolosa 70, 20018 San Sebastian, Spain. anderson. [email protected] Robert Meksin, Department of Psychology, New School for Social Research, 80

Fifth Avenue, New York, NY 10011, USA. [email protected] Lander Méndez, Department of Social Psychology and Methodology, University

of the Basque Country, Av. de Tolosa 70, 20018 San Sebastian, Spain. lander855@ gmail.com

Contributors  xiii

Dario Páez, Department of Social Psychology and Methodology, University of the Basque Country, Av. de Tolosa, 70, 20018 San Sebastian, Spain. [email protected] José Joaquín Pizarro, Department of Social Psychology and Methodology, Uni-

versity of the Basque Country, Av. de Tolosa, 70, 20018 San Sebastian, Spain. jose. [email protected] Jasmen Rice, The Chicago School of Professional Psychology, DC Campus, 901

15th St NW, Washington, DC 20005, USA. [email protected] David C. Rubin, Juanita M. Kreps Professor of Psychology and Neuroscience, Department of Psychology and Neuroscience, Duke University, 417 Chapel Drive, Durham NC 27708–0086, USA. [email protected] Giuseppe Sartori, Department of General Psychology, University of Padua, Via

Venezia, 8, 35121 Padua, Italy. [email protected] Charles B. Stone, Department of Psychology, John Jay College of Criminal Justice & The Graduate Center, The City University of New York, 524 West 59th Street, New York, NY 10019, USA. [email protected] Jennifer M. Talarico, Department of Psychology, Lafayette College, 350 Hamilton Street, Easton, PA 18042, USA. [email protected] Michelle J. Tat,VA Boston Healthcare System, 150 S. Huntington Ave, Boston, MA

02130, USA. [email protected] Qi Wang, Department of Human Development, Cornell University, MVR 209,

Ithaca, NY 14853, USA. [email protected] Daniel B. Wright, Aspire University, Redwood City, CA 94063, USA. Dbrook-

[email protected]

ACKNOWLEDGMENTS

The history of a book is always a collective adventure, the result of the hard work of several people. Most of them are involved in this volume. We would like first to thank all the authors for sharing our enthusiasm when preparing the reedition of the book. We would like to address a special thank you to Martin A. Conway who accepted with two co-authors to write a Discussion chapter. Martin has been highly involved in the research on Flashbulb memories (FBMs) since the early 1990s, being the first author to test a model predicting their formation. Since then, he has always followed this literature with great interest and attention. We are thus very happy to have his expert view on the current research in the field. There are additional people who were very important for making this book possible and whom we would like to thank. Firstly, David Pillemer and Robyn Fivush who wrote great chapters in the 2009 edition. They decided that they were not going to contribute to this new edition but their chapters from 2009 are still worth reading for all of us, and especially for the new researchers in the field. There are also people who motivated our interest for the research on FBMs a long time ago. The first is Catrin Finkenauer. Without her enthusiasm to conduct a study related to the FBMs for the sudden death of the Belgian king Baudouin in 1993, it is likely that one of the editors of the present book (OL) would not have started investigating these questions. Lydia Gisle is also strongly associated with the first steps of the research on FBM of the first editor of the book (OL). Thanks also to Martial van der Linden who initiated a seminar on autobiographical memory in the fall of 1993. The seminar contributed greatly to the study published in 1998 related to the Belgian king’s death. Finally, this research project would not have been possible without the generous support of the Belgian National Fund for Scientific Research (FRS-FNRS). Several times, the second editor of this book (AC) has been faced with roadblocks in running FBM projects. On many occasions, the generous advice of Bernard Rimé, his passionate curiousity of all scientific material,

Acknowledgments  xv

and his expertise in research work have given her the motivation to overcome these obstacles. We are deeply grateful to Marianne Bourguignon, who undertook the boring task of checking the references and manuscripts of this second edition, and preparing them for the final steps of the editorial process. Finally, we gratefully acknowledge the help of all undergraduate and postgraduate students, who enthusiastically cooperated in the many data collections concerning FBM. I (OL) would also like to thank close relatives. First, my parents who greatly contributed to my passion for psychological matters. And a special thank you to my father for his unbelievable knowledge of the history of Europe and North America, particularly of the 20th century. He transmitted to me this passion for history and for the processes explaining the building of collective memories. I would also like to thank my wife Annette and my two children, Leah and Pablo, for supporting a sometimes too busy husband and father.Your understanding helped a lot in making this book possible. I (AC) would like to thank my relatives, friends, colleagues, and students for being the co-protagonists of many FBMs of relevant benchmark events of my personal and professional life. I would also like to thank my family and my husband Giuseppe for providing a wise counterbalance to my hectic agenda. I dedicated the first edition of this book to my mother’s patience and in memory of my father. Now I add a hopeful thought on the future of my daughter Federica, for her to continue to enrich my life with new, unforgettable FBM experiences.

INTRODUCTION How research on Flashbulb memories has developed in the last ten years Olivier Luminet and Antonietta Curci

Definition, origins, and evolution of the interest for the construct Flashbulb memories (FBMs) are memories for the circumstances in which individuals have heard about particular news. Research has mostly studied public news but FBMs also occur for private events (Pillemer, 2009). We all have these types of memories, and the notion of FBMs is now very popular in the media, which is often replete with anniversary dates in which the reader is often asked about some canonical variables assessing FBMs, with “Where were you?” and “What were you doing when you heard about the 9/11 attacks?” being prototypical questions. The news about a particular event does not need to involve as many people and countries as the 9/11 attacks did. It can apply to any community, whatever its size, as long as the event holds an emotional tone and is distinctive and consequential for most members of the community (e.g. Talarico & Rubin, Chapter 4; Rice, Hamamouche, & Bohannon, Chapter 5, this volume). Although systematic research on FBMs only started at the end of the 1970s, it has already been reported in some preliminary work. The roots of this domain can be related to news of the assassination of President Lincoln in 1865. Almost 35 years later, F. W. Colegrove (1899) published an article in the American Journal of Psychology in which he examined the memories of US citizens upon learning of the news of the assassination. Despite the very long time gap, out of 179 participants, 127 (71%) were able to give many details of the circumstances in which they heard about the news. Although the accounts were not always extremely long or detailed, all of them provided at least some elements regarding the time and the location where people were when they heard the news. Colegrove concluded his paper by insisting on the abiding character of vivid experiences.

2  Luminet and Curci

The next publication on FBMs occurred almost 80 years later and was also related to the very long term memories citizens of the US had about another famous president’s assassination, the one of Kennedy in 1963 (Brown & Kulik, 1977). Rapidly after its publication, a wave of studies was published on the topic of FBMs from the 1980s to the 1990s. During that period, 22 papers were published (Scopus database). Many of these papers focused on fierce debates about FBMs being potentially a special case of memory. While Brown and Kulik (1977) advocated that FBMs were almost immune to changes, later studies emphasized that the reconstructive aspects of these memories were highly active, thus making FBMs similar to ordinary memories (e.g. Neisser & Harsch, 1992; Wright & Gaskell, 1995). If FBMs are similar to any other ordinary memory process, it is not surprising that interest in them vanished soon afterwards. After some quiet years, the time between 2003 until 2009 can be considered as ‘the golden age’ of FBM research, with 60 papers published in only seven years.The first reason for this revival was the 9/11 attacks in the US, which occurred in 2001. This event had all the characteristics that are supposed to create strong and longlasting FBMs: high distinctiveness and unexpectedness, strong consequences and emotional reactions, and extended covert and overt rehearsal. For the first time in the history of research into FBMs, long follow-up studies were conducted regarding a particular target event. They showed that for long retention intervals, the consistency rate was far above normal memory processes (e.g. Hirst et al., 2015). The second reason for the sudden popularity of FBM research was related to a stronger interest for more phenomenological aspects of FBMs, such as vividness and confidence. When considering these two aspects, there is a strong consensus that FBMs are exceptional (e.g. Talarico & Rubin, Chapter 4, this volume). A third reason was related to the more systematic use of statistical analyses (e.g. taxometric models, structural equation modeling) that allowed models for the formation and ­assessment of FBMs to be tested (see Curci, Chapter 1; Luminet, Chapter 3, this volume).

Why a new edition of the book? The publication of the first edition of this book in 2009 corresponded to that golden age. In the introduction of that edition, we justified the publication using five rationales. We will present them here below and consider whether new advancements have occurred since then. First, in 2009, models explaining the formation of FBMs were proposed and tested (Finkenauer et al., 1998; Er, 2003).These models examined the consistency and vividness dimensions of FBMs. Since 2009, new models have been suggested, including one that examines the predictors of confidence in FBMs (Day & Ross, 2014). Second, much progress in statistical methods has been made (e.g. Curci, 2005). Structural equation models have become very popular tools in the field of FBM investigation, allowing researchers to test specific hypotheses concerning the formation and maintenance of FBMs. On the other hand, studies focusing on the measurement of FBM have tested the appropriateness of dimensional vs. categorical models to assess

Introduction  3

the FBM construct with respect to similar constructs in the field of autobiographical memory. These psychometric studies have contributed to the theoretical debate on the nature and specific characteristics of FBMs (see Lanciano & Curci, 2012). A third rationale that was presented in the first edition of this book was that two public events that supposedly met the conditions needed for the creation of FBM (high novelty and surprise, high personal importance/consequentiality, intense feeling states, and extended rehearsal) were investigated in a rather systematic way by researchers. These events were the death of Princess Diana in 1997, and the September 11th attacks in the US in 2001.A parallel investigation of the 9/11 attacks by several research groups and sampling different national groups allowed for a better understanding of the key variables involved in the process of FBM formation (e.g. Luminet & Curci, 2009). One main advancement since the first edition of the book was that longer follow-ups were performed for 9/11, for instance three and ten years after 9/11 on a US sample (Hirst et al., 2009, 2015). Other more recent public events have also been investigated (e.g. the death of Bin Laden or Michael Jackson) although the number of publications (two in each case) never reached the number of papers published on the subject of Princess Diana’s death (five) or the September 11th attacks (25). Fourth, we pointed out that the new literature on FBMs published after the death of Princess Diana and the September 11th attacks allowed researchers to examine from new directions the special vs. mundane character of FBMs. On this matter, we can point out strong advancements in the last years of research. This is largely due to a consideration of the more varied ways of assessing FBMs via self-report (focusing on consistency, level of detail, confidence, or vividness), and through the adoption of an experimental approach (Lanciano, Curci, & Sartori, Chapter 6, this volume). Future research still needs to be conducted to test the concurrent validity of these different approaches. Fifth, we noticed that the topic of FBMs appears to be equally relevant for various fields of psychology (social, cognitive, cross-cultural, and clinical). This is even more the case now, and this new edition makes it obvious that the connections between social and cognitive psychology are becoming even stronger (e.g. Hirst & Meksin, Chapter 10; Stone & Jay, Chapter 8, this volume). We also expected in the former edition that the content would be of interest to other fields of social sciences such as sociology, political sciences, or history. On this matter, we would like to outline briefly the strong potential of considering FBMs in the field of history more systematically (for more detail, see Luminet & Spijkerman, in press). Although we found that references to FBMs were quite scarce in history publications, some exceptions were noted. For instance, Clark (2012) gave the description of FBMs for the assassination of the archduke Franz Ferdinand and his wife in Sarajevo in June 1914. We also found many FBMs for the day of the Armistice for World War I in Belgium on November 11, 1918. Their occurrence seemed to be widespread as they were observed whatever the condition (soldier or civilian), the place of living (city or countryside), gender, or age. We firmly believe that there are several advantages of considering also FBMs from the remote past. Firstly, it allows accessing more ‘spontaneous’ accounts of FBMs.

4  Luminet and Curci

Indeed, a major shortcoming in the current studies is that they orient people’s mind into the search for contextual memories, even when they did not consider them spontaneously. Another issue is to examine the hypothesis that events leading to FBMs will be particularly well retained by people and thus more likely to be part of the collective memory of a group or a nation. Studying the presence of FBMs shortly after public events could thus give insight into how an event obtains lasting historical significance. A third advantage is to extract highly relevant information from archives. We showed the possibility to corroborate people’s accounts of FBMs in diaries or memoirs with controllable information happening at the time of the event, like the weather. The current edition also includes some important modifications. Two new chapters have been included that represent important developments in the field. Chapter 6 examines the possibilities of conducting experimental lab-based studies in order to provide better control on the factors that determine the formation of FBMs. It also suggests a new implicit measure to assess autobiographical memory, providing new ways to avoid the biases involved in traditional explicit assessment (Lanciano, Curci, & Sartori). Chapter 9 explores the relevance of examining positive FBMs and investigates the functionality hypothesis for FBMs in general (Stone and Jay). It also means that two chapters are no longer present in the current edition. The authors (Pillemer and Fivush, Bohanek, Marin, & McDermott Sales) did not wish to write an update. The main reason for not including the chapter by Pillemer on comparing FBMs and recall of first-hand experiences was that no major changes have occurred in this topic in recent years. The other chapter dealt with emotional memory and memory for emotions. The chapter covered very important issues regarding one specific predictor of FBMs, the intensity of emotional feelings when hearing news. Fivush et al. considered that their chapter was not central in a book dealing with FBMs. This is only partly true, and we highly recommend new readers who are not familiar with the first edition to read these two extremely valuable chapters. For the other chapters, we asked each author to make updates related to any important contribution that occurred in their field between 2009 and 2016. A very nice surprise when we received the new chapters was that many authors went far further than a simple update. They took the opportunity to consider very new and innovative ideas. Some of these chapters aim to be thought-provoking and deliver clear messages regarding the program of research for the years to come. It means that we are not only happy to present a complete update of the state-of-the-art research on FBMs. In addition, we can provide the FBMs research community with ideas and hypotheses that are ready to be tested, either in the lab, or when public news that is distinctive, consequential and emotional occurs. Finally, these ideas will improve the general research domain of autobiographical memory.

Summary of the chapters In the first chapter of the book, Curci compares two measurement approaches for the assessment of FBMs: dimensional (Latent Trait Analysis) models and categorical (Latent Class Analysis). The former assumes that psychological constructs are latent

Introduction  5

dimensions, so that the individual’s characteristics might be quantified along these dimensions; the latter assumes that the individual’s characteristics can be classified into some latent categories corresponding to a categorical latent psychological construct.The mathematical assumptions that underlie each of these approaches are presented and references to empirical evidence are discussed. Measurement models have carried significant assumptions – either explicit or implicit – about the nature and inclusion of FBMs in the general domain of autobiographical memory. Studies reviewed in the chapter demonstrate that dimensional measurement models fit well with ordinary autobiographical data, while FBMs are better modelled through a categorical approach. The application of a taxonomic model to a large dataset concerning the September 11th attacks (Lanciano & Curci, 2012) provides empirical support for the idea that FBMs represent a discontinuity in the field of autobiographical memory, and lead to important conclusions on their special nature. In Chapter 2, Wright and Carlucci discuss the criteria for a memory to be considered as an FBM.The issue of memorability is presented in term of its conceptual and methodological implications for the assessment of the FBM construct and the research hypotheses that can be investigated. The authors propose that either associative or causal hypotheses can be tested in FBM research and different designs have been adopted to test each of these hypotheses. It follows that researchers have been interested in either comparing different groups of people on their memory for a specific event, or testing a relatively small number of people on a large number of events. These methodological choices have influenced the conclusions that have been drawn on the nature and characteristics of FBM. In the final part of the chapter, the authors discuss the strengths and limitations of the single-case approach (involving the assessment of the memory for one event), which is still the most popular strategy in FBM investigation, although it has been rarely employed in cognitive psychology. The reason that FBM scholars adopt a single-case approach is that it allows for accuracy and consistency to be more easily assessed, but new avenues are available for FBM research to overcome the limitations of this approach. In Chapter 3, Luminet emphasizes the need to use statistical tools in order to test different models for the formation of FBMs. The vast majority of studies on FBMs examine public events through large surveys. In general, a set of variables thought to influence the formation of FBMs is assessed from a few days to a few weeks after the event has occurred, and then at least one follow-up is administered some months afterwards.This follow-up provides the possibility to compute a consistency score. Structural equation modeling (SEM) is presented in this chapter as a very fruitful approach that provides a statistical test of the hypothesized relationships among FBM variables. SEM provides a powerful tool for examining how the data can fit with a theoretical model of relationships between different constructs. It also provides a test for the path between each predictor represented by a set of dimensions (referred to as latent variables). And it allows for a distinction between direct antecedents of FBMs (like novelty and surprise) and indirect ones (like emotional reactions through the activation of rehearsal and event-memory). The chapter compares theoretical models developed by Conway et al. (1994), Finkenauer et al. (1998) and Er (2003). Since the last edition of the book, a new model has

6  Luminet and Curci

been proposed by Day and Ross (2014) in which, together with FBM consistency, the phenomenological aspect of confidence is also considered. For each model, the author adopts a critical view by assessing its ability to provide a strong theoretical background, an empirical validation, and a test of the appropriateness of the events considered. In Chapter 4, Talarico and Rubin provide a systematic screening of the available data regarding the old question of the potential uniqueness of FBM as a special memory mechanism. Indeed, they are the first authors since Brown and Kulik (1977) who have advocated for a special memory mechanism accounting for these memories, identified as exceptionally consistent, vivid, and held with high confidence, as conveyed by the photographic metaphor of “flashbulb-ness”. Talarico and Rubin examine this question through three main criteria: memory properties, conditions necessary to produce these memories, and the way individuals process these memories. Regarding memory properties, they suggest an important difference between vividness and confidence on the one hand, and longevity, accuracy, and consistency on the other hand. For the second criterion of the necessary conditions for FBMs, consequentiality, distinctiveness, and emotional affect are considered. The authors examine which one of these variables would best predict FBMs. Regarding the third criterion of how FBMs are processed, significance, surprise, emotional intensity, and rehearsal are presented as each having an important role.The conclusions address different directions regarding the effect of these variables on the three main characteristics of FBMs: consistency, vividness, and confidence.The systematic review presented in Chapter 4 also points to a clear agenda for future research in order to answer some of the current critical issues in understanding the potential uniqueness of FBM. Chapter 5, by Rice, Hamamouche, and Bohannon, contains entirely different content from the 2009 chapter in which the authors discussed the relationships between accuracy and consistency in FBMs. In this new chapter, after reviewing some general considerations on autobiographical memories, they go into detail about one of the three conditions presented by Talarico and Rubin in the ­previous chapter as being necessary to produce FBMs, namely consequentiality. Then, they present a taxonomy divided into three axes. The first distinction relates to the ­distance between the event and the respondent. Three levels can be distinguished here, the victim, the observer, and the distance learner. In the second axis, the nature of the threat involved by FBMs is considered in a dichotomous way. The distinction occurs between a threat that is ­ thers) or social (involving a change either physical (involving injury to the self or o in social relationships or in social position). The last axis is defined according to the valence of the event, distinguishing between ­pleasant, appetitive events, from unpleasant, aversive events. The authors take the election of Barack Obama in 2008 as a case study to examine the impact of these axes and categories of consequentiality on FBM formation. The chapter provides innovative avenues that elaborate on the measurement of a key condition for the formation of FBMs. In Chapter 6, Lanciano, Curci, and Sartori portray state-of-art methods of studying FBM in order to establish how the supposed ‘special quality’ of these memories can be captured. Research work on FBMs has been mainly conducted over the years

Introduction  7

through naturalistic approaches (retrospective recollections, surveys, or longitudinal methods). As stated by Wright and Carlucci in Chapter 3, despite the advantages, this view has some limitations concerning the control procedures for FBM assessment. In the first part of the chapter, Lanciano et al. discuss the possibility of establishing an experimental approach and review studies adopting experimental settings in FBM investigation. In the second part of the chapter, the authors review recent research that uses a new procedure to assess autobiographical memory based upon implicit measures (the autobiographical Implicit Association test), along with explicit measures. An important implication of these studies is that they provide support for a consistencybased approach to FBM assessment. Finally, the evidence reviewed entails important consequences for evaluating the accuracy of eyewitness FBMs in forensic contexts. Tat, Gold, and Budson, in Chapter 7, discuss how memory for traumatic national events relates to clinical disorders such as Alzheimer’s disease (AD), mild cognitive impairment (MCI), Korsakoff syndrome, amnesia, temporal lobe epilepsy, or post-traumatic stress disorder (PTSD). The authors review studies concerning FBM and factual memory for relevant public events, such as the September 11th attacks. While MCI patients appear able to form FBMs, AD patients exhibit somewhat impaired personal and factual memory for traumatic national events. Among amnesic patients, Korsakoff syndrome was associated with memory inconsistency for personal details and a poor semantic memory for factual details; patients with medial temporal lobe/diencephalic lesions had global memory deficits for personal and factual memory for the September 11th attacks, while frontal lobe patients were selectively impaired in personal memory for a highly emotional event. Individuals with temporal lobe epilepsy still appear able to form an FBM, although they exhibit a greater memory degradation compared with controls. Finally, individuals affected by a previous PTSD tend to show distorted long-term memory for personal details over time. The authors conclude their review by proposing methodological improvements to FBM assessment in clinical populations. Chapter 8 by Stone and Jay addresses the potential differences in FBMs for negative and positive events and the functions both fulfill. Regarding four criteria for assessing FBMs (confidence, consistency, vividness and degree of detail), systematic reviews suggest that some of these characteristics are stronger in FBMs experienced as pleasant or unpleasant. But the data can only be preliminary, as very important sources of variation occur across studies regarding the intensity of the events considered, the comparison point (negative vs. neutral), whether the exposure to the event is direct or indirect, and the delay of reporting the memory that is considered. In addition, too many studies have a small sample size, which raises the question of whether the results in the field of FBMs may be underpowered, and the risk of false negatives. The authors also suggest that the huge bias in favor of FBMs for unpleasant events stem from a higher distinctiveness attributed to negative events.They also insist that FBMs, even for intense ­negative events, cannot be considered equivalent to traumatic memories. The last part of the chapter examines the functions that can be fulfilled by FBMs by referring to the functional approach of autobiographical memory (e.g. Pillemer, 2003), which examines why people remember something over long periods of time (Bluck, Alea, & Demiray, 2010).

8  Luminet and Curci

In Chapter 9, Berntsen presents an overview of FBM modeling starting from the pioneering work of Brown and Kulik (1977) to the more recent study of Finkenauer et al. (1998), to show that empirical investigation focusing on the cognitive and emotional predictors of FBM at the individual level has led to inconsistent results. The author proposes a new categorization of FBMs, which emphasizes the crucial role of the social relevance of the eliciting event. Based on the standard version of social identity theory (Tajfel & Turner, 1979; Tajfel, 1982), Berntsen maintains that events that are appraised as highly relevant for a person’s social identity will be highly memorable. The durability of these memories over time will depend on whether the original events support or challenge positive values of the group to which the individual belongs. As a result of this process, FBMs display an exceptional vividness and confidence although they are not necessarily complete. Evidence is reviewed concerning the role of group membership in promoting the formation of vivid memories for public events, such as World War II. In the process of FBM formation, it is the activation of social identity that supports the immediate and delayed effects of emotional and rehearsal factors at the individual level and explains the long-term persistence of FBMs in social groups via social practices and exposure to mass media. FBMs are sometimes also remarkable because they are unique moments during which the course of history and the course of our lives are aligned. As observed by Neisser (1982), people will later report they were alive when a particular event happened and that they were there. In Chapter 10, Hirst and Meksin go one step further from this observation by showing that although the formation of an FBM is first an individual act, it can only be fully understood when considering that it happens within a community. “I was there” then becomes “I was there together with other members of my community who are equally affected”. In addition, the precipitating event is interpreted and mnemonically represented through the community’s lens. By fully considering the collective nature of most FBMs, Hirst and Meksin can connect them with a collective identity. They draw special attention in this demonstration to the way media shapes the collective memory dimensions included in a flashbulb event and, because of the presence of this FBM, the way the media enhances the influence of collective memory on collective identity. In Chapter 11, Páez et al. examine the overlaps and differences in the determinants of FBMs and event-memory considered at a collective level, which they equate with collective memories (CMs). They first examine similarities and differences between the content of FBMs and CMs. They observe that in both cases, political and collective events related to violence or catastrophes are often reported. But clear differences occur for events like war or political change. In the selection of events, systematic biases such as socio-centrism often occur. Secondly, they investigate similarities in processes, like the relevance of the critical period of identity formation (“reminiscence bump”) for the creation of both types of memories. Next, they review similarities and differences in various categories of predictors (appraisals of novelty, importance, consequentiality, surprise, other emotional reactions, rehearsal). Aside from a systematic comparison of these predictors, they also pay attention to some biases such as the higher recall of negative events over

Introduction  9

positive ones, or the attenuation of the unpleasantness of an event with time, biases that occur for both collective memories and FBMs. A next step which they propose is to examine the role of collective predictors such as collective identity, collective self-esteem, and congruence with cultural values. They also consider the nature of both types of memories, with FBMs being largely autobiographical and individual, while CMs are by nature collective and have mainly a semantic dimension. Wang and Aydin, in Chapter 12, focus on FBMs and event memories of public events to define the cultural factors that shape both forms of memories.The authors outline a culturally embedded model of FBM and discuss various cultural factors that shape the formation and consequences of FBMs in macro and micro contexts. These factors influence how and what types of public event news individuals receive and subsequently remember. They also moderate the effects of individual factors that have been commonly assumed to be universal (i.e. appraisal of importance, emotional state and surprise, rehearsal and social sharing), and further affect the lifespan distributions of FBM and public event-memory. In line with Berntsen (Chapter 9, this volume), Wang and Aydin emphasize the self-defining function of FBMs in shaping the social identity with respect to the group to which the individual belongs. Additionally, the cultural constraints and beliefs of individualistic vs. collectivistic countries are supposed to either promote or diminish memory for idiosyncratic details of the reception context. Finally, the authors discuss recent findings on the influence of social media on remembering. Finally, we are particularly happy that Justice, Akhtar and Conway accepted to read all the chapters of the book for a concluding chapter (Chapter 13) that emphasizes the main contributions from each author, and that they suggest an agenda for future research on FBMs. Martin A. Conway in particular has been involved in the field of research on FBMs for more than 20 years. His contribution, together with that of his colleagues, is thus invaluable. This new edition provides a nice combination of different generations of scholars active in the field of FBMs. Some of them were already involved when the book Affect and accuracy in recall: Studies of “flashbulb” memories was published by Winograd and Neisser in 1992. Others ran their first studies at the end of the 1990s to early 2000s, while others started only recently.They represent the new generation showing that the field is still very active with new, exciting, and intriguing questions and challenges to deal with. We hope that the reader will share the same interest and enthusiasm for the different topics addressed in this book.

References Bluck, S., Alea, N., & Demiray, B. (2010).You get what you need: The psychosocial functions of remembering. In J. H. Mace (Ed.), The act of remembering:Toward an understanding of how we recall the past (pp. 284–307). Malden, MA: Wiley-Blackwell. Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. Clark, C. (2012). The sleepwalkers: How Europe went to war in 1914. London: Harper. Colegrove, F. W. (1899). Individual memories. American Journal of Psychology, 10, 228–255.

10  Luminet and Curci

Conway, M. A., Anderson, S. J., Larsen, S. F., Donelly, C. M., McDaniel, M. A., McClelland, A. G. R., Rawles, R. E., & Logie, R. H. (1994). The formation of flashbulb memories. Memory & Cognition, 22, 326–343. Curci, A. (2005). Latent variable models for the measurement of flashbulb memories: A comparative approach. Applied Cognitive Psychology, 19, 3–22. Day, M. V., & Ross, M. (2014). Predicting confidence in flashbulb memories. Memory, 22, 232–242. Er, N. (2003). A new flashbulb memory model applied to the Marmara earthquake. Applied Cognitive Psychology, 17, 503–517. Finkenauer, C., Luminet, O., Gisle, L., El-Ahmadi, A., van der Linden, M., & Philippot, P. (1998). Flashbulb memories and the underlying mechanism of their formation: Toward an emotional-integrative model. Memory & Cognition, 26, 516–531. Fivush, R., Bohanek, J. G., Marin, K., & McDermott Sales, J. (2009). Emotional memory and memory for emotions. In O. Luminet & A. Curci (Eds.), Flashbulb memories. New issues and new perspectives (pp. 163–184). Hove: Psychology Press. Hirst, W., Phelps, E. A., Buckner, R. L., Budson, A. E., Cuc, A., Gabrieli, J. D. E., . . .Vaidya, C. J. (2009). Long-term memory for the terrorist attack of September 11: Flashbulb memories, event memories, and the factors that influence their retention. Journal of Experimental Psychology: General, 138, 161–176. Hirst,W., Phelps, E. A., Meksin, R.,Vaidya, C. J., Johnson, M.K., Mitchell, K. J., . . . Mather, M. (2015). A ten-year follow-up of a study of memory for the attack of September 11, 2001: Flashbulb memories and memories for flashbulb events. Journal of Experimental Psychology: General, 144, 604–623. Lanciano, T., & Curci, A. (2012). Type or dimension? A taxometric investigation of flashbulb memories. Memory, 20, 177–188. Luminet, O., & Curci, A. (2009). The 9/11 attacks inside and outside the US: Testing four models of flashbulb memory formation across groups and the specific effects of social identity. Memory, 17, 742–759. Luminet, O., & Spijkerman, R. (in press). ‘11 November 1918, an exceptional day!’ Flashbulb memories of the First World War Armistice in Belgium from a psychological and a historical perspective. Memory Studies, Special issue on the social cognition of history. Neisser, U. (1982). Snapshots or benchmarks? In U. Neisser (Ed.), Memory observed: Remembering in natural contexts (pp. 43–48). San Francisco: Freeman. Neisser, U., & Harsch, N. (1992). Phantom flashbulbs: False recollections of hearing the news about Challenger. In E.Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb memories” (pp. 9–31). New York: Cambridge University Press. Pillemer, D. B. (2003). Directive functions of autobiographical memory: The guiding power of the specific episode. Memory, 11, 193–202. Pillemer, D. B. (2009). “Hearing the news” versus “being there”: Comparing flashbulb memories and recall of first-hand experiences. In O. Luminet & A. Curci (Eds.), Flashbulb memories. New issues and new perspectives (pp. 125–140). Hove: Psychology Press. Tajfel, H. (Ed.). (1982). Social identity and intergroup relations. Cambridge: Cambridge University Press. Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks/Cole. Winograd, E. & Neisser, U. (Eds.) (1992). Affect and accuracy in recall: Studies of “flashbulb” memories. New York: Cambridge University Press. Wright, D. B., & Gaskell, G. D. (1995). Flashbulb memories: Conceptual and methodological issues. Memory, 3, 67–80.

1 MEASUREMENT ISSUES IN THE STUDY OF FLASHBULB MEMORY Antonietta Curci

Measurement models have carried significant assumptions – either explicit or implicit – about the nature and inclusion of FBMs in the domain of autobiographical memory. The chapter compares the dimensional (Latent Trait Analysis) and categorical (Latent Class Analysis) measurement models of FBM and shows that dimensional models fit well with ordinary autobiographical data, while FBMs are better modelled through a categorical approach.The adoption of a taxonomic model to a large dataset on the September 11th attacks provides empirical support to the idea that FBMs represent a discontinuity in the field of autobiographical memory, and lead to important conclusions on their special nature.

Measuring Flashbulb memory (FBM) is one of the key issues in the investigation of the phenomenon. Put in another way, it is the problem of construct validity (Carmines & Zeller, 1979) applied to the investigation of FBM. Methodologists usually recommend a careful consideration of the degree to which a given measurement model matches that which it claims, or purports, to be measuring. In doing so, a clear theoretical definition of the construct under analysis needs to be provided along with an accurate specification of the employed measurement model. In the field of FBM, researchers have adopted different theoretical views regarding the phenomenon (see also Wright & Carlucci, Chapters 2; Talarico & Rubin, Chapter 4, this volume). Both explicitly and implicitly, measurement models of FBM have carried significant assumptions about the nature and inclusion of the phenomenon in the general domain of autobiographical memory. The present chapter firstly provides an overview of different measurement models traditionally adopted in assessing FBMs, by showing the limitations of these models and the various attempts to overcome these. Subsequently, a brief outline of the peculiarities of both dimensional and categorical models will allow the reader to evaluate their appropriateness in the investigation of FBM. Finally, from both the theoretical and

12 Curci

empirical point of view, similarities and differences between FBMs and other forms of ordinary autobiographical memories (such as memory for the original event) are considered. Generally, categorical models appear to best account for FBMs as particularly vivid and detailed autobiographical memory formations. A practical implication for research is the need to improve FBM assessment to take into account the categorical nature of the construct, also relying upon the current availability of many advanced statistical tools.

The measurement of FBM The problem of measuring FBMs dates back to the first research work on the topic. In their original paper, Brown and Kulik (1977) defined FBMs as memories from the attributes of the reception context of shocking public news. In other words, people may retain, for a long time, not only the original event itself, but also the reception context for this event, that is, the place where they were, the time when they learned of the event, their ongoing activity, the informant, the personal reactions and reactions of others, and the aftermath of the event (Bohannon, 1988; Brown & Kulik, 1977; Conway et al., 1994; Larsen, 1992). The present chapter reviews studies focusing on the way manifest indicators of the reception context of a shocking event are related to the latent construct of FBM. Indeed, when assessing FBMs, researchers not only make assumptions concerning the specific features to be measured (vividness, consistency, confidence, longevity, etc.), but also on the mathematical model connecting these observed features with the latent construct. This, in turn, has noteworthy implications for the theoretical advancement of research on FBMs. In their original study, Brown and Kulik (1977) operationalized FBMs in two ways. First, individuals were scored as having an FBM if they answered “yes” to the direct question: “Do you recall the circumstances in which you first heard that. . .?”. Second, an FBM was identified if people could remember at least one attribute of the reception context (Brown & Kulik, 1977; Pillemer, 1984). In assessing the phenomenon, Brown and Kulik (1977) assumed that a simple counting of attributes of the reception context would represent a good approximation to the construct of FBM.When applying such a procedure, the outcome of the assessment is an absolute scale, in principle without upper boundary, since in real life the total number of members of a collection cannot be a priori defined (Luce & Suppes, 2002). Thus, the number and characteristics of attributes chosen to define FBM is completely arbitrary. Brown and Kulik (1977) selected six attributes of the context, which they considered more indicative of the nature of FBMs, and all six attributes were given the same relevance in their model. However, the authors’ choice was neither theoretically justified nor empirically supported. Unfortunately, a simple counting of attributes is completely uninformative of the intrinsic nature of the construct in analysis. In spite of this, the procedure of summing up FBM attributes (Brown & Kulik, 1977) was subsequently adopted by researchers (Bohannon, 1988; Kaya Kizilöz & Tekcan, 2013; Kvavilashvili et al., 2003; Pillemer, 1984; Talarico & Rubin, 2007).

Measurement issues in Flashbulb memory  13

Other studies have attempted to reduce the major flaw of Brown and Kulik’s measurement model (1977) by prioritizing specific details of the context. Winograd and Killinger (1983) identified an FBM from the mention of ongoing activity in participants’ recollections. Wright (1993) required that each participant recalled at least one attribute from location, other present people, and ongoing activity. Despite having evident limitations, these approaches introduced the idea that FBM is more than a simple collection of irrelevant attributes, since some of them are more representative of the construct than others. Neisser and Harsch (1992), in their study on the Challenger disaster, employed a procedure called WAS (Weighted Attributes Scores), which assigned different weights to different attributes of the reception context. The authors assumed that the attributes were not all equally important. Some of them were defined as “major” (i.e. location, informant, and ongoing activity), since they seemed to be essential to identifying the reception context. Other attributes were considered “minor” (i.e. other present people, and time), since one could be wrong on these and still essentially accurate about the major details. The WAS procedure consists of assigning a score for each recalled major attribute, plus a bonus point when the subject scores above a given threshold on the set of minor attributes. More specifically, scores ranging from 0 to 2 were assigned to the individual’s memories for location, informant, and ongoing activity. Furthermore, an additional score of 1 was added if the participant scored at least 3 on the indicators assessing other present people and time (on two scales also ranging from 0 to 2). As a consequence, for each individual, the final WAS measure for FBMs ranged from 0 to 7 (Neisser & Harsch, 1992). The WAS system was subsequently used in more recent studies on memory for the September 11th attacks (Kvavilashvili et al., 2010; Kvavilashvili et al., 2009; Pezdek, 2003; Shapiro, 2006; Smith, Bibi, & Sheard, 2003; Tekcan et al., 2003), and it represented a clear advancement towards a measurement model which considered FBM as a qualitatively different phenomenon from ordinary memory formations. More recently, Curci et al. (2001), Curci and Luminet (2006), and Luminet et al. (2004) modelled FBM data through a statistical technique called CatPCA (Categorical Principal Component Analysis; see Gifi, 1990), which shares the same logic of the WAS approach (Luminet et al., 2004). CatPCA is a principal component analysis specifically aimed at scaling categorical or ordered categorical variables (van de Geer, 1993). In employing this technique, the authors aimed at getting a composite measure, which combines scores of different indicator variables corresponding to the attributes of the reception context, weighted with respect to their relevance in identifying FBMs. With respect to WAS, the main advantage of CatPCA is that the weights assigned to the scores are not decided a priori by the researcher, but derive from the empirical distributions of the indicator variables in the sample of respondents in the study (Greenacre, 1993). Luminet et al. (2004) discussed the similarities and differences between the WAS and CatPCA approaches, showing that, although the scores for the two procedures were highly correlated, CatPCA was preferred because its scores are obtained from an empirical basis and not from researchers’ aprioristic assumptions.

14 Curci

In sum, the literature reviewed thus far substantiates some relevant points: (1) Traditional measurement models appear inadequate to assess FBMs; (2) Evidence concerning the measurement of FBM have implications for the theoretical understanding of the nature of the phenomenon; (3) Significant implications can be derived for a general model of autobiographical memory.

The assessment of latent constructs Dimensional vs. categorical models Researchers in psychology usually deal with unobservable, or latent, constructs. For instance, intelligence, introversion, and mathematical ability cannot be directly measured as can height or weight, since they are concepts rather than physical entities. Nevertheless, some empirical attributes can be considered as observable indicators of underlying latent constructs (Baker, 1985). One basic assumption of latent variable models is the local independence of the indicator variables given the latent variable(s) (Lazarsfeld & Henry, 1968). Postulating the local independence means that, after removing the variability between the latent construct and its observed indicators, the individual’s responses to the items of a test are statistically independent of each other (Hambleton, Swaminathan, & Rogers, 1991; McCutcheon, 1987). The so-called dimensional models assume that a continuous, normally distributed latent construct accounts for variations among observed indicators (Moustaki, 1996). For instance, the ability to provide correct answers to a set of mathematical calculations might be conceptualized as a continuous trait (i.e. the mathematical ability) which each individual holds to a different degree. Psychologists are familiar with factor analysis models, which consider observed variables as linear combinations of latent factors, plus random error terms. Factor analysis models are most useful with continuous indicator variables, for which correlation coefficients can easily be computed. Fitting these models involves finding the values of the latent variable parameters which maximize the probability of reproducing a correlation matrix of indicators as close as possible to the observed product moment correlation matrix (Comrey & Lee, 1992; Kim & Mueller, 1978). When the aim of the researcher is to reduce a set of categorical observed indicators (binary, ordered categorical, simply categorical) into a smaller set of latent factors, the Latent Trait Analysis (LTA) is one of the most accepted dimensional approaches (Bartholomew et al., 2002; Rost & Langeheine, 1997). In the area of educational testing and psychological assessment, LTA is termed Item Response Theory (IRT). A considerable overlap exists between LTA and IRT, leading the authors to consider them as basically interchangeable approaches. Among psychologists, the acronym IRT is more popular, and it will be preferred in the following pages. Item Response Theory has two basic postulates: (1) The individual’s performance on a test item depends on some latent abilities called traits; (2) The relationship between the individual’s item performance and latent traits can be described

Measurement issues in Flashbulb memory  15

by a monotonically increasing function (Embretson & Reise, 2000; Hambleton et al., 1991).This function, called an IRT Item Characteristic Function, expresses the probability of an individual giving a correct answer to an item of a test as a function of his/her ability (trait) (Baker, 1985). IRT methods were originally unidimensional, assuming that only one latent trait accounts for variations of observed indicators, but generalizations to multidimensional models have been more recently proposed (Embretson & Reise, 2000). In the present chapter, unidimensional IRT models will be discussed, since the application of IRT methods to FBM data has been limited to this approach (Curci, 2005; Curci & Lanciano, 2009; Wright, Gaskell, & O’Muircheartaigh, 1998). The theoretical rationale of the models and some details on the procedure will only be presented here, since a technical discussion of their mathematical features is beyond the scope of the present chapter (for more details, see Bartholomew et al., 2002; Embretson & Reise, 2000; Hambleton et al., 1991). Unidimensional IRT models differ from each other with respect to the form of the Item Characteristic Function, and the number of parameters in the model. For example, in the investigation of a student’s mathematical ability, a starting assumption is that, in providing the correct answers to a set of mathematical computations, each individual possesses some amount of the ability under investigation. In other words, each student has a hypothetical score on the scale of a latent construct corresponding to her/his mathematical ability. This construct cannot be directly assessed, but, for each student, its amount is inferred from the number of correct responses provided to the set of mathematical computations. IRT models allow the researcher to estimate the probability of an individual to choose some response categories from a set of observed indicators (i.e. correct responses to the mathematical computations) as a function of an underlying latent trait (i.e. the mathematical ability). In doing so, IRT models describe each observed indicator with respect to some properties (difficulty, discrimination, etc.) which define the position of the individual along the continuum represented by the latent trait. If we denote the ability score as theta, then, at each level of theta, a different probability is associated to obtaining the correct answer to a given computation. This probability is denoted as P(theta). For each item of the set, plotting P(theta) as a function of theta will result in a smooth S-shaped curve, called an Item Characteristic Curve (ICC). At the lowest levels of theta, the probability of a student providing the correct answers to the considered mathematical computation is very low, and it increases as the amount of possessed mathematical ability increases following the shape of the ICC. The one-parameter logistic model (or Rasch model, Rasch, 1960) simply considers the location of the curve along the x-axis, by including in the equation for P(theta) only the so-called difficulty parameter. The higher the difficulty parameter for a given item, the higher the ability required for correctly answering that item, the more the ICC will be located along the right side of the x-axis. In Figure 1.1, the ICC on the left has a lower difficulty parameter than the ICC on the right, and this means that less ability is required in order to provide the correct answer to the item represented by the former ICC as compared with the item represented by the latter. To illustrate, much more mathematical ability is

16 Curci 1 0,9 0,8 0,7 0,6 p(theta) 0,5 0,4 0,3 0,2 0,1 -7

FIGURE 1.1 

-6

-5

-4

-3

-2

-1

0 1 theta

2

3

4

5

6

7

ICCs for one-parameter IRT models.

required to solve an algebraic equation than to perform the 12 x 8 computation. It follows that the ICC for the former operation will be located on the right side of the x-axis more than the ICC for the latter operation item. The two-parameter logistic model also includes in the equation the so-called discrimination parameter, which reflects the slope of the ICC. The steeper the ICC, the more efficiently the item discriminates respondents having high latent ability from those having low latent ability. In Figure 1.2, the ICC on the right side of the plot corresponds to an item with a higher level of discrimination than the ICC on the left side.To provide a concrete example of the discrimination parameter, solving an algebraic equation will better discriminate high school students with different levels of mathematical ability, than young primary school students or expert college students attending mathematics courses at the university. The three-parameter logistic model takes into consideration the probability of getting correct answers due to simple guessing by respondents. This means that the model equation also contains a pseudo-chance parameter, which represents the probability of an individual correctly answering the item only by guessing. In mathematical words, the pseudo-chance parameter corresponds to the lowest nonzero asymptote for the ICC. In Figure 1.3, the two ICCs on the right side include this pseudo-chance parameter, since the probabilities of getting correct answers to the corresponding items are never lower than some thresholds, while the ICC on the left side does not enclose this parameter. To provide an example of what the pseudo-chance parameter represents, the probability of distinguishing chemical formulas from mathematical operations is reasonably above zero for high school students. It is worth noting that the IRT modelling includes both the individual’s responses to a set of items, and the properties of the items, which constitute the test.

Measurement issues in Flashbulb memory  17 1 0,9 0,8 0,7 0,6 P (theta) 0,5 0,4 0,3 0,2 0,1 -7

-6

-5

-4

-3

-2

-1

-0

1

2

3

4

5

6

7

theta

FIGURE 1.2 

ICCs for two-parameter IRT models. 1 0,9 0,8 0,7 0,6

P (theta) 0,5 0,4 0,3 0,2 0,1 -7

-6

-5

-4

-3

-2

-1

-0

1

2

3

4

5

6

7

theta

FIGURE 1.3 

ICCs for three-parameter IRT models.

The Item Information Function (IIF) and Test Information Function (TIF) indices specify the amount of information on the latent construct provided respectively by every single item and the total set of the items of the test (Baker, 1985). Finally, general fit indices can be computed to evaluate the correspondence between the whole estimated model and the data entered in the analyses (Akaike, 1974; Myung, 2003; Schwarz, 1978). IRT and LTA modelling is available in many software packages for data analysis, among others BILOG-MG (Zimowski et al., 1996), GENLAT (Moustaki, 2001), IRTPRO (Paek & Han, 2013), MULTILOG (Thissen, 1991), PARSCALE (Muraki & Bock, 1998), Xcalibre (Guyer & Thompson, 2011), and R (R Core Team, 2014) package ltm (Rizopoulos, 2007). A different approach in psychological assessment is represented by categorical modeling, which considers individuals to be classified into discrete classes having specific characteristics. To illustrate, at odds with the dimensional approach, the

18 Curci

stage theory of mathematical learning is based upon the idea that mathematical abilities are related to the peculiar phase of the child development, and are not simply the outcome of an incremental acquisition of knowledge (Dienes, 1960). In the field of clinical psychology, individuals might be classified as belonging or not to a class of patients based upon the co-occurrence of a set of observed symptoms. Following this categorical perspective, data can be modelled following two general approaches: cluster analysis and Latent Class Analysis (LCA) (Meehl, 1992). Cluster analysis divides individuals on account of their similarity/dissimilarity from each other. The procedure terminates once a division is reached, based upon empirical features of the sample. Conclusions cannot be drawn regarding the latent structure of data (Meehl, 1992). For this reason, cluster methods cannot be included with latent variable models. LCA models assume that a discrete latent construct accounts for covariations among observed indicators (McCutcheon, 1987). The latent construct has different latent classes with different probabilities of occurrence. Each class of the construct corresponds to a specific response pattern of indicators of the test. This means that the probability of occurrence for each latent class of the construct is determined by the percentage of individuals in the sample exhibiting a given response pattern to the observed indicators.The meaning of a latent class is defined by the response categories of the indicators associated with that latent class of the construct (McCutcheon, 1987). This is the same as saying that the latent class t of a construct X is defined by the response category i to the item A, j to the item B, k to the item C, and so forth. To illustrate, the latent class “depression” is defined by the positive answers provided by respondents to the items of a Major Depression scale (i.e. feeling depressed mood, insomnia/hypersomnia, fatigue, diminished ability to concentrate, etc.), while the “non-depression class” is defined by negative answers to the listed symptoms. Associations between latent classes and response categories of indicators are probabilistic. Indeed, for each response category, conditional probabilities are estimated, which correspond to the probability of choosing that response category of the observed indicator, given the latent construct. Also for LCA, indices of fit can be computed to assess their appropriateness to account for the empirical associations among observed indicators in analysis (McCutcheon, 2002). Among the most common statistical packages to run LCA, there are lEM (Vermunt, 1997), Latent GOLD (Vermunt & Magidson, 2005), MPLUS (Muthen & Muthen, 1998), PANMARK (Van der Pol, Langeheine, & de Jong, 1998), SAS macros (Dziak et al., 2015), Stata plug-in (Huang et al., 2016), and R (R Core Team, 2014), package e1071 (Meyer et al., 2015). Besides the mathematical specifications of the latent variable approaches so far reviewed, the main theoretical distinction between dimensional and categorical models refers to the nature of the underlying latent variable responsible for covariations among observed indicators. The application of one model instead of another implies a theoretical assumption on the nature of the construct under investigation. Furthermore, given that latent variable models are confirmative, and fit indices

Measurement issues in Flashbulb memory  19

can generally be computed to evaluate their appropriateness to a given dataset, the researcher has the opportunity to select the best fitting model to the data in analysis, so that the theory can be submitted to scrutiny on empirical basis (Kim & Mueller, 1978).

Beyond dimensionality: Categories as “real” entities The traditional measurement approach in psychology has favored dimensional models, especially in the investigation of personality characteristics (Eysenck, 1953). As specified above, factor analysis models are still the most used methods by researchers interested in combining observed indicators into theoretically meaningful constructs and their outcomes are currently interpreted as continuous dimensions. A categorical interpretation of latent constructs is still considered as an oversimplification of the reality, or, what is worse, an arbitrary account. This is due to the fact that psychological reality is considered so complex and fluctuating that it can hardly be constrained into discrete classes (Gangestad & Snyder, 1985). This consideration – very common among personality researchers – has also influenced many areas of psychological research, and might also be extended to the investigation of autobiographical memory. Indeed, one of the main arguments against categorical models relies upon the fact that classes are usually considered phenetically. This means that classification mainly serves descriptive purposes, as a summarizing function. No underlying meaning is assigned to the classes, which are completely arbitrary or quasi-arbitrary constructions (Gangestad & Snyder, 1985). On the other hand, classes might be established in genotypic terms, when empirical evidence suggests there is a latent entity, event, or construct which aggregate some observed characteristics. As a consequence, causal relationships might be assessed between this latent construct and the observed indicators (Gangestad & Snyder, 1985). Classes are thus meaningful, they really exist, and the goal of empirical investigation is to access the latent categorical structure of observed reality.Taxometric analysis takes for granted this point, since “taxa” are defined as natural kinds or species, not merely arbitrary categories (Ruscio, Haslam, & Ruscio, 2006; Schmidt, Kotov, & Joiner, 2004; Waller & Meehl, 1998). Three sets of reasons justify the concept of taxonicity in reality; common sense explanations, causal interpretation of reality, and mathematical demonstrations (Meehl, 1992), and all these reasons are well suited to classification procedures in medicine, biology, or chemistry. This kind of categorization appears acceptable when classifying animals, stones, organic diseases, and so on. In psychological research, accepting the conjecture of a discrete latent categorization is much more complicated, therefore empirical evidence might provide large support to this approach. Empirical distributions of indicators, and observed relationships among them, can suggest considering the taxonicity hypothesis for a given phenomenon, to be supported by stringent statistical demonstrations (Meehl, 1992). In psychological research, taxonomic models have been adopted to investigate relevant constructs for the study of personality and

20 Curci

personality disorders by Meehl (1992, 1999, 2004; Meehl and Golden, 1982). In his work on taxometrics, the author has criticized the dogmatic and a priori preference for the dimensional approach in personality assessment, and has proposed sophisticated procedures for identifying latent taxa in psychopathology. The work by Waller, Putnam, and Carlson (1996) on dissociation represents an interesting application of these procedures, to distinguish pathological from ­non-pathological dissociation. Evidence from this work supports the idea that non-pathological dissociation is a dimensional construct to be assessed through dimensional modelling, while pathological dissociation is a typological construct best modelled by a taxometric approach. Clinical research has applied latent class models to account for a large set of disorders, such as autism spectrum disorders (Kyriakopoulos et al., 2015), depression (Hankin et al., 2005), anxiety (Rhebergen et al., 2014), narcissism (Fossati et al., 2005), eating disorders (Gleaves et al., 2000; Williamson et al., 2002), substance addiction (Mauro et al., 2016), and dissociation (Waller et al., 1996;Waller & Ross, 1997). In all of these studies, empirical support is provided for the hypothesis of a categorical latent structure underlying covariations of a set of discrete indicators. To sum up, in order to define a phenomenon as a real, non-arbitrary categorical formation, two requirements are needed: (1) A theoretical model concerning the categorical nature of the phenomenon; (2) Evidence of a kind of discontinuity implanted in a network of empirical relationships. The next section is intended to apply these principles to theoretical modelling and empirical research on autobiographical memory and FBMs.

FBM and autobiographical memory Semantic vs. episodic memories FBMs have been defined as memories for the reception context of relevant public events. However, individuals’ recollections of their private reception context are undoubtedly associated with recollections of the original event itself (eventmemory, EM). Empirical evidence shows that emotional and reconstructive factors have a different impact upon FBMs and EMs. In their study on FBMs for the death of President Mitterrand, Curci et al. (2001) found that EMs did not significantly decrease over time, whilst memory for FBM attributes was impaired. Moreover, French respondents, who were more concerned by the death of Mitterrand than Belgian respondents, appeared to be more consistent in their EM, but no effect of provenance was found on FBM consistency. In another study, Pezdek (2003) compared memories for the September 11th attacks in two samples of people living in New York and in California-Hawaii, respectively. New York participants reported the most accurate EMs, and the least accurate memory for the reception context. The opposite pattern was found for the California-Hawaii sample (Pezdek, 2003). Finally, Hirst et al. (2009, 2015) observed different patterns of changes in FBMs and

Measurement issues in Flashbulb memory  21

EMs for the September 11th attacks in a very large sample of US respondents, and their findings account for a significant difference in retention and forgetting of the two classes of memory. More specifically, inconsistent FBMs were more likely to be repeated rather than corrected over the years, while inaccurate EMs were more likely to be corrected. From a theoretical point of view, FBMs and EMs have been considered as belonging to different memory systems (Bohannon & Symons, 1992; Finkenauer et al., 1998; Pezdek, 2003; Smith et al., 2003).While FBMs have been conceptualised as a form of episodic memory, EMs have been considered as a form of semantic memory (Bohannon & Symons, 1992). In their study on the Challenger explosion, Bohannon and Symons showed that the degree of involvement in a public event is a better predictor of the persistence of episodic than semantic material. By contrast, Larsen (1992) has argued that Bohannon and Symons’ view “rests on a questionable application of the episodic-semantic distinction” (p. 37). In his opinion, it is difficult to consider a memory for the explosion unfolding on the television screen as more generic or semantic than the individual’s memory for the other people present when the news was announced. It follows that distinguishing FBMs from EMs on the basis of the occurrence of semantic vs. episodic material is ambiguous. In a more recent study involving different samples of individuals recalling different public events, Bohannon, Gratz, and Cross (2007) proposed that the presence of semantic and episodic elements in FBM accounts could be ascribed to the effect of different sources of information, in that the experience of learning the news from another people implies a privileged retention of idiosyncratic details of the reception context (i.e. a typical FBM). By contrast, the experience of learning from the mass media leads the individuals to keep sematic information of the news event itself. The theoretical debate just reviewed is reflected in the measurement approach to FBMs. Nachson and Zelig (2003) suggested that to avoid confusion on the assessment of FBMs vs. EMs, a narrative measure of EM has to be constructed, identical to Brown and Kulik’s (1977) measure of FBM for the reception context. Unsurprisingly, in their study on the assassination of Israel’s Prime Minister, Itzhak Rabin, the authors found no difference between FBM and EM, since the two measures were designed to be equal. Conversely, studies that have explicitly compared measurement models for FBM and EM (Curci, 2005; Curci & Lanciano, 2009) have demonstrated that both forms of memory include semantic and episodic materials. Finally, in an experimental study adopting the implicit association method drawn from the Social Cognition paradigm (Greenwald, Nosek, & Banaji, 2003; Sartori et al., 2008), Curci et al. (2015) overcame the classical distinction between episodic and semantic elements in FBMs and EMs. The authors showed that different processes are activated when individuals were requested to recall either the reception context of a news item or the original event, with FBM recalls being activated by automatic and implicit processes and EMs mainly relying upon ordinary reconstructive processes.

22 Curci

In sum, research work so far reviewed has failed to distinguish FBMs from EMs solely from the presence of episodic vs. semantic details in the individual’s accounts, given that episodic and semantic pieces of information are included in both forms of memory. FBM assessment thus needs to be based on other distinctive features of the underlying latent construct.

The construction of autobiographical memories Research on autobiographical memories has pointed to some distinctive characteristics of these representations. First, autobiographical memories differ across the life-span for the same individual. A young teenager’s memory of the first day at school is different when the same individual is 40 years old (Conway & Rubin, 1993). Second, autobiographical memories vary in vividness (Anderson & Conway, 1997). Finally, a mix of general and specific knowledge can be found in every autobiographical recollection (Conway & Pleydell-Pearce, 2000; Conway, Singer, & Tagini, 2004). With respect to the first point, it should be noted that a constructivist view on autobiographical memory has put a great emphasis on a process called “cyclic retrieval” (Williams & Hollan, 1981). This is a constructive process, which mediates the formation of autobiographical memories, and it is based upon an iterative comparison between the outcomes of a search phase and the elaboration of memory descriptions from available cues (Conway & Rubin, 1993). To illustrate, when recalling my first day at school, I repetitively verify my recollection against criteria taken from the goals I am currently pursuing and from external sources (i.e. parents’ narratives, friends’ accounts, cultural rules, feeling states). Moreover, I might fill in any memory gaps with these pieces of information. As a consequence of this process, memories cannot be considered stable structures, which can be accessed at any time; rather they are transitory patterns of activation within an autobiographical knowledge base. The construction of autobiographical memory is dependent upon the activation of personally relevant goals, and issues related to the self (Conway & Rubin, 1993). This means that individuals reconstruct their past with respect to the goals they are pursuing in the present, and the process of retrieval develops in conformity with the organization of the working self (Carver & Scheier, 1990; Higgins, 1987). Access to knowledge that corresponds to the goal structure of the working self is facilitated, while access to knowledge incompatible with these goals is inhibited. On the other hand, autobiographical memories ground the self, in that they provide consistency and plausibility constraints on the goals that can be held (Conway et al., 2004). For this reason, a teenager’s memory for the first day at school, highly concerned with the goal of being a good pupil, is fairly different from the memory retrieved by the same individual when 40 years old when other goals have become relevant. In regard to FBMs, recollections of individuals personally or socially concerned by a given event appear to be more vivid and consistent than memories of low concerned individuals (for a review, see Rice, Hamamouche, & Bohannon, Chapter 5, this volume).

Measurement issues in Flashbulb memory  23

With respect to the second and third issues, concerning the vividness and specificity of FBMs, drawing from the concept of Memory Organization Packets (MOPs) by Schank (1982), Conway and colleagues have proposed a model which considers autobiographical memories as generated from a multilevel knowledge base through the process of activation described above (Anderson & Conway, 1997; Conway & Bekerian, 1987; Conway et al., 2004; Conway & Pleydell-Pearce, 2000; Conway & Rubin, 1993). The structure of this knowledge base encompasses at least three layers. The first contains thematic knowledge about extended periods of time, the life-time periods (e.g. when I was at University, when I lived at X); the second – general events – contains records of extended and repeated events occurring over periods of months or weeks (e.g. my holidays in France, Saturday nights at the disco); the third layer – event-specific knowledge – comprises highly specific and vivid details, as well as forms of images and associated feelings. FBMs can be considered forms of event-specific knowledge since they exhibit features of vividness and specificity that characterizes this layer of autobiographical knowledge (Conway & Pleydell-Pearce, 2000).This model has the advantage of preserving the specificity of some forms of episodic recollections, such as FBMs, but at the same time it considers the dynamic interconnection existing between these unitary representations and the most general level of thematic knowledge. In line with the arguments presented above, FBMs, more than ordinary autobiographical memories, correspond to recollections of event-specific sensory-perceptual details, thus they appear more vivid than any other autobiographical memory (Conway, 1995). Unlike ordinary autobiographical memories, which convey a distributed form of knowledge representation involving the three layers of the autobiographical knowledge base, FBMs appear to arise from more densely integrated regions of such a knowledge base. What is immediately salient about FBMs is the fact that people report vivid accounts of their experiences of learning important news and rate their recollections stable and consistent even over long periods. From the theoretical point of view described above, the idea of a categorical nature of FBM appears to be preferable to account for these characteristics. FBMs are not simply ordinary memories with an unusual feature of vividness, but, given their unusual phenomenological characteristics, they can be considered as integrated cores of sensory-perceptual event-specific information in the space of autobiographical memory (“whole” units; Conway, 1995).

Evidence of discontinuity in the domain of autobiographical memory The latent variable approach was adopted to model FBM data in a small set of studies (Curci, 2005; Curci & Lanciano, 2009; Wright et al., 1998). Following this approach, FBM is considered a latent construct, which cannot be directly accessed, and all relationships within a set of indicators are accounted for by covariances between those indicators and the latent construct. Wright et al. (1998) were the first to employ such modelling in a study on the memories of Mrs. Thatcher’s

24 Curci

resignation and the Hillsborough disaster, where FBM was considered as a continuous trait underlying a set of observed categorical indicators. Curci (2005) compared three latent variable approaches, which only differed from each other with respect to the mathematical assumptions concerning the nature of the latent construct: Confirmatory Factor Analysis (CFA; Bollen, 1989), and Latent Trait Analysis (LTA; Bartholomew et al., 2002) hypothesize that a continuous latent construct underlies a set of observed indicators; the Latent Class Analysis (LCA; McCutcheon, 1987, 2002) assumes that the latent construct is a categorical variable. The hypothesis under investigation was the following: if FBMs are triggered from more densely integrated regions of the autobiographical knowledge base (Conway, 1995), then the measurement model for FBMs should differ in its assumptions and outcome from the model used to assess ordinary autobiographical memories. While dimensional measurement models would fit ordinary autobiographical memory data well, FBMs would be better modelled through a categorical approach (Curci, 2005).The three models were adopted on data concerning memory for the reception context in both a correlational and experimental study. Overall, findings confirmed the idea that FBMs convey an integrated form of knowledge representation (Conway, 1995), while ordinary autobiographical memories require the combination of distributed information.The LCA model appeared to be a good approximation to the clustered nature of FBMs, while the LTA model represents more reasonably the alternative account of ordinary autobiographical memories as continuous memory formations (Curci, 2005). A similar conclusion was reached by Curci and Lanciano in the 2009 study on the September 11th attacks, where both dimensional and categorical models were applied on data collected on a large sample of US and European respondents. The authors found that FBM indicators of the two subsamples could be adequately modelled through a categorical approach, while for EMs both LTA and LCA exhibited some inadequacies. Results from this study provided further confirmation for the idea that FBMs represent very integrated cores in the field of autobiographical memory. It follows that any dimensional assessment of FBM may be insufficient to capture the phenomenon, and FBM investigation needs to be implemented by categorical measurement approaches. With respect to EM, the authors suggested that the confusion in the measurement model might arise from an unclear operationalization of the construct, which would encompass not only semantic information of the original event, as originally postulated by Bohannon and Symons (1992), but also episodic material, strictly embedded with personal details of the reception context. Finally, Lanciano and Curci (2012) presented a taxometric investigation of data concerning the death of Pope John Paul II, by applying three different taxometric procedures (Ruscio, 2007, 2016) to corroborate evidence on the latent structure of FBMs. Through a sophisticated statistical analysis, the authors confirmed that from the empirical point of view, FBM indicators aggregate into a categorical construct, and that a dimensional model is barely applicable to FBM data. The FBM taxa correspond to dense autobiographical formations in which different

Measurement issues in Flashbulb memory  25

phenomenological features converge, i.e. the richness of perceptual details, a sensory vividness, the individual’s feeling of confidence, stability over time, and slow forgetting (Conway, 1995). It should however be considered that the conclusions on the measurement models for FBM do not exclude the role that reconstructive processes have on the persistence of these formations. In the 2005 study, Curci showed that contrary to expectations, the construct of FBM was found to be significantly associated with the detail of the context corresponding to the mass media as a source of learning. This was probably due to the fact that the original event considered in the study was expected and the mass media contributed considerably in spreading information about its protagonist. Furthermore, the data collection for the study took place three months after the original event, thus distortions could have affected individuals’ memory (Neisser & Harsch, 1992). Indeed, it could be the case that people watched TV and read newspapers in the days immediately following the event and that their memory for their first source of learning shifted towards the mass media script.Thus, they would have simply remembered a source of information they had encountered later, and this detail would have become strongly interconnected with the other details of the context to create a unique memory formation. As a consequence, FBMs appeared vivid and consistent over time, although deeply determined by reconstructive factors. This point is also clearly in line with the constructivist models (Conway et al., 2004; Conway & Pleydell-Pearce, 2000; Conway & Rubin, 1993), which emphasise the dynamic organization within the structures involved in the formation of autobiographical memories. FBMs seemed to be the outcomes of constructive processes, which apply on event-specific sensory-perceptual material, thus preserving its features of specificity and vividness. In sum, evidence so far reviewed suggests that the process of retrieving FBMs is based upon the availability of some integrated memory units, particularly vivid and detailed in their nature. However, these units are formed and maintained due to constructive and reconstructive factors, so that in the long run they are not immune to forgetting and distortions. Questioning the existence of FBMs based upon the sole consideration of reconstructive processes would undervalue evidence concerning the nature of FBM, as assessed by measurement studies so far described (see Curci & Conway, 2013, for a detailed discussion).

Conclusion The aim of this chapter has been to provide an overview of the measurement models adopted in the investigation of FBMs, as well as their implications for the theoretical consideration of the phenomenon.The reviewed literature has suggested that traditional linear models have progressively appeared as inadequate to reproduce the object of analysis. On the other hand, categorical models have offered a promising tool for modelling FBM data. Furthermore, taxometric investigation has allowed the researchers to test hypotheses concerning the peculiar characteristics of this class of autobiographical memories. The choice between dimensional and categorical

26 Curci

models, far from being a simple methodological dispute, is of fundamental relevance for the general debate about the nature of FBM (Curci & Conway, 2013). From the empirical point of view, evidence of discontinuity in the domain of autobiographical memory has provided support for the theoretical account of FBM as arising from the activation of “whole” units within the autobiographical knowledge base, which integrate sensory-perceptual vivid details and general aspects of thematic knowledge into unitary representations (Conway, 1995; Lanciano & Curci, 2012). Finally, the so-called constructivist model of autobiographical memory (Conway & Pleydell-Pearce, 2000; Conway & Rubin, 1993) has received empirical scrutiny from the research work on FBM. The organization of the structures involved in autobiographical memory is dynamic, in that reconstructive factors determine the intrinsic nature of the phenomenon, and impose some constraints on its operational measurement. The issue of assessing FBMs represents a significant illustration for this general theory. Future studies should provide more empirical support to both issues of FBM as a categorical memory formation, and the effects of reconstructive factors in determining the intrinsic nature of the phenomenon.

References Akaike, H. (1974). A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19, 716–723. Anderson, S. J., & Conway, M. A. (1997). Representations of autobiographical memories. In M. A. Conway (Ed.), Cognitive models of memory (pp. 217–246). Hove, UK: Psychology Press. Baker, F. B. (1985). The basics of item response theory. Portsmouth, NH: Heinemann. Bartholomew, D. J., Steele, F., Moustaki, I., & Galbraith, J. (2002). The analysis and interpretation of multivariate data for social scientists. Boca Raton, FL: Chapman & Hall/CRC. Bohannon, J. N. III. (1988). Flashbulb memories for the Space Shuttle disaster: A tale of two theories. Cognition, 29, 179–196. Bohannon, J. N. III, Gratz, S., & Cross,V. S. (2007). The effects of affect and input source on flashbulb memories. Applied Cognitive Psychology, 21, 1023–1036. Bohannon, J. N. III., & Symons, V. L. (1992). Flashbulb memories: Confidence, consistency, and quantity. In E. Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb memories” (pp. 65–91). New York: Cambridge University Press. Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley. Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. Carmines, E. G., & Zeller, R. A. (1979). Reliability and validity assessment. Newbury Park, CA: Sage. Carver, C. S., & Scheier, M. F. (1990). Origins and functions of positive and negative affect: A control-process view. Psychological Review, 97, 19–35. Comrey, A. L., & Lee, H. B. (1992). A first course in factor analysis. Hillsdale, NJ: Lawrence Erlbaum Associates. Conway, M. A. (1995). Flashbulb memories. Hove, UK: Lawrence Erlbaum Associates. Conway, M. A., Anderson, S. J., Larsen, S. F., Donelly, C. M., McDaniel, M. A., McClelland, A. G. R., . . . Logie, R. H. (1994).The formation of flashbulb memories. Memory & Cognition, 22, 326–343. Conway, M. A., & Bekerian, D. A. (1987). Organization in autobiographical memory. Memory & Cognition, 15, 119–132.

Measurement issues in Flashbulb memory  27

Conway, M. A., & Pleydell-Pearce, C.W. (2000).The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261–288. Conway, M. A., & Rubin, D. C. (1993). The structure of autobiographical memory. In A. F. Collins, S. E. Gathercole, M. A. Conway, & P. E. Morris (Eds.), Theories of memory (pp. 103–137). Hove, UK: Lawrence Erlbaum Associates. Conway, M. A., Singer, J. A., & Tagini, A. (2004).The self and autobiographical memory: Correspondence and coherence. Social Cognition, 22, 491–529. Curci, A. (2005). Latent variable models for the measurement of flashbulb memories: A comparative approach. Applied Cognitive Psychology, 19, 3–22. Curci, A., & Conway, M. A. (2013). Playing the flashbulb memory game: A comment on Cubelli and Della Sala. Cortex, 49, 352–355. Curci, A., & Lanciano, T. (2009). Features of autobiographical memory: Theoretical and empirical issues in the measurement of flashbulb memory. The Journal of General Psychology, 136, 129–152. Curci, A., Lanciano,T., Maddalena, C., Mastandrea, S., & Sartori, G. (2015). Flashbulb memories of the Pope’s resignation: Explicit and implicit measures across differing religious groups. Memory, 23, 529–544. Curci, A., Luminet, O., Finkenauer, C., & Gisle, L. (2001). Flashbulb memories in social groups: A comparative study of the memory of French President Mitterrand’s death in a French and a Belgian group. Memory, 9, 81–101. Curci, A., & Luminet, O. (2006). Follow-up of a cross-national comparison on flashbulb and event memory for the September 11th attacks. Memory, 14, 329–344. Dienes, Z. P. (1960). Building up mathematics. London: Hutchinson Educational Ltd. Dziak, J. J., Yang, J., Tan, X., Bray, B. C., Wagner, A. T., & Lanza, S. T. (2015). Lca distal SAS macro users’ guide (version 3.0). University Park, PA: The Methodology Center, Penn State. Embretson, S. E., & Reise, S. P. (2000). Item response theory for psychologists. Mahwah, NJ: Lawrence Erlbaum Associates. Eysenck, H. J. (1953). The structure of human personality. London: Methuen. Finkenauer, C., Luminet, O., Gisle, L., El-Ahmadi, A., van der Linden, M. & Philippot, P. (1998). Flashbulb memories and the underlying mechanism of their formation: Toward an emotional-integrative model. Memory & Cognition, 26, 516–531. Fossati, A., Beauchaine,T. P., Grazioli, F., Carretta, I., Cortinovis, F., & Maffei, C. (2005). A latent structure analysis of Diagnostic and Statistical Manual of Mental Disorders, Fourth E ­ dition, Narcissistic Personality Disorder criteria. Comprehensive Psychiatry, 46, 361–367. Gangestad, S., & Snyder, M. (1985). “To carve nature at its joints”: On the existence of discrete classes in personality. Psychological Review, 92, 317–349. Gifi, A. (1990). Nonlinear multivariate analysis. New York: John Wiley & Sons. Gleaves, D. H., Lowe, M. R., Snow, A. C., Green, B. A., & Murphy-Eberenz, K. P. (2000). Continuity and discontinuity models of bulimia nervosa: A taxometric analysis. Journal of Abnormal Psychology, 109, 56–68. Greenacre, M. J. (1993). Correspondence analysis in practice. London: Academic Press. Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the implicit association test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85, 197–216. Guyer, R., & Thompson, N. A. (2011). Item response theory parameter recovery using Xcalibre 4.1. Saint Paul, MN: Assessment Systems Corporation. Hambleton, R. K., Swaminathan, H., & Rogers, H. J. (1991). Fundamentals of item response theory. Newbury Park, CA: Sage. Hankin, B. L., Fraley, R. C., Lahey, B. B., & Waldman, I. D. (2005). Is depression best viewed as a continuum or discrete category? A taxometric analysis of childhood and adolescent depression in a population-based sample. Journal of Abnormal Psychology, 114, 96–110.

28 Curci

Higgins, E. T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94, 319–340. Hirst, W., Phelps, E. A., Buckner, R. L., Budson, A. E., Cuc, A., Gabrieli, J. D., . . .Vaidya, C. J. (2009). Long-term memory for the terrorist attack of September 11: Flashbulb memories, event memories, and the factors that influence their retention. Journal of Experimental Psychology: General, 138, 161–176. Hirst,W., Phelps, E. A., Meksin, R.,Vaidya, C. J., Johnson, M. K., Mitchell, K. J., . . . Mather, M. (2015). A ten-year follow-up of a study of memory for the attack of September 11, 2001: Flashbulb memories and memories for flashbulb events. Journal of Experimental Psychology: General, 144, 604–623. Huang, L., Dziak, J. J., Wagner, A. T., & Lanza, S. T. (2016). LCA bootstrap Stata function users’ guide (version 1.0). University Park, PA: The Methodology Center, Penn State. Kaya Kizilöz, B., & Tekcan, A. I. (2013). Canonical categories in flashbulb memories. Applied Cognitive Psychology, 27, 352–359. Kim, J. O., & Mueller, C. W. (1978). Factor analysis: Statistical methods and practical issues. Newbury Park, CA: Sage. Kvavilashvili, L., Mirani, J., Schlagman, S., & Kornbrot, D. E. (2003). Comparing flashbulb memories of September 11 and the death of Princess Diana: Effects of time delays and nationality. Applied Cognitive Psychology, 17, 1017–1031. Kvavilashvili, L., Mirani, J., Schlagman, S., Foley, K., & Kornbrot, D. E. (2009). Consistency of flashbulb memories of September 11 over long delays: Implications for consolidation and wrong time slice hypotheses. Journal of Memory and Language, 61, 556–572. Kvavilashvili, L., Mirani, J., Schlagman, S., Erskine, J. A., & Kornbrot, D. E. (2010). Effects of age on phenomenology and consistency of flashbulb memories of September 11 and a staged control event. Psychology and Aging, 25, 391–404. Kyriakopoulos, M., Stringaris, A., Manolesou, S., Radobuljac, M. D., Jacobs, B., Reichenberg, A., . . . Frangou, S. (2015). Determination of psychosis-related clinical profiles in children with autism spectrum disorders using latent class analysis. European Child & Adolescent Psychiatry, 24, 301–307. Lanciano, T., & Curci, A. (2012). Type or dimension? A taxometric investigation of flashbulb memories. Memory, 20, 177–188. Larsen, S. F. (1992). Potential flashbulb: Memories of ordinary news as the baseline. In E. Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb memories” (pp. 32–64). New York: Cambridge University Press. Lazarsfeld, P. F., & Henry, N.W. (1968). Latent structure analysis. Boston: Houghton Mill. Luce, D. R., & Suppes, P. (2002). Representational measurement theory. Stevens’ Handbook of Experimental Psychology, 4, 1–41. Luminet, O., Curci, A., Marsh, E., Wessel, I., Constantin, T., Gencoz, F., & Yogo, M. (2004). The Cognitive, emotional, and social impacts of the September, 11 attacks: Group differences in memory for the reception context and the determinants of flashbulb memory. The Journal of General Psychology, 131, 197–224. Mauro, P. M., Furr-Holden, C. D., Strain, E. C., Crum, R. M., & Mojtabai, R. (2016). Classifying substance use disorder treatment facilities with co-located mental health services: A latent class analysis approach. Drug and Alcohol Dependence, 163, 108–115. McCutcheon, A. L. (1987). Latent class analysis. Newbury Park, CA: Sage. McCutcheon, A. L. (2002). Basic concepts and procedures in single and multiple group latent class analysis. In J. A. Hagenaars & A. L. McCutcheon (Eds.), Applied latent class analysis (pp. 56–88). New York: Cambridge University Press. Meehl, P. E. (1992). Factors and taxa, traits and types. Differences of degree and differences of kind. Journal of Personality, 60, 117–174.

Measurement issues in Flashbulb memory  29

Meehl, P. E. (1999). Clarifications about taxometric method. Applied and Preventive Psychology, 8, 165–174. Meehl, P. E. (2004) What’s in a taxon? Journal of Abnormal Psychology, 113, 39–43. Meehl, P. E., & Golden, R. (1982). Taxometric methods. In P. Kendall & J. Butcher (Eds.), Handbook of Research Methods in Clinical Psychology (pp. 127–181). New York: Wiley. Meyer, D., Dimitriadou, E., Hornik, K., Weingessel, A. & Leisch, F. (2015). E1071: Misc functions of the Department of Statistics, Probability Theory Group, TU Wien. R package version 1.5–15. Retrieved from https://cran.r-project.org/web/packages/e1071/index.html Moustaki, I. (1996). A latent trait and a latent class model for mixed observed variables. British Journal of Mathematical and Social Psychology, 49, 313–334. Moustaki, I. (2001). GENLAT: A computer program for fitting a one- or two-factor latent variable model to categorical, metric and mixed observed items with missing values.Technical report, London School of Economics and Political Science, London, UK. Muraki, E. & Bock, R. D. (1998). PARSCALE: Parameter scaling of rating data (version 3.5). Chicago, IL: Scientific Software, Inc. Muthen, L. K., & Muthen, B. O. (1998). Mplus user’s guide. Los Angeles: Muthen & Muthen. Myung, I. J. (2003). Tutorial on maximum likelihood estimation. Journal of Mathematical Psychology, 47, 90–100. Nachson, I., & Zelig, A. (2003). Flashbulb and factual memories: The case of Rabin’s assassination. Applied Cognitive Psychology, 17, 519–531. Neisser, U., & Harsch, N. (1992). Phantom flashbulbs: False recollections of hearing the news about Challenger. In E.Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb memories” (pp. 9–31). New York: Cambridge University Press. Paek, I., & Han, K. T. (2013). IRTPRO 2.1 for Windows (item response theory for patientreported outcomes). Applied Psychological Measurement, 37, 242–252. Pezdek, K. (2003). Event memory and autobiographical memory for the events of September 11, 2001. Applied Cognitive Psychology, 17, 1033–1045. Pillemer, D. B. (1984). Flashbulb memories of the assassination attempt on President Reagan. Cognition, 16, 63–80. R Core Team (2014). R: A language and environment for statistical computing. R Foundation for Statistical Computing,Vienna, Austria. Retrieved from www.R-project.org/ Rasch, G. (1960). Probabilistic models for some intelligence and attainment tests. Copenhagen: Danmarks Paedogogiske Institut. Rhebergen, D., van der Steenstraten, I. M., Sunderland, M., de Graaf, R., Ten Have, M., Lamers, F., . . . Andrews, G. (2014). An examination of generalized anxiety disorder and dysthymic disorder by latent class analysis. Psychological Medicine, 44, 1701–1712. Rizopoulos, D. (2007). ltm: An R package for latent variable modelling and item response theory analyses. Journal of Statistical Software, 17, 1–25. Rost, J., & Langeheine, R. (Eds.). (1997). Applications of latent trait and latent class models in the social sciences. Münster: Waxmann. Ruscio, J., Haslam, N., & Ruscio, A. M. (2006). Introduction to the taxometric method: A practical guide. Mahwah, NJ: Lawrence Erlbaum Associates. Ruscio, J. (2007). Taxometric analysis: An empirically grounded approach to implementing the method. Criminal Justice and Behavior, 24, 1588–1622. Ruscio, J. (2016). Taxometric programs for the R computing environment: User’s manual. Retrieved from http://ruscio.pages.tcnj.edu/quantitative-methods-program-code/ Sartori, G., Agosta, S., Zogmaister, C., Ferrara, S. D., & Castiello, U. (2008). How to accurately detect autobiographical events. Psychological Science, 19, 772–780. Schank, R. C. (1982). Dynamic memory: A theory of reminding and learning in computers and people. New York: Cambridge University Press.

30 Curci

Schmidt, N. B., Kotov, R., & Joiner, T. E. (2004). Taxometrics:Toward a new diagnostic scheme for psychopathology. Washington, DC: American Psychological Association. Schwarz, G. (1978). Estimating the dimension of a model. Annals of Statistics, 6, 461–464. Shapiro, L. (2006). Remembering September 11th: The role of retention interval and rehearsal on flashbulb and event memory. Memory, 14, 129–147. Smith, M. C., Bibi, U., & Sheard, D. E. (2003). Evidence for the differential impact of time and emotion on personal and event memories for September 11, 2001. Applied Cognitive Psychology, 17, 1047–1055. Talarico, J. M., & Rubin, D. C. (2007). Flashbulb memories are special after all; in phenomenology, not accuracy. Applied Cognitive Psychology, 21, 557–578. Tekcan, A., Ece, B., Gülgöz, S., & Er, N. (2003). Autobiographical and event memory for 9/11: Changes across one year. Applied Cognitive Psychology, 17, 1057–1066. Thissen, D. (1991). MULTILOG User’s Guide: Multiple, categorical item analysis and test scoring using item response theory. Chicago, IL: Scientific Software, Inc. van de Geer, J. P. (1993). Multivariate analysis of categorical data: Applications. Newbury Park, CA: Sage. Van der Pol, F., Langeheine, R., & de Jong, W. (1998). PANMARK 3 user manual. Voorburg, The Netherlands: Netherlands Central Bureau of Statistics. Vermunt, J. (1997). Log-linear and event history analysis with missing data (version 1.0). Tilburg, The Netherlands: Tilburg University. Vermunt, J. & Magidson, J. (2005). Latent GOLD 4.0 user’s guide. Belmont, MA: Statistical Innovations Inc. Waller, N. G., & Meehl, P. E. (1998). Multivariate taxometric procedures: Distinguishing types from continua. Thousand Oaks, CA: Sage Publications. Waller, N. G., Putnam, F. W., & Carlson, E. B. (1996). Types of dissociation and dissociative types: A taxometric analysis of dissociative experiences. Psychological Methods, 1, 300–321. Waller, N. G., & Ross, C. A. (1997). The prevalence and biometric structure of pathological dissociation in the general population:Taxometric and behaviour genetic findings. Journal of Abnormal Psychology, 106, 499–510. Williams, D. M., & Hollan, J. D. (1981). The process of retrieval from very long-term memory. Cognitive Science, 5, 87–119. Williamson, D. A., Womble, L. G., Smeets, M. A. M., Netemeyer, R. G., Thew, J. M., Kutlesic, V., & Gleaves, D. H. (2002). Latent structure of eating disorder symptoms: A factor analytic and taxometric investigation. American Journal of Psychiatry, 159, 412–418. Winograd, E., & Killinger, W. A. (1983). Relating age at encoding in early childhood to adult recall: Development of flashbulb memories. Journal of Experimental Psychology, 112, 413–422. Wright, D. B. (1993). Recall of the Hillsborough disaster over time: Systematic biases of ‘flashbulb’ memories. Applied Cognitive Psychology, 7, 129–138. Wright, D. B., Gaskell, G. D., & O’Muircheartaigh, C. (1998). Flashbulb memory assumptions: Using national surveys to explore cognitive phenomena. British Journal of Psychology, 89, 103–121. Zimowski, M. F., Muraki, E., Mislevy, R. J., & Bock, R. D. (1996). BILOG-MG. Multiplegroup IRT analysis and test maintenance for binary items. Chicago, IL: Scientific Software International.

2 FLASHBULB MEMORY METHODS Daniel B. Wright and Marianna E. Carlucci

New and old methods for investigating Flashbulb memories (FBMs) are explored. We discuss what distinguishes FBMs from other events, how the measurement strategies should relate to the underlying constructs of FBMs, and detail of methods used in FBM research. We focus on the relationship between the goals of the research and whether the researchers sample a lot of people or sample a lot of events. Recent research, particularly studies using experimental designs, provide innovative ways to understand the underlying mechanisms and to predict the consequences of FBMs.

The observation that some events are more memorable than others, and that some of the events on the high side of the memorability distribution tend to be important news events, are the empirical basis of the Flashbulb memory (FBM) phenomenon. There are many interesting scientific questions that can be asked about FBMs, but these questions must go beyond this initial observation. Indeed, one methodological problem in FBM research is regarding the types of questions asked and the methods used to answer those questions. Many, though not all, of the scientific questions of interest to researchers about FBMs are causal, while the methods used in the prototypical FBM study are more appropriate for addressing associative questions. Furthermore, the methods that tend to be used are well suited for differentiating among people and groups, which is arguably of more interest to sociologists and oral historians than to psychologists, who are usually more interested in differences between situations and tasks (Wright & Gaskell, 1995). The purpose of this chapter is to explore the methods used in FBM research and to introduce new ways to explore FBMs. The two basic requirements for the scientific process are measurement and theory. You must have measurement and some measurement theory (Hand, 2004) and explore a set of hypotheses that are part of a complex network of hypotheses

32  Wright and Carlucci

 Emotion  high Now Print → Good memory Importance →  low   Survival  → Normal processes → Less good memory FIGURE 2.1 A simplified

FBM model, based on Brown and Kulik (1977) and subsequent models.

(Wright, 2006). This chapter is divided into two sections on the basis of these two requirements.The first section examines what FBMs are and what they are not, and how researchers can decide whether a memory is an FBM. The chapter also delves into various choices researchers have to make in deciding what their domain of study is and some of the methods that could be used. The second section focuses on the main scientific questions that can be asked once the conceptual issues of the first section have been addressed. Different FBM models are not discussed here because there are excellent reviews in other parts of this book (for example Luminet, Chapter 3, this volume) and it is our belief that some of the methodological issues that we address should precede a detailed discussion of cognitive theory, though introducing a simple model will aide discussion. Consider a simplified schematic of the basic FBM model (Figure 2.1). Most FBM researchers describe how some combination of emotion, importance, and survival blend so that events that score highly on these attributes evoke some special mechanism that produces good memories. How these event characteristics combine is where the detailed discussions of FBM theorizing lie.

Conceptual and measurement questions about Flashbulb memories What is the population under investigation? Figure 2.2 shows a distribution of memorability for a set of events. For the purposes of this chapter, assume that “memorability” is a variable about the quality of memories; if the value is above this threshold then the memory is an FBM, and if the value is below this threshold then the memory is not an FBM. Special mechanism advocates (those who support the idea that FBMs are created through a special cognitive mechanism) might argue that the distribution for memorability should be bi-modal (a distribution with two peaks), a mixture of a distribution for “normal” memories and a distribution for FBMs. People have used different ways to conceptualize this and to measure it, and these are discussed below. As with any distribution, before asking any empirical questions it is necessary to ask, “What is the population of events that form this distribution?” Is it all events experienced (and what is an event?)? All news events? All events that people recall vividly? All assassinations? Similarly, is the distribution for all people, for some selected sub-population, or just

Flashbulb memory methods  33

Non-FBs FBs Memorability FIGURE 2.2 A distribution, for

some population of events, of memorability with events beyond some value being called FBMs.

for an individual? For a detailed discussion of event-memory please see Rubin and Umanath (2015). This question was pondered in Brown and Kulik’s (1977) seminal study on FBMs. While they mostly asked their sample about news events, they also asked people to report about personal events. Thus, they were interested in all events that might be recalled vividly. Following Brown and Kulik’s study, some researchers (e.g. McCloskey,Wible, & Cohen, 1988; Neisser, 1982) questioned their assumption that FBMs were accurate (how to differentiate between the good and less good memories from Figure 2.1), while others (e.g. Conway et al., 1994; Finkenauer et al., 1998) went on to develop Brown and Kulik’s FBM model: how the emotion, importance, and survival qualities from Figure 2.1 combine into a variable denoted as “memorability”, which makes FBMs different from other memories. Unfortunately, Brown and Kulik presented little data in their original study to support their model, and it is unclear how much further insight subsequent research has given to the concept.1 While memory accuracy and models of memory are clearly important, making statements about memories for events in some part of Figure 2.2 without considering the population being examined makes inference difficult. At this point, it is worth thinking about some of the possible populations of events. This is a difficult task, since there are no agreed definitions of what is an event or what is a memory.The only aspect that seems agreed by FBM researchers is that FBMs involve recollection of the context when one learns of an event (Hirst &

34  Wright and Carlucci

Phelps, 2016) rather than just experiences of the actual event. There are questions about how microscopic or extended an event should be, and what relationship the person needs to have with the event. These are general questions for memory research (Rubin, 1992). One possibility is considering all events that somebody with a typical adult’s cognitive abilities would be able describe soon after the event. This set of events is what many autobiographical memory researchers might consider their domain of interest. Another possibility is all events that can, given the right prompt, be recalled at testing.2 One of the differences between these two approaches is that many of the events that occurred will not be remembered, so the former would include events low in memorability (e.g. below some threshold for recall) and the latter would not. These can be considered as two populations (events that are not easily recalled vs. events that are more easily recalled).The first is the population defined above and often assumed, though not always explicitly, for much autobiographical memory research: all events for which a typical adult could have given a reasonable account at the time of the event.This allows the population under investigation to include events that some people would forget. The second population is one where the researchers are interested in a single event and more in line with seminal FBM research. Indeed, using single events can be helpful in understanding many aspects relevant to FBMs. Consider the vivid memories astronomers at the Anglo-Australian observatory had at the discovery of Supernova 1987-A (Wright & Gaskell, 1995). While not many people outside of the observatory remember this event, this research illustrated that strong personal feelings can be associated with FBM memories for an event of relatively little societal importance. Indeed, the population does not have to be large (e.g. a whole country) as people can have FBMs of a family event (Rubin & Kozin, 1984). To summarize, when people embark on FBM research it is important to consider what population of events to include as this decision will have implications on the resulting methodological choices.

What is memorability and how is it measured? Memorability is perhaps the hallmark of FBMs and is said to distinguish FBMs from other memories. Brown and Kulik (1977) and subsequent FBM researchers suggest that the prototypical FBM is an extremely long-lasting and vivid recollection. This chapter describes three popular ways of conceptualizing memorability (see also Curci, Chapter 1, this volume):Tulving’s (1985) remember/know memory distinction, clarity ratings, and latent variable analysis. Tulving (1985) differentiated two types of memories. He described “remember” memories as those that allow the person to mentally re-live the event and that evoke autonoetic consciousness, a way to place oneself in the memory (similar to what one might expect an FBM memory to be). In comparison, “know” memories are those that lack the re-living aspect of remember memories (similar to what one might expect an everyday memory to be).This distinction is used in many memory studies (see also Gardiner, 1988) where participants are shown, for example, a list of words and later asked if they have seen these words before. If they say “yes”, they are asked to write R if it is a remember memory and K if it is a know memory. Within

Flashbulb memory methods  35

autobiographical memory research there has been a move away from having participants choose between just a K (for Know) and an R (for Remember) response to choosing among K, R, and F or G (for Familiar or Guess) responses, because researchers felt that some participants were using K responses inappropriately (Gardiner et al., 2002). These different categories show how complex it is to capture or measure different types of memories. Notably, the re-living, self-referential description of remember memories suggests these might be similar to what researchers mean when they describe FBMs. The problem for FBM research is that in practice remember (R) responses do not capture what people usually think of as a true FBM. In a typical recognition memory study people may be shown 50 words and later asked if their memory is a K or an R memory. Many respond R, but it is questionable whether these memories are FBMs. Thus, Tulving’s taxonomy may not capture or measure the true essence of an FBM. FBMs are often described as vivid, live-like, clear memories. As such, one way that researchers have measured whether a memory is an FBM is to ask participants in FBM studies how vivid or clear their memories are for a particular event. While this is a good way of describing one’s memory there are some methodological issues to consider. First, this approach assumes that the clarity scale is unidimensional.That is, that the scale is measuring one discrete construct that can be captured with one number. Second, this approach hinges on the assumption that people’s individual interpretations of the clarity scale are similar. Finally, measuring a construct (i.e. clarity) with a single item or question may be unreliable. If someone asked you a single question about your current mood (e.g. “how good of a mood are you in?”) would that be enough to capture your current state? The same issues apply to capturing FBMs with a single item or question. Furthermore, while researchers have traditionally thought of recollection as univariate, there is evidence that shows that this view is “incomplete” (p. 589) and that a more nuanced view of recollection is needed (Brainerd, Gomes, & Moran, 2014). Indeed, some researchers have argued that recollection is not univariate. Memory has several dimensions that cannot be captured by a single item or question. Finally, many researchers believe that “clarity” is not the only aspect that should contribute to the flashbulb-ness dimension, but that consistency, details of memory, etc., should also be used in the measurement. Interestingly, FBMs tend to decrease in consistency and accuracy over time so the utility of these measures is questionable for defining FBMs. The third approach circumvents the problem of having only a single question. Brown and Kulik (1977) suggest that one of the defining characteristics of an FBM is that people can recall the four Ws: what they were doing, how they heard, whom they were with, and where they were. Responses to these questions can then be aggregated to define overall memorability. Three ways of aggregating the data are considered here: counting the number of responses, conducting a latent trait model, and conducting a latent class model. Moustaki (1996) provides detailed comparisons among these three methods in the context of FBMs, so her paper is particularly valuable for FBM researchers (see also Curci, Chapter 1, this volume). Here these methods are briefly described.

36  Wright and Carlucci

The “counting the number of responses” approach involves either counting the number of times someone positively responds to each of four Ws or counting them and allowing some partial credit. The latent trait and latent class models deserve further explanation. Bartholomew and Knott (1999) describe how several different latent variable models all have a similar form, and they also use FBM data to aid their explanation. Latent trait models take several categorical observed variables and try to account for covariation in responses by assuming that there are a small number (often one) of continuous latent variables underlying the responses. They are popular in education, where they are often called item response models (or item response theory), and are analogous to exploratory factor analysis when the observed variables are categorical. This procedure would create a dimension for memorability. Latent class models account for the responses by assuming that there are groups of people, and that people within the same group have similar probabilities for responding in certain ways, but there are large differences between people of different groups. These are popular in sociology and are becoming more popular in psychology. As discussed in the next section, it is sometimes difficult to differentiate these approaches on an empirical basis.

Are there differences between these approaches? What memorability is and how to measure it are important and interrelated questions. First, consider whether to use participants’ responses on a clarity scale or the number of positive responses to the what, how, where, and who questions. In Wright, Gaskell, and O’Muircheartaigh (1998, Exp. 2) a large number of participants from a representative sample of the UK adult population were asked about Margaret Thatcher’s resignation (1990). They were asked how clear their memory was of the event on a 1 (Cannot remember the event) to 5 (Completely clear) rating scale, and also asked for their responses on the four W-questions. One would expect that people who could recall all four Ws would give thwemselves very high ­ratings on the clarity scale. However, those responding positive to all the W-questions tended to respond only in the middle of the clarity scale. Middle of the scale clarity responses are not commensurate with the live, photographic re-living descriptions often given to FBMs. Thus, the choice of using responses to the Ws or rating scales can provide very different interpretations for what FBMs are and produce very different estimates for the number of people with FBMs for any given event. The second comparison is between latent trait and latent class analysis. Moustaki (1996; see also Curci, 2005, and Chapter 1 of this volume) goes through the Thatcher resignation data in detail. She shows how the data can be used in these models to produce either a single latent trait along which people vary (continuous) or two latent classes to which people can belong (categorical). Bartholomew and Knott (1999; for an introductory textbook on these procedures, see Bartholomew et al., 2008) discuss how it is very difficult to differentiate these models empirically and that the choice often depends on the particular theory the researcher has for FBMs. For example, there is much discussion about whether memory retrieval is based on a threshold model, where people either remember the event or not, vs. whether there is some continuous memory strength dimension that predicts the probability

Flashbulb memory methods  37

of remembering the event. If you assume that people either have or do not have an FBM, as if some trigger either fires or does not fire, then a latent class model is likely more appropriate (categorical). However, if you assume that FBMs are simply at one end of a memorability scale, then the latent trait model is likely more appropriate (continuous).Brown and Kulik’s original description of FBMs, where they adopt Livingston’s (1967) Now Print! mechanism, suggests a threshold model that is more consistent with the latent class models. Under this model it is suggested that FBMs are created to preserve memories of important events or events with personal relevance through a special mechanism in the brain. Thus, the class vs. trait question could be of great importance to the special mechanism debate if one believes the mechanism is like Brown and Kulik’s description. However, other types of special mechanisms could be more continuous in nature, and therefore the choice of class vs. trait does not strictly depend on the special mechanism debate.The choice of class vs. trait still depends on the theoretical arguments being made. There is also a statistical difficulty because it is often not possible to distinguish between class and trait models. Bartholomew et al. (2008) point out that often the covariance matrix of observed variables can be equally well recreated by either assuming a set of discrete latent classes or by assuming continuous latent traits. This means that either model will fit equally well. However, Curci (2005) found that FBM is estimated best by latent class models. Additionally, Curci and Lanciano (2009) compared latent class and latent trait model approaches by applying them to FBM and to event memory. They found that data showed a better fit for latent class models to FBM data than to event memory data.When the latent classes are very different from each other (and within the classes they are relatively homogeneous), then it is possible to differentiate classes and traits. The taxometric approach (Ruscio, Haslam, & Ruscio, 2006; Waller & Meehl, 1998; see also De Boeck, Wilson, & Acton, 2005) has become a popular method for doing this in clinical psychology and it could also be used in FBM research. For example, Lanciano and Curci (2012) used the taxometric approach and found evidence to support the categorical hypothesis for FBM.

Summary on measurement Science requires several decisions about measurement. To help FBM researchers go beyond the observation that some people have good memories (which have various correlated attributes) for some events, it is necessary to be explicit about: 1 2 3

What the population under investigation is (all events vs. single events). What distinguishes FBMs from other events (memorability). Measurement strategies regarding the underlying construct.

Types of FBM hypotheses The following is a typical explanation of research methods in psychology. There is some population, P, about which the researchers want to make inference. It is

38  Wright and Carlucci

usually impossible to study this entire population and therefore a sample, S, is taken from it. Making inference about a population from data of a sample has a long tradition. For example, on the first page of Gossett’s (1908) classic t test paper, any result “is only of value in so far as it enables us to form a judgment as to the statistical constants of the population” (p.398) – but this requires certain rules. Depending on whether the researchers assume S is representative of P, there are two types of inference that can be made from data of S about P. If S is assumed to be representative of P, then provided that sampling and measurement error are taken into account, the researchers can make positive statements about all of P. For example, suppose you have a sample of memories that you think are representative of the population of interest, and find the correlation between the emotional reaction at the time of the event and the clarity of the subsequent memory has a 95% confidence interval of 0.3 to 0.4. This would allow you to conclude, with some confidence (although what exactly “confidence” means is another thorny issue), that the correlation in the population is likely within the calculated band. Suppose instead that you are not willing to assume that S is representative of P, either because the people chosen are not representative of the population of people or the memories chosen are not representative of the population of these people’s memories. Instead you assume the sample is just representative of some (often unknown) subset of the population. Finding an interval of 0.3 to 0.4 only allows you to say that you are confident that the correlation is in this range for this particular subset.This is called local inference (Lunneborg, 2000). If someone has put forward the hypothesis that “for all subsets of the population the correlation is 0.5”, then the finding that this does not hold for any particular subset, including the observed subset, allows this “for all subsets” hypothesis to be rejected. Thus, the hypothesis that the motions of all bodies conform to Newton’s laws can be rejected by precise observation of the orbit of the planet Mercury. This is why Popperian falsification is used in many sciences. However, while “for all subsets” hypotheses are popular in some of the sciences, it is difficult to argue for their applicability with reference to FBMs. It is likely that people would expect correlations to vary in different subsets of a population, as is usually true in the medical and social sciences (Engels et al., 2000). Both of these are associative hypotheses, in that they are about the population joint distribution (often measured by the correlation) of emotion and clarity for either the entire population or for some subset of the population. They are not about causality, although as discussed below they can be used to inform causal theories. There are questions about the role of associative hypotheses in science (Cronbach, 1957; Fodor, 1991; Spearman, 1904; Wright, 2006), but it is clear that, particularly within FBM research, associate hypotheses are often discussed. The sample, S, is often divided into groups. Suppose there are two groups, S1 and S2. This division could be based on something outside the control of the researcher, such as whether the participants were watching TV when some FBM event occurred, or the participants’ gender. In these cases it is a “quasi-experiment” and associative inferences are usually made (i.e. emotion and clarity are associated,

Flashbulb memory methods  39

rather than emotion causes clarity). In some cases, researchers might use the associative finding (TV is associated with a certain types of memories) to infer some causal relationship (the graphic TV images cause certain types of memories), but this requires further assumptions (such as that those watching the television are similar to those not watching the television). In other cases, making causal inference from an association is not possible (for example, Holland, 1986, argues gender cannot cause anything). If the division into S1 and S2 is random, as often occurs with experiments, the researchers are able to make causal inferences more easily, albeit with some caveats (Cook & Campbell, 1979). The critical aspect of drawing causal inferences is that the two groups should differ only by sampling error (which can be estimated if random assignment is used) and whatever the researchers want to make the causal inference about. Philosophers call this the ceteris paribus conditional, and as Cook and Campbell (1979) put it: “random assignment is the great ceteris paribus – that is, other things being equal – of causal inference.” (p. 5). It is not possible to randomly allocate somebody into a condition where they think, for example, that the death of Muhammad Ali (2016) was an important event. Thus, this is a prime reason that most FBM studies cannot assess causality. Most FBM studies compare naturally occurring groups, which may differ in many ways other than just what they thought about the event. Much sociology and social psychology aims to document which groups of people are most affected by particular events. To make causal inference about, for example, thinking Ali’s death was an important event requires some assumptions. While it would clearly be wrong for researchers to assume that, for example, people who thought Ali’s death was a very important event were the same as others in all other ways, a researcher might feel that it is plausible to assume that any of the other differences are not associated with the dependent variable, memory quality.This assumption is of course testable if enough of the other possible confounding variables have been accurately measured. Purists might argue that because most FBM data arise from quasi-experimental research they cannot unequivocally lead to causal conclusions. Causal statements require certain assumptions to be met (e.g. randomization) while most FBM studies do not meet those assumptions. Strictly speaking, finding a correlation between, say, emotion and clarity does not imply that emotion causes clarity, but finding a correlation does show that there almost certainly exists some causal relationship between various parts of the complex network of hypotheses within which these two attributes are embedded (see Meehl & Waller, 2002; Pearl, 2009; Wright, 2006). If this network is viewed spatially, where the number of intervening nodes is a measure of distance, then because, all other things being equal, simpler models should be preferred to more complex models (i.e. Occam’s razor, the bet on sparsity principle), it is best to consider causal relationships between the nearest events as more likely than more distant relationships. Further, because some events are separated in time (e.g. emotion at the time of an event and clarity of a subsequent memory) and, outside of some physics hypotheses, causes precede effects; this can also help in locating possible causal relationships.

40  Wright and Carlucci

FBM research is usually correlational/quasi-experimental. The data are about associations. While specific causal statements can be made on the basis of correlational data, there are often other plausible explanations, including that some other variables cause variation in both attributes (Simon, 1954). For example, while most FBM/autobiographical memory research shows that emotion and clarity are positively correlated, it appears this may be due to the event’s importance being causally related to both (Wright & Nunn, 2000). However, the causal relationship may be in the opposite direction. This is supported by numerous laboratory studies which indicate that heightened emotion negatively affects memory clarity (­Deffenbacher et al., 2004). This is an example where the associative hypothesis (that emotion and clarity are positively related) may be in the opposite direction from the causal hypothesis (that emotion impairs memory) as evidenced by contradicting results. While causal inference from correlational data is possible, it has numerous pitfalls, it is open to alternative interpretations, and requires caveats and assumptions (Pearl, 2009). This does not mean that the data from a typical FBM study are not valuable, much of them are, but it does question whether alternative designs could address many of the causal questions in better ways. Moving forward, it may also be necessary for FBM researchers to devise new ways to study FBMs. For example, creating new ways of communicating about FBMs and new mathematical approaches to the study of FBMs.

The single-case study approach Using single events is a popular way of assessing FBM. By using the same event, some control is provided for the study. However, it does mean that the researcher needs to be cautious about generalizing to other events. The difficulty with the single-case study approach with FBM research is that differences in the events are often the key variables under investigation. Therefore, controlling the event is counter-productive or even impossible. If a researcher is interested in making inferences about all events, then studying only, for example, the Hillsborough football stadium disaster limits what can be said about all events. Wright (1993) found systematic differences in people’s memories of this disaster across time. He could not conclude anything in particular about events, even those of FBM calibre, he could only conclude that one particular event, which was arguably of FBM calibre for many of his participants, yielded memories incompatible with the hypothesis that for all subsets of events of flashbulb calibre, there are no systematic biases in memories. This research does say something about the Hillsborough event. As such, it is of interest to people who actually want to know about the event. This is different from the aims of most cognitive psychologists, but in line with researchers from other disciplines. Some of the best singlecase studies relevant to FBMs are by anthropologists studying memories for some of the bizarre and often horrific rituals in other cultures (Whitehouse, 2000).There have also been valuable studies of particular political and cultural events (Curci & Lanciano, 2009; Lanciano & Curci, 2012).

Flashbulb memory methods  41

In a very interesting single-case study relevant to FBMs, Judith Zur (1998) examined La Violencia, the violence surrounding the ethnic war in Guatemala. The authorities forbid discussion of the atrocities: “The entire history of la violencia can be read as a war against memory, an Orwellian falsification of memory, a falsification of reality” (Zur, 1998, p. 159). Zur’s research provides great insight into this often forgotten war, but it also provides insight regarding the workings of memory in a situation that would be immoral to mimic in a laboratory. She provides graphic illustrations that memories, some accurate and some distorted, prevail even when people are not allowed to talk about the kidnappings and murders. Zur’s work provides an example, outside cognitive psychology, of using the single-case approach applied to memory. As the single-case approach is relatively rare within cognitive psychology (outside of neuropsychology, where the individual is often the object), the question is why it became the method of choice for FBM researchers. Brown and Kulik’s (1977) original study used multiple events and this allowed them to make valuable comparison across events. They used part of these data as evidence for their model. Although, as discussed in note 1, some of their data do not support this model, their design was a good approach. Most cognitive psychologists are interested in how different events elicit different types of memories.Therefore, it is surprising that the FBM literature has tended to use single events, or a small number of events, rather than asking about several events that differ by the attributes of interest. There may be two reasons why cognitive psychologists have opted for the single-case approach when researching FBMs. The first is the desire to examine consistency or accuracy. In order to do this, researchers have to question participants soon after an event and then again weeks, months, or even years later (this is called the test-retest approach). This is only practical for a single event if the relevant events are defined as important news events. The second reason is that since Neisser’s (1982) and McCloskey et al.’s (1988) single-case approach, this method has gained acceptance beyond reproach as the way to study the phenomenon in general. Regardless, the growing awareness of the FBM concept has made the study of FBMs acceptable in its own right. Figure 2.3 shows that citations of Brown and Kulik and use of the phrase “Flashbulb memory” have increased dramatically in recent years. The study of FBMs has become its own area of research but, as mentioned in the first half of this paper, more attention is needed in several areas. For example, it is not clear what FBMs are, how they manifest, how to measure them, and how to extrapolate the information from one event to a larger pool of events. Thus, more research is needed.

Summary on hypotheses There are two broad types of hypotheses: causal and associative. The design of the study should be congruent with the type of hypothesis of most importance because it is difficult to make inference of one type from a study designed to investigate the other type. Wright (2006) describes the differences between these types of hypotheses and their associated methods with reference to eyewitness testimony

42  Wright and Carlucci

+ Brown and Kulik (1977) cites FB or FBs cites

30

Number of citations

+ +

25

+

20

+

15

+

10

+ ++

5 0

+

++

++

+ ++ +

1980

+

+

+

+ +

+

+ + +

+

+

+

1985

1990

1995

2000

2005

Year FIGURE 2.3 The

number of times Brown and Kulik (1977) was cited each year since publication (shown with a +) and the number of times the phrase “Flashbulb memory” or “Flashbulb memories” was used in titles and abstracts (shown with a ●). The data are from the ISI Web of Knowledge collected on 13 March, 2008 and go up to 2007.

research, and many of the same issues apply here. FBM researchers also have to decide whether they are interested in differences between people or between events. For example, Wright et al. (1998) were explicit that they were interested in group differences. Therefore, they asked a large representative sample of UK adults about a couple of events so that they could make positive statements about the associations for these events. If the interest is in group differences, then this is a viable approach. If the interest is in how somebody would remember different types of events, then the better approach is to ask a relatively small number of people about a large number of events (Wright & Gaskell, 1995), as is often done in other areas of autobiographical memory research (Wright & Nunn, 2000). Nevertheless, the single-case approach is seemingly preferred amongst FBM researchers. Indeed, in doing research for the update of this chapter, the authors found something striking. Researching FBM in the past five years yielded studies revolving around major events like Osama Bin Laden’s death (2011), Michael Jackson’s death (2009), Barack Obama’s inauguration (2009), and follow-up studies

Flashbulb memory methods  43

on 9/11 FBMs (Demiray & Freund, 2015; Hirst et al., 2015; Koppel, Brown, Stone, Coman, & Hirst, 2013).The single-case approach remains the dominant method of inquiry into FBMs.

New directions Though single-case studies are ubiquitous in the FBM literature, they lack the experimental control needed to make causal statements. As a result, researchers have attempted the experimental approach in FBM research (see Curci, 2005, Study 2 for one of the first known experimental studies of FBM). The experimental approach deepens the conversation regarding the measurement of FBMs (the focus of this chapter) and signals new ways in which researchers are thinking about FBMs. In a novel paradigm, MacKay et al. (2004) created an experiment to assess memory for emotional events. In these studies, participants were asked to complete a memory task that included taboo words. In this modified Stroop task, participants were asked to recall the color (context) of each word. The researchers found that, like with FBMs, participants in these studies remembered taboo words better, with higher confidence ratings, and faster (via response times) when compared to their memories for neutral words. Notably, none of these effects were due to other characteristics of FBMs, such as rehearsal. In a 2005 study, MacKay and Ahmetzanov, using the same paradigm, created an FBM analogue in the laboratory by including contextual features such as location. Results showed better location memory for taboo words compared to neutral words. Similarly, Lanciano, Curci, and Semin (2010) used an experimental design to investigate the influence of emotional factors on FBMs (see also Lanciano, Curci, & Sartori, Chapter 6, this volume). They manipulated emotional valence by asking participants to either visualize the death of a parent (and the consequences that death would put into motion) or visualize a benign task (shopping at the supermarket). Both groups were asked to study a set of words. The recall of those words was the main dependent variable. The results indicate that those in the emotional condition (those who pictured the death of a parent) remembered more words than those in the neutral condition. While ecological validity is a limitation of this study, the results showed that a key component of FBMs (emotional valence) does causally impact memory. Understanding the underlying mechanisms of FBMs (e.g. how emotional valence influences memory) will undoubtedly lead to more advanced measurement of FBMs. This volume makes clear that FBMs are special, at least insofar as they have attracted a lot of research attention. One reason for this may be that we all have an FBM. Indeed, “each of us carries in our memory accounts of those events that make up the recent history of our society” (Schuman & Converse, 1984, p. 38).This communal understanding of what it felt like, for example, to see the horrific images on 9/11, the Boxing Day Tsunami, and other events draws people together into community. It allows people to place their personal life onto a historical time-line. It creates a collective picture book that we all share and know. However, we also know that many FBMs do not appear remarkable in many ways. They are prone to

44  Wright and Carlucci

decay, tend to be inconsistent (Curci et al., 2001), and are subject to errors (Pezdek, 2003) in much the same way that non-FBMs are.Thus, it may be that FBM research is ripe for a new approach to its study and measurement.

Conclusion As a postgraduate at the London School of Economics (LSE), one of us (DBW) was researching what had become known as everyday memories. He would go to the library and dutifully photocopy lots of papers, some of which he would read. So that he could read without interruption, he would go off to other departments’ seminar rooms. One day he was reading Banaji and Crowder’s (1989) attack on everyday memory research in a philosophy seminar room under a picture of Imre Lakatos. Dan was angry. How dare somebody attack the sacrosanct topic on which he was conducting his PhD research! The irony of his physical location only became apparent after counting to ten and in subsequent reads of the article.While at the LSE, Lakatos (1977) had argued that research programs have a set of core beliefs that go largely unchallenged. Programs can be either degenerative or progressive. In the case of the everyday memory movement, Banaji and Crowder (1989) were arguing that the everyday memory movement was degenerative: “no theories that have unprecedented explanatory power have been produced; no new principles of memory have been discovered, and no methods of data collection have been developed that add sophistication or precision” (p. 1185). Dan now believes the Banaji and Crowder (1989) paper was very positive for the study of everyday memories. It forces people to step back and ask questions about the methods they use, and they point people toward ways of making everyday memory research generalizable. Just because people use a particular type of stimulus does not excuse them from asking if there are better methods. It is important to make sure that FBM research avoids becoming degenerative. As Figure 2.3 shows, FBM research is an increasingly popular topic, but this would be problematic if its popularity protected researchers from exploring the core of what it is to be an FBM. It is necessary to develop a theory of measurement that allows FBMs to be studied in relation to other memories and to use methods that are appropriate to the research questions.The current research has shown that there exist some events that produce good memories and some events that produce bad memories.

Notes 1 This statement deserves further explanation. Brown and Kulik’s (1977) main group comparisons, which they use to justify their model that emotion combined with consequentiality produces FBMs, were that black people had more FBMs than white people for civil rights events, and these events were more consequential events for black people than for white people. However, in their study black people also rated both Robert Kennedy’s assassination and Ted Kennedy’s Chappaquiddick incident as more consequential than white people did, but black people had fewer FBMs for these events. Similarly, black ­people rated Martin Luther King’s assassination as more consequential than John F. K ­ ennedy’s, but had fewer FBMs for King’s assassination. See Wright et al. (1998) for further examples of this

Flashbulb memory methods  45

type. These observations do not falsify the Brown and Kulik model, but this difficulty in falsifying a causal model with correlation data is a theme of this chapter and one of the reasons why the amount written about FBMs exceeds the amount of scientific progress in understanding them. 2 There are lots of other possibilities. For example, the population in the typical memory recognition study is all the events that participants are given the opportunity to recognize.

References Banaji, M. R., & Crowder, R. G. (1989). The bankruptcy of everyday memory. American Psychologist, 44, 1185–1193. Bartholomew, D. J., & Knott, M. (1999). Latent variable models and factor analysis (2nd ed.). London: Arnold. Bartholomew, D. J., Steele. F., Moustaki, I., & Galbraith, I. (2008). Analysis of multivariate social science data (2nd ed.). London: Chapman & Hall/CRC. Brainerd, C. J., Gomes, C. F. A., & Moran, R. (2014). The two recollections. Psychological Review, 121, 563–599. Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. Conway, M. A., Anderson, S. J., Larsen, S. F., Donnelly, C. M., McDaniel, M. A., McClelland, A. G. R., . . . Logie, R.H. (1994). The formation of flashbulb memories. Memory & Cognition, 22, 326–343. Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: Design & analysis issues for field settings. Boston: Houghton Mifflin Company. Cronbach, L. J. (1957). The two disciplines of scientific psychology. American Psychologist, 12, 671–684. Curci, A., Luminet, O., Finkenauer, C., & Gisle, L. (2001). Flashbulb memories in social groups: A comparative study of the memory of French President Mitterrand’s death in a French and a Belgian group. Memory, 9, 81–101. Curci, A. (2005). Latent variable models for the measurement of flashbulb memories: A comparative approach. Applied Cognitive Psychology, 19, 3–22. Curci, A., & Lanciano, T. (2009). Features of autobiographical memory: Theoretical and empirical issues in the measurement of flashbulb memory. The Journal of General Psychology, 136, 129–152. De Boeck, P., Wilson, M., & Acton, G. S. (2005). A conceptual and psychometric framework for distinguishing categories and dimensions. Psychological Review, 112, 129–158. Deffenbacher, K. A., Bornstein, B. H., Penrod, S. D., & McGorty, E. K. (2004). A  meta-analytic review of the effects of high stress on eyewitness memory. Law & Human Behavior, 28, 687–706. Demiray, B., & Freund, A. M. (2015). Michael Jackson, Bin Laden and I: Functions of positive and negative, public and private flashbulb memories. Memory, 23, 487–506. doi:10.1080/ 09658211.2014.907428 Engels, E. A., Schmidt, C. H., Terrin, N., Olkin, I., & Lau, J. (2000). Heterogeneity and statistical significance in meta-analysis: An empirical study of 125 meta­analyses. Statistics in Medicine, 19, 1707–1728. Finkenauer, C., Luminet, O., Gisle, L., El-Ahmadi, A., van der Linden, M., & Philippot, P. (1998). Flash bulb memories and the underlying mechanisms of their formation: Toward an emotional-integrative model. Memory & Cognition, 26, 516–531. Fodor, J. A. (1991). You can fool some of the people all of the time, everything else being equal; hedged laws and psychological explanations. Mind, 397, 19–34.

46  Wright and Carlucci

Gardiner, J. M. (1988). Functional aspects of recollective experience. Memory & Cognition, 16, 309–313. Gardiner, J. M., Ramponi, C., & Richardson-Klavehn, A. (2002). Recognition memory and decision processes: A meta-analysis of remember, know, and guess responses. Memory, 10, 83–98. Gossett,W., writing as “Student” (1908).The probable error of the mean. Biometrika, 6, 1–25. Hand, D. J. (2004). Measurement theory and practice: The world through quantification. London: Edward Arnold. Hirst, W., & Phelps, E. A. (2016). Flashbulb memories. Current Directions in Psychological Science, 25, 36–41. Hirst, W., Phelps, E. A., Meksin, R.,Vaidya, C. J., Johnson, M. K., Mitchell, K. J., . . . Olsson, A. (2015). A ten-year follow-up of a study of memory for the attack of September 11, 2001: Flashbulb memories and memories for flashbulb events. Journal of Experimental Psychology: General, 144, 604–623. Holland, P. W. (1986). Statistics and causal inference. Journal of the American Statistical Association, 81, 945–960. Koppel, J., Brown, A. D., Stone, C. B., Coman, A., Hirst, W. (2013). Remembering President Barack Obama’s inauguration and the landing of US Airways Flight 1549: A comparison of the predictors of autobiographical and event memory. Memory, 21, 798–806. Lakatos, I. (1977). The methodology of scientific research programmes: Philosophical papers (Vol. I). Cambridge, UK: Cambridge University Press. Lanciano, T., & Curci, A. (2012). Type or dimension? A taxometric investigation of flashbulb Memories. Memory, 20, 177–188. Lanciano, T., Curci, A., & Semin, G. (2010). The emotional and reconstructive determinants of flashbulb memory: An experimental approach. Memory, 18, 473–485. Livingston, R. B. (1967). Reinforcement. In G. C. Quarton, T. Melnechuck, & F. O. Schmitt (Eds.), The neurosciences: A study program (pp. 499–514). New York: Rockefeller University Press. Lunneborg, C. E. (2000). Data analysis by resampling: Concepts and applications. Pacific Grove, CA: Duxbury Press. McCloskey, M., Wible, C. G., & Cohen, N. J. (1988). Is there a special flashbulb­memory mechanism? Journal of Experimental Psychology: General, 117, 171–181. MacKay, D. G. & Ahmetzanov, M. V. (2005). Emotion, memory, and attention in the taboo Stroop paradigm: An experimental analogue of flashbulb memories. Psychological Science, 16, 25–32. MacKay, D. G., Shafto, M.,Taylor, J. K., Marian, D. E., Abrams, L., & Dyer, J. (2004). Relations between emotion, memory and attention: Evidence from taboo Stroop, lexical decision, and immediate memory tasks. Memory & Cognition, 32, 474–488. Malmberg, K. J. (2002). On the form of ROCs constructed from confidence ratings. Journal of Experimental Psychology: Learning, Memory, & Cognition, 28, 380–387. Meehl, P. E., & Waller, N. G. (2002). The path analysis controversy: A new statistical approach to strong appraisal of verisimilitude. Psychological Methods, 7, 283–300. Moustaki, I. (1996). A latent trait and a latent class model for mixed observed variables. British Journal of Mathematical and Statistical Psychology, 49, 313–334. Neisser, U. (1982). Snapshots or benchmarks? In U. Neisser (Ed.), Memory observed: Remembering in natural contexts (pp. 43–48). San Francisco: Freeman. Pearl, J. (2009). Causal inference in statistics: An overview. Statistics Surveys, 3, 96–146. Pezdek, K. (2003). Event memory and autobiographical memory for the events of September 11, 2001. Applied Cognitive Psychology, 17, 1033–1045.

Flashbulb memory methods  47

Rubin, D. C. (1992). Definitions of autobiographical memory. In M. A. Conway, D. C. Rubin, H. Spinnler, & W. A. Wagenaar (Eds.), Theoretical perspectives on autobiographical memory (pp. 495–499). The Netherlands: Kluwer Academic Publishers. Rubin, D. C., & Umanath, S. (2015). Event memory: A theory of memory for laboratory, autobiographical, and fictional events. Psychological Review, 122, 1–23. Rubin, D. C., & Kozin, M. (1984).Vivid memories. Cognition, 16, 81–95. Ruscio, J., Haslam, N., & Ruscio, A. M. (2006). Introduction to the taxometric method: A practical guide. Mahwah, NJ: Lawrence Erlbaum Associates. Schuman, H., & Converse, P. (1984). The intersection of personal and natural history. In T. B. Jabine, M. L. Staf, J. M. Tanur, & R. Tourangeau (Eds.), Cognitive aspects of survey methodology: Building a bridge between disciplines (pp. 38–43). Washington, DC: National Academy Press. Simon, H. A. (1954). Spurious correlation: A causal interpretation. Journal of the American Statistical Association, 49, 467–479. Spearman, C. (1904). “General intelligence”: Objectively determined and measured. American Journal of Psychology, 15, 201–293. Tulving, E. (1985). How many memory systems are there? American Psychologist, 40, 385–398. Waller, N. G., & Meehl, P. E. (1998). Multivariate taxometric procedures: Distinguishing types from continua. Thousand Oaks, CA: Sage. Whitehouse, H. (2000). Arguments and icons: Divergent modes of religiosity. Oxford, UK: Oxford University Press. Wright, D. B. (1993). Recall of the Hillsborough disaster over time: Systematic biases of “flashbulb” memories. Applied Cognitive Psychology, 7, 129–138. Wright, D. B. (2006). Causal and associative hypotheses in psychology: Examples from eyewitness testimony research. Psychology, Public Policy, and Law, 12, 190–213. Wright, D. B., & Gaskell, G. D. (1995). Flashbulb memories: Conceptual and methodological issues. Memory, 3, 67–80. Wright, D. B., Gaskell, G. D., & O’Muircheartaigh, C. A. (1998). Flashbulb memory assumptions: Using national surveys to explore cognitive phenomena. British Journal of Psychology, 89, 103–122. Wright, D. B., & Nunn, J. A. (2000). Similarities within event clusters in autobiographical memory. Applied Cognitive Psychology, 14, 479–489. Zur, J. N. (1998). Violent memories: Mayan war widows in Guatemala. Oxford, UK: Westview Press.

3 USING STRUCTURAL EQUATION MODELLING APPROACHES TO BETTER UNDERSTAND THE FORMATION OF FLASHBULB MEMORIES1 Olivier Luminet

We compare different models for the formation of Flashbulb memories (FBMs) based on Structural Equation Modelling (SEM). SEM involves a confirmatory approach for testing hypotheses concerning both the measurement of latent constructs and the structural relationships between them. It allows us to test how well empirical data fit with a pre-existing theoretical model. We compare different models (Conway et al., 1994; Curci & Luminet, 2006; Er, 2003; Finkenauer et al., 1998).We also present recent papers in which new models are tested a) for predicting confidence in FBMs and b) in the case of positive events.

Flashbulb memories (FBMs) are autobiographical memories for the circumstances in which one learned of a public event. In most definitions, researchers do not include first-hand experiences (but see Er, 2003) but they include events that happened within smaller circles, e.g. an event experienced within the family setting (Bohannon, 1988; Brown & Kulik, 1977; Conway et al., 1994; Neisser & Harsch, 1992). The context includes location, time, the presence of others, the informant, activity before and after learning, own affect, own thoughts, and other idiosyncratic details which often include perceptual aspects (Bohannon, 1988; Brown & Kulik, 1977; Conway et al., 1994; Finkenauer et al., 1998; Kizilöz & Tekcan, 2013; Larsen, 1992). FBM is operationalized most of the time by its vividness and consistency of detail. More recently, a model predicting confidence in FBMs was proposed (Day & Ross, 2014). There is controversy, however, on the status of FBM as a consistent memory. In some studies, FBMs have been described as a special class of memories, which persist almost unchanged and are consistent over time (Brown & Kulik, 1977; Conway et al., 1994; Curci & Luminet, 2006; Pillemer, 1984), even after a very long time (Berntsen & Thomsen, 2005). Other studies have concluded that FBMs are neither immune to forgetting nor are they uncommonly consistent as time passes (Christianson, 1989;

Models for formation of FBMs  49

McCloskey, Wible, & Cohen, 1988; Nachson & Zelig, 2003; Neisser & Harsch, 1992; Weaver, 1993). One explanation for these inconsistencies in the definition of FBMs is that an important decrease in consistency can occur in the first year, but then FBMs become remarkably stable as shown recently in a ten-year follow-up study (Hirst et al., 2015). Despite the abundant literature describing FBMs for public events, many papers only report descriptives related to FBM predictors, or only include correlations among these predictors and the degree of vividness of FBMs. Although informative, such analyses only partially explain which variables are necessary for the formation of FBMs, and the way that they are interrelated For instance, they do not account for the different causal paths that can influence in parallel the formation of FBMs. One main goal of this chapter is therefore to present statistical models (Structural Equation Modelling, see details below) that consider multiple paths in predicting the formation of FBMs. We will compare these models based on how comprehensive and broad they are at a theoretical level, and how well they are supported by empirical data. A general consensus is found in the literature about the main emotional, cognitive, and social factors affecting FBM (e.g. Conway, 1995, Finkenauer et al., 1998, see Talarico & Rubin, Chapter 4, this volume). Disagreements still exist, however, not only regarding the emphasis to place on one factor over another, but also on the order of the sequence of relations between variables. Modelling is particularly useful for solving these issues. Indeed, modelling helps in determining which factors are involved and how they relate to each other. The most common statistical tool used for testing models predicting FBM formation is called Structural Equation Modelling (SEM). SEM involves a confirmatory approach, which is useful for testing hypotheses concerning both the measurement of latent constructs, and the structural relationships between them (Bollen, 1989). The main advantage of using the SEM approach is that it allows for the theorized relationships among FBM variables to be tested. In other words, the SEM approach is designed to compare a theoretically established model with empirical evidence to assess the degree to which data do or do not fit the theory, instead of simply deriving a model based upon empirical evidence. It is very important to realize that the results of an SEM approach cannot be considered as the main support for advocating the superiority of one model over another one. SEM is only a statistical tool, not a theoretical rationale. Our conclusions about preferable models accounting for the formation of FBMs can thus only be drawn by simultaneously considering theoretical accounts and statistical information. Some important modifications have been realized since the publication of the first edition of this volume in 2009. Firstly, we no longer consider Brown and Kulik’s (1977) theory as a model of FBM formation. Scholars have interpreted the findings of Brown and Kulik when they developed their own models, but Brown and Kulik did not themselves test a model. We will thus only consider studies which explictly test a model for the formation of FBMs using SEM (Conway et al., 1994; Er, 2003; Finkenauer et al., 1998, and Curci & Luminet, 2006, for

50 Luminet

the particular case of expected events). In addition, we will present two recent papers in which a model of FBM formation was explicitly elaborated and tested. In the first one, Tinti et al. (2014) provided a new model in order to explain the formation of positive FBMs. Although innovative, this model presents important limitations due to the way it operationalized FBM. Then, we present a model that predicts confidence in FBMs (Day & Ross, 2014). The chapter will also compare different models that were tested after 9/11 in a US and a non-US sample (Luminet & Curci, 2009). Although not testing full models, some relevant papers deal with a test of subparts of models and are therefore considered. We will first present the distinction between the types of events considered (extraordinary vs. mundane, Kvavilashvili et al., 2010). Another paper contrasts personally and culturally relevant vs. irrelevant events (Koppel, Brown, Stone, Coman, & Hirst, 2013). We will conclude the chapter with a brief agenda of the important issues that need to be considered for improving our current understanding of FBM formation. In this new edition, we also no longer include the section on comparing fit indices because no recent paper has systematically made this comparison since Luminet and Curci (2009). We do not report, either, the systematic table comparing path models but rather refer in the text to the path models that are the most strongly supported by empirical data. For each model, we will present the basic principles, the type of design adopted to assess FBMs (one shot or test-retest), the main empirical findings that were obtained using these models, the strengths and limitations regarding the model itself and its theoretical background, its ability to fit the data (using the SEM approach), and the appropriateness of events that were analyzed for testing these models.

Conway et al. (1994): The first SEM model for FBM Conway et al. (1994) were the first to test a model predicting the formation of FBMs using SEM. The event investigated was the unexpected resignation of the British Prime Minister Thatcher in 1990 in a large sample of UK and non-UK citizens. Such comparison between groups allows for the examination of whether the level of involvement for the event and/or the activation of social identity are important moderators that require the consideration of different models. FBMs were operationalized as highly consistent memories from one data collection phase (i.e. after 2 weeks) to another (i.e. after 11 months). Memories that did not show a high level of consistency were considered as nonFBMs. In order to be considered as an FBM, a very strict criterion was applied to a list of five FBM attributes (to have at least 90% consistency on this list). An important distinction was made in this model between encoding factors (affect, importance, and prior knowledge) and rehearsal factors. While encoding factors are supposed to predict FBM formation, rehearsal factors are predicted to influence FBM maintenance.

Models for formation of FBMs  51

Main findings The authors first separately examined the model of formation of FBMs and of non-FBMs. The FBM model showed that knowledge/interest was a central variable that had an effect on all the other latent variables (importance, surprise/emotionality and rehearsal). Prior knowledge is suggested to facilitate the organization and assimilation of the incoming information to existing semantic structures in memory. Additionally, importance was also a key variable that directly predicted intensity of affect and rehearsal. Notably, no association was found between affect and rehearsal. In the non-FBM model, knowledge/interest still played a central role. But this time, the association between importance on the one hand, and affect and rehearsal on the other hand disappeared, indicating that in this case, these three constructs have separate independent effects on memory. Finally, they tested a model that involved all participants. In this combined model, FBMs had two direct predictors, affect and rehearsal. Knowledge/interest and importance/consequences were two indirect predictors. Indirect paths indicate that the effect of a variable on FBMs is mediated by one or several other variables. Again, surprise/emotionality and rehearsal were found to be unrelated (see Figure 3.1). This last finding is opposite to most empirical evidence showing that intensity of affect is related to extent of rehearsal (e.g. Rimé et al., 1998). Conway et al. (1994)’s model emphasizes the central role of the encoding factors (affect, importance, and prior knowledge) for the formation of FBMs, while rehearsal does not play a critical role. It would only help the elaboration of memory reports by establishing multiple access routes to FBMs. These authors also shared the assumption of Brown and Kulik (1977) that an event must reach a threshold of significance in order to trigger an FBM. If such a level is not reached, then an association between affect and importance is not present and an FBM is not formed.

Strengths and limitations Conway et al.’s (1994) first study using an SEM approach had a number of strengths. It investigated an event that involved high ratings on two key variables for the formation of FBMs (importance and surprise). It considered group comparisons, which allowed them to show that the structure of the relationships among variables was different for the FBM and the non-FBM groups. Conway et al. (1994) were also the first to investigate the role of prior knowledge (in terms of political knowledge) and attitudes towards the event on FBMs. There were, however, several limitations in this work.The model was data-driven, although it is required that Structural Equation Modelling is built upon theoretical assumptions (e.g. Bollen, 1989). The authors did not consider the appraisal of novelty as a potential key predictor (see below on the model developed by Finkenauer et al., 1998, which emphasizes the role of novelty). Overall, they ignored the central distinction in cognitive theories of emotion between the antecedent of appraisal

52 Luminet

Knowledge/Interest

Importance

Rehearsal

Surprise/Emotionality

FBM FIGURE 3.1 

FBM formation according to Conway et al. (1994).

and the emotional feeling state that follows (e.g. Frijda, 1986; Scherer, 1984). In addition, the global intensity of emotional feelings and the specific state of surprise were considered together, while emotion theories insist on their distinction (e.g. Frijda, 1986; Lazarus, 1982; Leventhal, 1984; Scherer, 1984). Emotional intensity (or arousal) is one fundamental dimension of any emotional state, while surprise is a specific emotional state, involving specific appraisals. Also, the sample mainly consisted of first-year psychology students, which precludes generalizability to other populations. The event itself did not conform to the subjective conditions that typically produce FBMs, which are personal importance, surprise, emotional intensity and rehearsal (see Talarico & Rubin, Chapter 4, this volume). Indeed, although high ratings were observed for importance and surprise, only low levels of affective reactions were recorded, although it is assumed that a threshold for emotionality needs to be reached in order to have FBMs. Additionally, rehearsal was of low magnitude. But for Conway the key issue is the presence of the phenomenological features of FBMs rather than high scores on the various predictors.

Models for formation of FBMs  53

Finkenauer et al. (1998): A two-path model According to Finkenauer et al. (1998), the process of formation and maintenance of FBMs develops through two pathways. The first connects the cognitive evaluations of novelty to FBMs through the effect of surprise and emotion. This pathway is representative of the direct effect of emotion. The second pathway represents the indirect effect of emotion. The experience of learning a relevant piece of public news induces high levels of personal importance and consequentiality, which in turn lead to intense (negative) emotions. Intensity of emotion does not directly affect FBMs. Rather, its impact is mediated by two consecutive processes. First, intensity of emotion triggers rehearsal, which strengthens the memory trace of the original event. This eventually enhances memory for the reception context. In this process, a central role is also played by the latent variable involving both background knowledge and affective attitudes, which influence importance/ consequentiality, emotional reactions and rehearsal (Conway et al., 1994; Finkenauer et al., 1998; see Figure 3.2).

Novelty

Importance

Surprise

Emotionality

Attitudes/Knowledge

Rehearsal

FBM

Event memory

formation according to Finkenauer et al. (1998). Dashed lines are paths that were found in a subsequent test of the model (Luminet & Curci, 2007).

FIGURE 3.2  FBM

54 Luminet

Main findings Finkenauer et al.’s (1998) study involved a large group of French-speaking Belgians (N = 399) of various age groups and backgrounds. Seven to eight months after the unexpected death of the Belgian King in 1993, participants recalled the circumstances in which they first heard the news, specific details about the original event (memory of the original event) and about the following predictors of FBM: (1) personal consequentiality and importance, (2) novelty, (3) surprise, (4) emotional feelings, (5) overt rehearsal (i.e. frequency of talking about the event and following the media), and (6) personal attitude, which was assessed through the degree to which the person reported having a favorable attitude toward the Royal family in general.2 It is worth noting that although Finkenauer et al. (1998)’s results were based on a single time measurement, later papers testing this model were based on a consistency measure (Curci & Luminet, 2009; Luminet & Curci, 2009). The authors compared their model with the ones developed by Brown and Kulik (1977) and Conway et al. (1994) using the SEM approach. Results showed that Finkenauer et al.’s (1998) model evidenced the best values of fit (CFI, AIC) and was often the only model that had indices above standard thresholds (such as chi-square/ df ratio or RMSEA).3 Also, all (except two) causal path linking factors were significant, while some critical paths suggested by the other models were not significant.

Strengths and limitations One main advantage of this model over the previous one was a more refined analysis of the emotional process that included a distinction between appraisals and feeling states and a distinction between surprise and other emotional feelings. This allowed certain appraisals (like novelty) to be more closely linked with certain specific emotional feelings, such as surprise. Finkenauer et al. (1998) also laid out a detailed analysis of the relationship between immediate emotional responses and rehearsal.They predicted that rehearsal improves FBMs indirectly through the consolidation of event memory. At the structural level, these data are the first to support a two-path model, in which a distinction is made between direct and indirect determinants of FBMs (see later in this chapter and Tinti et al., 2014, for another two-path model). In addition to the direct path between surprise and FBMs, the model includes an indirect path. It starts with an appraisal of importance of the event, which then triggers emotional feelings. In turn, intensity of emotion predicts the extent of rehearsal. Rehearsal processes are likely to address information about the original event rather than information relative to the reception context. Hence, Finkenauer et al. (1998) proposed that the intensity of emotional feeling state determines event rehearsal and, as a consequence, the memory of the original event or event memory. The event memory, in turn, influences the maintenance of the FBM. Theoretically, it is suggested that when learning about an original event, all currently activated information – including the reception context, sensory information, and the original

Models for formation of FBMs  55

event – is encoded in memory (Tulving & Kroll, 1995).Thus, both the original event and the reception context would immediately be encoded. Subsequently, during rehearsal of the original event, the reactivation of this information in memory would spread to any associated information, and thus to an FBM (for a similar perspective see Larsen, 1992 for his notion of rehearsal displacement). In this way, rehearsal strengthens the associations between the different elements constituting the memory of the entire experience (Johnson & Chalfonte, 1994). Finkenauer et al.’s (1998) model was the first to consider event memory as a mediating variable between rehearsal and FBM, which is now widely accepted in the literature (e.g. Hirst & Phelps, 2016). It also assigned an equally central role to the direct and indirect paths in the formation of FBMs, assuming that some FBMs can be formed when only one path is activated (see below Curci & Luminet, 2009 for an expected event, and Luminet & Curci, 2009 for differences in US and non-US respondents after 9/11). Finally, from a statistical viewpoint, this was the first attempt to compare different models of FBM formation at the structural level. This statistical comparison between different model fits is crucial for selecting the variables that are central for the formation of FBMs, and also for understanding the structural dynamics of the relationships between them. Further, it allows researchers to decide if the models as they are can be generalized to all types of events or if models need to be specific with respect to the type of event investigated. For instance, we will show later in this chapter that small adaptations of the original Finkenauer et al.’s (1998) model were implemented when applied to the 9/11 attacks. Despite these numerous strengths, there were also some notable limitations in this study. A major one is that FBM was not considered as a consistency measure, because memory for the reception context was assessed only once, seven to eight months after the original event. Also, only one group (Belgian citizens) was considered, while it would have been interesting to include comparison groups of people for whom the news was less relevant, such as citizens from neighboring countries. Finally, only attitudes were investigated, while a combined investigation of attitudes and prior knowledge would be preferable, as the first aspect involves mainly affective components and the second mainly cognitive ones. These three limitations were addressed, however, in more recent studies testing Finkenauer’s model (see below Curci & Luminet, 2009; Luminet & Curci, 2009).

A specific application of Finkenauer et al.’s (1998) model to expected events Although surprise is often described as a necessary condition for eliciting FBMs, some studies have found FBMs for expected news (Bellelli, 1999; Curci, 2005; Curci et al., 2001; Neisser, 1982; Ruiz-Vargas, 1993; Winograd & Killinger, 1983). For instance, Curci et al. (2001) compared FBMs for the death of the former President of France, François Mitterrand, in two national groups, i.e. French and Belgian citizens.The death was expected since Mitterrand suffered from prostate cancer and

56 Luminet

he was in a terminal state. Nevertheless, the authors found that FBMs developed for the event in both groups. Curci and Luminet (2009) have tested the model of Finkenauer et al. (1998) on these data. Due to the nature of the event that was expected, they predicted that the direct impact of novelty and surprise on FBMs would be non-significant, but that both cognitive-emotional and social factors would influence FBMs through the indirect pathway.

Main findings As predicted, the direct path was not found, but all the relationships in the indirect path were significant. These results suggest that an event that is expected could also lead to FBM formation. One specific feature, however, is that only the indirect path is activated, emphasizing a stronger role for reconstructive processes, through rehearsal of the original news. This means that FBM consistency can in some cases be explained only by post-encoding elaboration.

Strengths and limitations This study is the first to examine the formation of models for expected events. Other studies are thus necessary to test whether the model can hold across situations. This study also suggests some important issues for the implication of novelty and surprise in the formation of FBMs and for their measurement in future studies. Novelty refers to two aspects: unexpectedness and exceptionality (Frijda, Kuipers, & ter Schure, 1989; Leventhal & Scherer, 1987; Scherer, 1988). In other words, an event might be considered novel not only in terms of its sudden and unexpected occurrence, but also in terms of its disruptiveness with respect to one’s ordinary routine. This distinction was not tested by Curci and Luminet (2009). Had it been done, we could surmise that respondents would have rated the event as exceptional and out of the ordinary, particularly in the French group, but not as unexpected due to the former French president’s terminal disease. New testing of FBM models (especially in the case of predictable events) should consider this twofold nature of novelty, thus taking into account the effects of both evaluations, especially for events foreseeable by the public. Recently, Lanciano, Curci, and Soleti (2013) examined FBMs for the expected death of Pope John Paul II. The authors did not use a SEM approach but they examined in a regression analysis the variables predicting FBM consistency. They found that positive attitudes and greater prior knowledge were the only two significant predictors. These findings are consistent with Conway et al.’s model (1994), which also assigned a central role to these variables for predicting FBM consistency.

Er (2003): A model for people who directly experienced the event Er (2003) investigated models of FBM formation related to a severe earthquake that occurred in Turkey in 1999. She suggested two models as a function of

Models for formation of FBMs  57

whether one was a victim of the event or not. She assumed that the recollection of victims would be markedly different from those who had no direct experience of the event. It follows that two different models would be necessary to account for memories of these two groups. For the victim group only, she suggested a model in which personal consequences had a direct effect on the formation of FBMs, while the effect of personal consequences on FBMs was only indirect in the non-victim group. Another major difference between the two groups is related to the distinction between FBMs and event memories. In the victim group, the two aspects cannot be separated. If the person directly experiences a major event, he/she cannot differentiate between the memory for the event itself and the memory for the specific details for the context in which the event was received.

Main findings In these analyses, FBM was operationalized by its vividness. For each group considered, a comparison was made between the new model (labelled “importance-driven emotional reactions”) and the previous ones (Brown & Kulik, 1977; Conway et al., 2004; Finkenauer et al., 1998). For the victim group (see Figure 3.3), SEM revealed that the new model had better fit indices than previous models. Results also showed a direct relationship between importance and FBMs, which is consistent with the hypothesis of Brown and Kulik (1977) but contradicts the findings of Conway et al. (1994) and Finkenauer et al. (1998). For the non-victim group, the path diagrams were very close to Finkenauer et al.’s (1998) model, except that surprise and novelty were considered as a single variable. The comparison of fit indices revealed that they were above standard thresholds for the models of Er (2003) and Finkenauer et al. (1998), while they were below these thresholds for the models of Brown and Kulik (1977) and Conway et al. (1994). One aspect that made this study unique is that the group involved was made of participants who experienced the event itself rather than just having heard about it. Furthermore, experiencing a major natural disaster, such as an earthquake, is likely to leave people traumatized for a long period of time. The intense fear that was experienced may also have lead some of them to develop post-traumatic stress disorder (PTSD) symptoms. Unfortunately, the author did not provide information related to the prevalence of PTSD symptoms in the victim group.We can, however, assume that a proportion of them formed traumatic memories and their forgetting rate was less steep than for non-traumatic ones (e.g. Neisser et al., 1996). Also, traumatic memories are assessed as being more emotional (e.g. Fivush, Bohanek, Marin, & McDermott Sales, 2009), and people with PTSD exhibit an inflation of the emotional aspects of their memory, (e.g. Qin et al., 2003, see Tat, Gold, & Budson, Chapter 7, this volume). These data suggest that victims are likely to have a different memory content than non-victim groups. Er’s model thus seems specifically relevant when examining FBM formation for people directly exposed to a traumatic situation.

58 Luminet

Novelty/Surprise

Importance

Emotionality

Rehearsal

FBM/Event Memory

FIGURE 3.3 

FBM formation according to Er (2003) for the victim group.

Strengths and limitations Er (2003) was the first to provide a model of FBM formation for events in which people were directly involved. In this case, the event itself and the reception context cooccur in space and time.This situation is rarely examined in the literature (but see Pillemer, 2009). Many scholars restrict the definition of FBM to remembering the specifics of the context in which they first heard a piece of news rather than directly experiencing the newsworthy event, implying that this study may not be a study of FBMs at all. But, if an extended view on FBMs is considered, Er’s (2003) model is certainly worth considering as it emphasizes important features present when people are directly exposed to the event. One such feature is very high emotional intensity. Another is the enhanced effect of importance/consequences of the event on FBMs. Consistent with these findings, Neisser et al. (1996) already showed that personal involvement was the main predictor of recall in a study that compared three groups of subjects on their memories of a major earthquake in California. Three other strengths of the Er’s study are worth mentioning. First, the selection of the event was consistent with the FBM hypothesis that requires high ratings on

Models for formation of FBMs  59

personal importance and consequentiality, affective reactions, and surprise (Conway, 2002). Second, a large comparison group (N = 320) was recruited, which allowed the researcher to test separate models for more and for less-involved people. Third, this was the first time that the questionnaire measuring FBM and its predictors was closely based on another study (Finkenauer et al., 1998) that tested model formation. This allows for more direct comparisons across studies because it is always possible that differences across studies could be explained by the different ways the variables were operationalized. Some important limitations also need to be mentioned. First, the initial data collection took place six months after the event. This means that the “original” memory that was investigated at that time had already been reconstructed through media exposure and social communication. Also, consistency measures were available only for a small proportion of the total sample (about 12%), with many analyses based on the single initial measure collected after six months. These methodological issues lead us to cautiously consider the validity of the assessment of the FBM latent variable. Various concerns relate to the measurement of predictors of FBM. First, there was no direct assessment of prior knowledge, despite the fact that previous models emphasized its role in FBM formation (Conway et al., 1994; Finkenauer et al., 1998). Second, the appraisal of novelty and the emotional feeling state of surprise that followed were not conceptually distinguished.Third, there was a confounding variable in the measure of surprise that made it more conceptually related to the measure of emotionality. Finally, the author improved the model by considering mathematical indices (modification indexes) that were proposed by the statistical software in order to increase the overall fit, rather than following a theoretical rationale, as is required when using SEM.

Towards a systematic comparison of the four models of FBM formation An important question is to decide which model(s) need to be selected for predicting the formation of FBMs. The above sections mainly described studies in which one model at a time was considered. Systematic comparisons are necessary, however, for determining which model should be preferred, and under which conditions. In this second part of the chapter, we first present a study in which four models predicting FBM formation were compared (Luminet & Curci, 2009). We also review two recent models (Day & Ross, 2014;Tinti et al., 2014) that emphasize other important paths. We then review other data that did not test a full model but that considered how certain characteristics of the event can modify paths that predict the formation of FBMs (Koppel et al., 2013; Kvavilashvili et al., 2010).

A model comparison of the 9/11 attacks Luminet and Curci (2009) conducted a study in which they compared four models of FBM formation related to the 9/11 terrorist attacks in the US in 2001 (Brown & Kulik, 1977; Conway et al., 1994; Er, 2003; Finkenauer et al., 1998). They used

60 Luminet

SEM for testing separate models based on the level of social identity activated in participants. The group that more strongly identified with the victims of the 9/11 attacks (high social identity group) was composed of 120 US participants, and the group that identified less with the victims (low social identity group) included 716 participants from different European countries.

Main findings FBMs were considered on a consistency basis with data collected on average 21 and 524 days after the 9/11 attacks. FBM measures included time of day, location of the respondents, informant, other people present, and ongoing activity. Each item was assessed on a score from 0 (least consistent) to 2 (most consistent) and a total consistency score was computed based on the sum of the five FBM items. Regarding fit indices, a model derived from Finkenauer et al. (1998) applied to the data for the 9/11 attacks provided the best indices for both the US and the non-US samples. The main modifications from the original model of Finkenauer et al. (1998) were related to the additional effects of novelty of the event on the other predictors.While the initial model only showed a relationship between novelty and surprise, the present data also showed a relationship in both groups between novelty and other emotional feeling states. This suggests that novelty has a more general impact on the emotional response than was originally assumed. In addition, novelty was positively related to event memory, although only in the non-US group. This suggests a purely cognitive effect, as novelty is related to an initial orientation reaction that directs attention and favors the encoding of novel information. For the US group, the derived model of Finkenauer et al. (1998) was the only one for which all fit indices reached the required thresholds. We will now turn to a brief description of the model of prediction. A major difference was observed in the two groups regarding the activation of the direct and the indirect paths. In the US group, strong significant relationships were found for the direct path (novelty – surprise – FBMs), while the indirect path was incomplete, with no significant relationship between rehearsal and event memory. In the nonUS group, the direct path was not significant, while the indirect path was complete, with all the predicted relationships being significant. In both groups of respondents FBMs were formed, suggesting that the activation of either the direct or the indirect path is sufficient for the elicitation of FBM. One difference between the two groups is thus related to the relationship between rehearsal and event memory. We suggest that the content of rehearsal might explain these results. Large differences occurred between the US and nonUS groups on the impact of the event on people’s personal lives (Luminet et al., 2004). It is likely that this impact was more emphasized in the US than the European media. It is also likely that this difference extended to other aspects of rehearsal (rumination and social communication). Thus, the content of rehearsal was less likely to strengthen event memory as a large proportion of this information was not related to the content of the news itself. In the non-US countries, there was less need to focus the content of rehearsal on the impact on people’s

Models for formation of FBMs  61

personal lives. It is thus likely that a higher proportion of rehearsal was centered on the event itself. The more event-centered focus of rehearsal in non-US countries was more likely to strengthen event memory. Another important finding was the non-significant direct path in the non-US group. We suggest that the activation of social identity (in this case only for US respondents) could be a central component for the activation of the direct path (Luminet & Curci, 2009). Depending on circumstances, the individual’s personal identity (self-definition derived from close personal relationships and idiosyncratic personality attributes) or social identity (self-definition in terms of the properties of a specific group and degree of an individual’s identification with this group) will be more highly activated. For events that only trigger personal identity (see for example, Conway et al., 1994), previous studies show that no direct link between surprise and FBMs is found. However, when social identity is central, the positive link between surprise and FBMs is found (e.g. Finkenauer et al., 1998). For more details on the effect of social identity on the formation of FBMs, see Berntsen, Chapter 9, this volume, and Luminet & Curci, 2009. We will now look at the structural results of applying other models to these data. Overall, the results revealed that many important predictions were not supported. The failure of Conway’s model to account for the present data is particularly noteworthy for the US group, as the two direct predictors of FBM (affect and rehearsal) were found to be non-significant. Moreover, prior knowledge failed to predict affect and importance. Only the relationship between prior knowledge and rehearsal was significant. The application of Er’s (2003) model to the present data showed that for the US group, three out of six predicted paths were non-significant. For the non-US data, the model that was suggested for the non-victim group was actually very close to the Finkenauer et al.’s (1998) model.The only main difference is that Er (2003) predicted a direct relationship between affect and FBM that is not predicted in Finkenauer et al.’s (1998) model. The results contradicted Er’s (2003) assumptions about this relationship for the events of 9/11.

Strengths and limitations The event that was investigated is highly prototypical for eliciting FBMs as it involved strong responses on the key predictors of novelty/surprise, importance/ consequentiality, and affective reactions. In this respect, it corresponds perfectly to the conditions for an appropriate test of the formation of FBMs. This study also represents a strong improvement in modelling FBM when compared to the previous studies that used SEM. In comparison with Conway et al. (1994), the present sample is more representative of the general population in terms of age and background. We also made a distinction between intensity of affect and surprise and we had high ratings on all the key variables, which was not the case for affect in Conway’s model. In comparison with the data collected by Finkenauer et al. (1998), the data collection for this study took place much sooner after the event, the assessment of FBMs was based on a test-retest consistency measure, and we had a group comparison.

62 Luminet

Some future improvements need to be considered, however. It would be important in the future to have a larger sample size, particularly for the high social identity group. Also, the first data collection should have been completed more quickly in order to minimize the possibility of reconstructive processes taking place.

Tinti et al. (2014): Positive events and a distinction between two separate forms of rehearsal Tinti et al. (2014) tested a model for the formation of FBMs related to the winning of the Italian football team during the 2006 World cup. They made a distinction between two types of rehearsal: collective (through the media) vs. individual (social sharing and rumination).They hypothesized that two types of rehearsal would have independent effects on event memory and on FBM. Using SEM, they showed that only individual rehearsal predicted consistency of FBMs, while collective (media) rehearsal was a unique predictor of event memory accuracy (See Figure 3.4).

formation according to Tinti et al. (2014). Plain arrows: paths expected to be significant; short-dashed arrow: path expected not to be significant; long-dashed arrow: no prediction.

FIGURE 3.4  FBM

Models for formation of FBMs  63

Media rehearsal mainly refers to factual aspects of the event (e.g. who scored a goal, the score when the game time elapsed, who failed to make the decisive penalty kick).This would involve repeated consultation of information about the event across different media, thereby enhancing only event memory. The other type of rehearsal (i.e. talking and thinking) focus on the person’s experience (e.g. his or her own happiness and the happiness of others around, the noise of the honking cars and bursting firecrackers, etc.). In this case, in which a more personal experience is involved, thinking and talking about the event strengthen FBM consistency. These results emphasize that while FBMs are determined by specific, idiosyncratic, and subjective emotional experiences and rehearsal, event memories are determined by the more objective measure of the degree of media exposure and also by level of pre-existing knowledge (Tinti et al., 2014). Another important difference between Tinti et al.’s (2014) model and the one developed by Finkenauer et al. (1998) is that the link between FBMs and event memories was not significant in their data.This result is very intriguing because it is in sharp contrast with the existing literature. Indeed, Luminet (2009) found that all five studies using SEM that tested for this path reported a significant positive association between FBMs and event memories. A number of reasons can be advanced for this difference. Firstly, Tinti et al. (2014) made a distinction between two aspects of rehearsal (descriptive through the media vs. experiential through thinking and talking, see above). It is possible that if previous studies had made such a distinction, they would have also found a non-significant association between memories for the context and memories for the event. Secondly, Tinti et al. tested their model on a positive event. It is possible that event memory is associated with FBM only for negative events. But there are also some conceptual and methodological issues suggesting that these results should be considered with caution. Indeed, the way event memories and FBMs were assessed in Tinti et al.’s study is very different from other studies and is also highly questionable. For FBMs, a composite score was computed based on degree of vividness, number of details, and degree of certainty. It is rather unusual to mix these three aspects, especially considering that one correlation (between the number of details and degree of certainty) was very low (r = .04) and that the two aspects that are assumed to be highly correlated (vividness and details) had only a modest correlation of r = .25. In addition, loadings on the FBM latent variable were very low for both the number of details (.27) and certainty (.39). These data suggest strong independence between the three items and call for examining them as separate dimensions of FBMs. Finally, the index of details was obtained by simply summing up dichotomous items concerning the presence/absence of a given detail in participants’ accounts. Data from a large body of studies from Neisser and Harsch (1992), call into question the additive model underlying this simplistic measurement procedure for assessing FBMs (see Curci, Chapter 1, this volume, for a review). For event memory, the authors used a composite measure including accuracy and certainty. Although it is important to consider certainty in event memory in addition to accuracy, this

64 Luminet

perspective prevents any comparison with previous studies in which only accuracy was considered. We thus face the problem of comparability with such different operationalizations of both FBM and event memory. But beyond these conceptual differences, at a psychometric level it is difficult to defend the idea of combining within the same score measures that do not correlate at all with each other.

Strengths and limitations Clearly more studies are necessary to support Tinti et al.’s (2014) model as a general model for FBM formation or as a specific one for predicting positive events. One asset of their model is that it considers how the content of rehearsal can have independent effects on event memories and on FBMs. Some studies already support this dissociation by showing a positive association between media coverage and event memories and not between media coverage and FBMs (e.g. Curci & Luminet, 2006; Er, 2003). But this is the first study that fully considers the differential effect of a form of rehearsal that is more collective and factual from one that is more individual and subjective. Another contribution of Tinti et al.’s (2014) model is that it accounts for positive FBMs. There is clearly a lack of systematic investigation of pleasant FBMs in the literature (see Stone & Jay, Chapter 8, this volume). But this strength also reveals a main limitation of the model. There is no a priori reason to predict that differentiating between collective-factual and individual-personal rehearsal would better account for positive rather than negative events. Obviously, future studies would need to compare, in the same sample, FBMs triggered by a negative event with those triggered by a positive event and examine whether the valence of the event leads to different outcomes. To conclude, the conceptual and psychometric issues raised above call for considering Tinti et al.’s. (2014) model with a high degree of caution.

Day and Ross model: Considering confidence in FBMs Recently, Day and Ross (2014) proposed a new model (Figure 3.5) in which they predicted FBM consistency (i.e. measured with the canonical categories about the news of Michael Jackson’s death shortly after his death and 18 months later) and respondents’ confidence in their FBM account. They considered affective attitude towards the event (i.e. the degree of social bond with Michael Jackson) as the first condition for FBM formation. Stronger affective attitude was predicted to be associated with more personal reactions to the event, such as greater surprise, emotional intensity and rehearsal. Similar to Finkenauer et al. (1998), they expected these three variables to be related: elevated surprise should lead to greater emotional intensity, which in turn would predict more rehearsal. Finally, they expected only surprise to predict FBM consistency. Their results confirmed these assumptions. Participants’ affective attitude towards the pop star influenced their response to the news: The more they were attached

Models for formation of FBMs  65

Surprise

Social Bond

Emotional Intensity

Flashbulb Memory Consistency

Flashbulb Memory Confidence

Rehearsal

FIGURE 3.5 

FBM formation according to Day and Ross (2014).

to him, the more they reacted to the news with surprise and intense emotions, and the more they talked and thought about their memory. Consistent with previous studies, they found that emotional intensity and rehearsal are not direct predictors of FBM consistency. An important dissociation was also evidenced with surprise, predicting specifically FBM consistency, while rehearsal predicted only FBM confidence. Finally, FBM consistency and FBM confidence were significantly, but moderately, associated.

Strengths and limitations An important contribution of Day and Ross’s (2014) model is to add FBM confidence and to consider it concurrently with FBM consistency. This is an important improvement, although future studies should ideally consider also FBM vividness. Doing so will allow a comparison of models that conceptualize FBMs in different ways and an examination of whether the structure of predictors differ as a function of FBM consistency, vividness, or confidence. One limitation of Day and Ross’ (2014) study is the low sample size (N = 135), which did not allow for a full SEM in which each latent variable is assessed by a set of observed variables. Instead, each variable was assessed either with a single item, or, when multiple items were considered, they were averaged to form a single index. Another limit is that the authors did not consider event memory, which precludes testing whether rehearsal affects FBM consistency indirectly, through event memory. Adding event memory would be a way to assess how strong the indirect path (linking surprise to emotional intensity, emotional intensity to rehearsal, rehearsal to event memory, and event memory to FBM) would be, in comparison with the direct path (surprise to FBM consistency) for which a significant, but weak relationship was found.

66 Luminet

Neither is it clear why affective attitude is the first step of Day and Ross’ (2014) model. Finkenauer et al.’s (1998) model would rather consider that surprise is not dependent on affective attitudes but is a more automatic feeling triggered by the appraisal of novelty. In this regard, it is also a limitation of their study that appraisal dimensions such as novelty or importance were not measured. A recent study sheds some light on the variables predicting confidence in FBMs (Stone, Luminet, & Takahashi, 2015).The authors investigated FBMs for two events that occurred almost concurrently when Japan and Australia experienced the resignation of their respective Prime Ministers in June 2010. The study allowed Stone et al. to examine whether cultural differences can account for variations in the predictors of FBM formation (Wang & Aydin, Chapter 12, this volume, for a more detailed review of culture and FBM). It is important to keep in mind that this study did not adopt an SEM approach. Major differences were found across countries for the association between FBM predictors and FBM confidence. In the Japanese sample, the only significant correlation with confidence was found with national importance. For the Australian sample, however, novelty, surprise, importance (being personal, national or international), rehearsal, event memory accuracy, attitudes and political involvement were all found to be correlated with confidence. These results suggest that confidence in FBMs is predicted by different variables depending on whether the respondents are from a country which is considered to be more collectivistic, such as Japan, or whether they are from a more individualistic country, such as Australia. For the Japanese sample, the main predictors of FBMs still need to be identified (see also Wang & Aydin, Chapter 12, this volume).

How the type of event can matter In addition to the difference between types of rehearsal emphasized by Tinti et al. (2014), other recent studies have shown the importance of making distinctions between the types of events considered (extraordinary vs. mundane, Kvavilashvili et al., 2010; personally and culturally relevant vs. irrelevant, Koppel et al., 2013).These last two studies did not test models of FBM formation. They do, however, provide relevant information regarding some important predictors. Kvavilashvili et al. (2010) examined two negatively valenced events, the first being extraordinary but also public (i.e. the 9/11 terrorist attacks), the second being mundane and private (i.e. participants learning that they had not won a small prize). The correlation between rehearsal and FBM consistency was positive for the mundane event whereas it was negative for the exceptional event. The explanation is that extraordinary events, such as 9/11, are constantly rehearsed through conversations and media. These numerous rehearsals create multiple different “memories” of the first time individuals heard the news of 9/11 and thereby lead to inconsistent FBMs (see also the notion of wrong time slice, Brewer, 1992).

Models for formation of FBMs  67

Koppel et al. (2013) distinguished events according to their level of cultural significance. They selected two public events which occurred almost at the same time: the inauguration of Barack Obama on January 20, 2009 and the emergency landing of US Airways Flight 1549 off the coast of Manhattan on January 15, 2009. The former was considered culturally and personally significant given its historic nature, while the latter was considered noteworthy but less culturally significant. Through multiple regression analyses, these authors examined which variables predicted both event memory accuracy and FBM consistency. The results showed that variations on the two memory variables occurred only across events but not across memory measures, suggesting that the type of event can be a more central issue than the type of memory considered. For the plane landing, which is an event that is less culturally important, rehearsal predicted both event-memory accuracy and FBM consistency. For the inauguration of Obama, which was more culturally important, rehearsal was not a predictor of either accuracy of event-memory or consistency of FBM, while personal importance and emotional intensity predicted both memory types. These results suggest that when an event has a high level of cultural relevance, more immediate, spontaneous affective reactions determine the formation of FBMs, while when the event is also noteworthy but less culturally relevant, factors that are later activated, such as rehearsal, take a more central role.

Comparison across models In the 2009 edition, we presented a detailed account of the proportion of studies that supported direct predictors of FBMs (emotions, rehearsal, surprise, importance, event memory) as well as support for intermediate links in the models (e.g. importance => rehearsal, knowledge => importance, rehearsal => event-memory, etc.). New studies published since then have considered some of these relations.The only relevant relationship for which there is substantial new data is the predicted direct link between rehearsal and FBMs. The 2009 review showed that six out of nine studies supported such a link.This relationship was also supported by Tinti et al. (2014), by Kvavilashvili et al. (2010), but only if the event was mundane, and by Koppel et al. (2013), but only if the event was of low cultural relevance. At the same time, Day and Ross (2014) did not find a significant link between rehearsal and FBMs. In addition, no relation was found by Koppel et al. (2013) when the event was of high cultural relevance and even a negative relation was found by Kvavilashvili et al. (2010) when the event was exceptional. These new findings emphasize that the direct link between rehearsal and FBMs might be among the most complex ones to understand. One issue for the future will be to investigate the individual versus collective dimensions of rehearsal and their effects on both types of memories in addition to the more common distinction between covert (rumination) and overt (social communication, following the media) rehearsal. Recent studies also suggest that additional predictors for event memory accuracy need to be considered, such as knowledge, media rehearsal and novelty (Luminet & Curci, 2009; Tinti et al., 2014).

68 Luminet

Conclusion We showed in this chapter that the use of an SEM approach represents an important improvement for the understanding of FBM formation. It allows for the comparison of theoretically-founded models with data from events in which FBMs were observed, and provides two levels of information: the overall fit and the relationships among predictors. The review of the events that lead to the formation of FBMs showed that a large number of situations can be investigated, even when they are related to low surprise (Curci & Luminet, 2009) or to low rehearsal (Conway et al., 1994). The comparison of models indicates that the Finkenauer et al. (1998) model provided the best fit with the data. The comparison of the Finkenauer et al. (1998) model across different studies revealed that although the same structure is observed across studies, the level of activation of the paths can change from one study to another. One important conclusion is that the activation of either only the direct path or only the indirect path is sufficient for the formation of FBMs (Luminet & Curci, 2009). Another important conclusion is that the following eight predictors need to be considered: novelty, surprise, importance, emotional feelings, attitudes, prior knowledge, rehearsal, and event memory. The recent study by Tinti et al. (2014) also suggests differentiating between different forms of rehearsal (e.g. among media rehearsal, social communication, and thoughts). Future studies should continue to compare models in order to identify the most relevant variables to consider. A paradoxical result evidenced by two recent studies showed that events that were more mundane or less culturally relevant were more likely to activate associations between rehearsal on the one hand and FBM consistency and event memory accuracy on the other hand (Koppel et al., 2013; Kvavilashvili et al., 2010), while these associations did not occur for exceptional or highly culturally relevant events. It suggests that results related to these latter events might be explained by models other than the ones currently tested. The findings obtained by Luminet and Curci (2009) already suggest the need to adapt Finkenauer et al.’s (1998) model to account for the level of social identity activated in the group. Finally, we would like to emphasize some recommendations for future studies that will allow for the improvement of the models of FBM formation. First, we think that there is a strong need for a standardized questionnaire. This will provide the opportunity to make direct comparisons across studies when important events such as the 9/11 attacks happen. It will be important to collect the first data set very soon after the event, using different groups that vary in their level of involvement for the news. The two components of novelty (unexpectedness and exceptionality) also need to be considered as they are assumed to have a different impact on the other predictors of FBMs.The generalizability of the models that were presented in this chapter can be confirmed only when a large set of events of different natures (assassinations of public figures, natural disasters, terrorist attacks, etc.) are assessed with SEM. It will also be important to test more extensively if some particular situations (e.g. being a victim rather than being an observer) can be accounted for

Models for formation of FBMs  69

by the models. A closer examination of the type of emotional feelings involved can also be an interesting avenue for future research. Until now, only the overall intensity of emotional feelings has been considered, but not the type of emotions involved. It will be important to test if events related to strong positive emotions or to some particular negative states (e.g. shame, guilt, disgust) also lead to the same structure of FBM formation. An important future step will be to systematically examine consistency, vividness, and confidence within the same study in order to provide a more comprehensive view of models predicting FBM formation. Current research suggests, for instance, that there is a moderate, positive relation between consistency and confidence of FBMs. But a closer examination indicates that the situation is more complex, ranging from studies finding strong (Schmolck, Buffalo, & Squire, 2000; Weaver, 1993) to moderate associations (Kvavilashvili et al., 2010;Winningham, Hyman, & Dinnel, 2000), and also including some studies reporting non-significant associations (Neisser & Harsch, 1992; Talarico & Moore, 2012; Talarico & Rubin, 2003). We can thus conclude from these findings that although higher confidence, in most instances, helps to strengthen consistency, the often modest magnitude of the relationship suggests that confidence cannot be taken as a pure proxy for consistency (see Lanciano, Curci, & Sartori, Chapter 6, this volume). This suggests that the different operationalizations of FBM would be predicted by different variables. Although using the vividness or the consistency approach for Finkenauer et al.’s (1998) model does not result in major differences (Finkenuer et al., 1998; Luminet & Curci, 2009), the recent model developed by Day and Ross (2014) suggests that the type and the structure of predictors for FBM confidence might differ.

Notes 1 The author thanks Antonietta Curci, who computed SEM analyses that are presented in this chapter, and Antonietta Curci and Jennifer Talarico for their helpful comments on an earlier draft of the chapter. 2 This latter variable is not equivalent to Conway et al.’s prior knowledge about politics. It was believed that both variables reflect closely-related constructs. More recent studies usually assess both attitudes and knowledge under the same latent variable. 3 The Comparative Fit Index (CFI) is used as an incremental fit index derived from the comparison of the hypothesized model with the null model, in order to determine the amount by which the fit is improved by using the hypothesized model over the null model (Bentler, 1989). The CFI is a measure of covariation in the data reproduced by the model. Values range from 0 to 1. Values greater than .90 can be considered as satisfactory (Bentler & Bonett, 1980). Akaike’s Information Criterion (AIC; Akaike, 1987) assesses how parsimonious a model is. This criterion takes the goodness-of-fit as well as the number of estimated parameters into account. The model that yields the smallest value of AIC is considered best. The ratio between the model chi-square and the degrees of freedom indicates the fit of the model per degree of freedom used. A threshold value of 2 or 3 (Carmines & McIver, 1981) is usually required. The Root Mean Square Error of Approximation (RMSEA) is a measure of the approximation between the model and the population covariance matrix. A perfect fit will yield a RMSEA of zero. Values equal to or lower than 0.05 are considered good. Values exceeding .10 are indicative of a poor fit (Browne & Cudeck, 1993).

70 Luminet

References Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52, 317–332. Bellelli, G. (1999) Ricordo di un giudice. Uno studio sulle flashbulb memories [Remembering a judge: A study on flashbulb memories]. Napoli: Liguori. Bentler, P. M. (1989). EQS: Structural equations program manual. Los Angeles: BMDP Statistical Software. Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness-of-fit in the analysis of covariance structures. Psychological Bulletin, 88, 588–606. Berntsen, D., & Thomsen, D. K. (2005). Personal memories for remote historical events: Accuracy and clarity of flashbulb memories related to World War II. Journal of Experimental Psychology: General, 134, 242–257. Bohannon, J. N. III. (1988). Flashbulb memories for the Space Shuttle disaster: A tale of two theories. Cognition, 29, 179–196. Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley. Brewer,W. F. (1992).The theoretical and empirical status of the flashbulb memory hypothesis. In E. Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb” memories (pp. 27–305). New York: Cambridge University Press. Brown, R., & Kulik, J. (1977). Flashbulb Memories. Cognition, 5, 73–99. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 136–162). Thousand Oaks, CA: Sage. Carmines, E. G., & McIver, S. P. (1981). Analyzing models with unobserved variables: Analysis of covariance structures. In G. W. Bohrnstedt & E. F. Borgatta (Eds.), Social measurement: Current issues (pp. 65–115). Beverly Hills, CA: Sage. Christianson, S. A. (1989). Flashbulb memories: Special, but not so special. Memory & Cognition, 17, 433–443. Conway, M. A. (1995). Flashbulb memories. Hove, UK: Lawrence Erlbaum. Conway, M. A. (2002). Testing the flashbulb memory hypothesis. Unpublished manuscript, University of Bristol, UK. Conway, M. A., Anderson, S. J., Larsen, S.F., Donelly, C. M., McDaniel, M. A., McClelland, A. G. R., . . . Logie, R. H. (1994).The formation of flashbulb memories. Memory & Cognition, 22, 326–343. Curci, A. (2005). Latent variable models for the measurement of flashbulb memories: A comparative approach. Applied Cognitive Psychology, 19, 3–22. Curci, A., & Luminet, O. (2006). Follow-up of a cross-national comparison on flashbulb and event memory for the September 11th attacks. Memory, 14, 329–344. Curci, A., & Luminet, O. (2009). Flashbulb memories for expected events: A test of the emotional-integrative model. Applied Cognitive Psychology, 23, 98–114. Curci, A., Luminet, O., Finkenauer, C., & Gisle, L. (2001). Flashbulb memories in social groups: A comparative test-retest study of the memory of French President Mitterand’s death in a French and a Belgian group. Memory, 9, 81–101. Day, M. V., & Ross, M. (2014). Predicting confidence in flashbulb memories. Memory, 22, 232–242. Er, N. (2003). A new Flashbulb memory model applied to the Marmara earthquake. Applied Cognitive Psychology, 17, 503–517. Finkenauer, C., Luminet, O., Gisle, L., El-Ahmadi, A., van der Linden, M., & Philippot, P. (1998). Flashbulb memories and the underlying mechanism of their formation: Toward an emotional-integrative model. Memory & Cognition, 26, 516–531.

Models for formation of FBMs  71

Fivush, R., Bohanek, J. G., Marin, K., & McDermott Sales, J. (2009). Emotional memory and memory for emotions. In O. Luminet & A. Curci (Eds.), Flashbulb memories. New issues and new perspectives (pp. 163–184). Hove: Psychology Press. Frijda, N. H. (1986). The emotions. Cambridge: Cambridge University Press. Frijda, N. H., Kuipers, P., & ter Schure, E. (1989). Relations among emotion, appraisal, and emotional action readiness. Journal of Personality and Social Psychology, 57, 212–228. Hirst, W., & Phelps, E. A. (2016). Flashbulb memories. Current Directions in Psychological Science, 25, 36–41. Hirst,W., Phelps, E. A., Meksin, R.,Vaidya, C. J., Johnson, M. K., Mitchell, K. J., . . . Mather, M. (2015). A ten-year follow-up of a study of memory for the attack of September 11, 2001: Flashbulb memories and memories for flashbulb events. Journal of Experimental Psychology: General, 144, 604–623. Johnson, M. K., & Chalfonte, B. L. (1994). Binding complex memories: The role of reactivation and the hippocampus. In D. L. Schacter & E.Tulving (Eds.), Memory systems 1994 (pp. 311–350). Cambridge: A Bradford Book/The MIT Press. Kizilöz, B. & Tekcan, A. I. (2013). Canonical categories in flashbulb memories. Applied Cognitive Psychology, 27, 352–359. Koppel, J., Brown, A. D., Stone, C. B., Coman, A., & Hirst, W. (2013). Remembering President Barack Obama’s inauguration and the landing of US Airways Flight 1549: A comparison of the predictors of autobiographical and event memory. Memory, 21, 798–806. Kvavilashvili, L., Mirani, J., Schlagman, S., Erskine, J. A. K., & Kornbrot, D. E. (2010). Effects of age on phenomenology and consistency of flashbulb memories of September 11 and a staged control event. Psychology and Aging, 25, 391–404. Lanciano,T., Curci, A., & Soleti, E. (2013). “I knew it would happen . . . And I remember it!”: The flashbulb memory for the death of Pope John Paul II. Europe’s Journal of Psychology, 9, 220–230. Larsen, S. F. (1992). Potential flashbulbs: Memories of ordinary news as the baseline. In E. Winograd and U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb” memories (pp. 32–63). New York: Cambridge University Press. Lazarus, R. S. (1982). Thoughts on the relations between emotion and cognition. American Psychologist, 37, 1019–1024. Leventhal, H. (1984). A perceptual-motor theory of emotion. Advances in Experimental Social Psychology, 17, 117–182. Leventhal, H. & Scherer, K. (1987). The relationship of emotion to cognition: A functional approach to a semantic controversy. Cognition and Emotion, 1, 3–28. Luminet, O. (2009). Models for the formation of flashbulb memories. In O. Luminet & A. Curci (Eds.), Flashbulb memories. New issues and new perspectives (pp. 51–76). Hove, UK: Psychology Press. Luminet, O., & Curci, A. (2009). The 9/11 attacks inside and outside the US: Testing four models of flashbulb memory formation across groups and the specific effects of social identity. Memory, 17, 742–759. Luminet, O., Curci, A., Marsh, E. J., Wessel, I., Constantin, T., Gencoz, F., & Yogo, M. (2004). The cognitive, emotional and social impacts of the September 11th attacks: Group differences in memory for the reception context and the determinants of flashbulb memory. The Journal of General Psychology, 131, 197–224. McCloskey, M., Wible, C. G., & Cohen, N. J. (1988). Is there a special flashbulb memory mechanism? Journal of Experimental Psychology: General, 177, 171–181. Nachson, I., & Zelig, A. (2003). Flashbulb and factual memories: The case of Rabin’s assassination. Applied Cognitive Psychology, 17, 519–531.

72 Luminet

Neisser, U. (1982). Snapshots or benchmarks? In U. Neisser (Ed.), Memory observed: Remembering in natural contexts (pp. 43–48). San Francisco: Freeman. Neisser, U., & Harsch, N. (1992). Phantom flashbulbs: False recollections of hearing the news about Challenger. In E. Winograd, & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb memories” (pp. 9–31). New York: Cambridge University Press. Neisser, U., Winograd, E., Bergman, E. T., Schreiber, C. A., Palmer, S. E., & Weldon, M. S. (1996). Remembering the earthquake: Direct experience vs. hearing the news. Memory, 4, 337–357. Pillemer, D. B. (1984). Flashbulb memories of the assassination attempt on President Reagan. Cognition, 16, 63–80. Pillemer, D. B. (2009). “Hearing the news” versus “being there”: Comparing flashbulb memories and recall of first-hand experiences. In O. Luminet & A. Curci (Eds.), Flashbulb memories. New issues and new perspectives (pp. 125–140). Hove, UK: Psychology Press. Qin, J., Mitchell, K. J., Johnson, M. K., Krystal, J. H., Southwick, S. M., Rasmusson, A. M., & Allen, E. S. (2003). Reactions to and memories for the September 11, 2001 terrorist attacks in adults with posttraumatic stress disorder. Applied Cognitive Psychology, 17, 1081–1097. Rimé, B., Finkenauer, C., Luminet, O., Zech, E., & Philippot, P. (1998). Social sharing of emotion: New evidence and new questions. European Review of Social Psychology, 9, 145–189. Ruiz-Vargas, J. M. (1993). ¿Cómo recuerda usted la noticia del 23-F? Naturaleza y mecanismos de los “recuerdos-destello” [How do you remember the 23-F news? Nature and mechanisms of “flashbulb memories”]. Revista de Psicología Social, 8, 17–32. Scherer, K. R. (1984). On the nature and function of emotion: A component process approach. In K. R. Scherer & P. Ekman (Eds.), Approaches to emotion (pp. 293–318). Hillsdale, NJ: Lawrence Erlbaum. Scherer, K. R. (1988). Criteria for emotion-antecedent appraisal: A review. In V. Hamilton, G. H. Bower, & N. H. Frijda (Eds.), Cognitive perspectives on emotion and motivation (pp. 82–126). Norwell, MA: Kluwer Academic. Schmolck, H., Buffalo, E. A., & Squire, L. R. (2000). Memory distortions develop over time: Recollections of the O. J. Simpson trial verdict after 15 and 32 months. Psychological Science, 11, 39–45. Stone, C. B., Luminet, O., & Takahashi, M. (2015). Remembering public, political events: A cross-cultural and -sectional examination of Australian and Japanese public memories. Applied Cognitive Psychology, 29, 280–290. Talarico, J. M., & Moore, K. M. (2012). Memories of “the rivalry”: Differences in how fans of the winning and losing teams remember the same game. Applied Cognitive Psychology, 26, 746–756. Talarico, J. M., & Rubin, D. C. (2003). Confidence, not consistency, characterizes flashbulb memories. Psychological Science, 14, 455–461. Tinti, C., Schmidt, S., Testa, S., & Levine, L. J. (2014). Distinct processes shape flashbulb and event memories. Memory & Cognition, 42, 539–551. Tulving, E., & Kroll, N. (1995). Novelty assessment in the brain and long-term memory encoding. Psychonomic Bulletin & Review, 2, 387–390. Weaver, C. A. (1993). Do you need a “flash” to form a flashbulb memory? Journal of Experimental Psychology: General, 122, 39–46. Winningham, R. G., Hyman, I. E., & Dinnel, D. L. (2000). Flashbulb memories? The effects of when the initial memory report was obtained. Memory, 8, 209–216. Winograd, E., & Killinger,W. A. (1983). Relating age at encoding in early childhood to adult recall: Development of flashbulb memories. Journal of Experimental Psychology: General, 112, 413–422.

4 ORDINARY MEMORY PROCESSES SHAPE FLASHBULB MEMORIES OF EXTRAORDINARY EVENTS A review of 40 years of research Jennifer M. Talarico and David C. Rubin

We review the shifting definition of “Flashbulb memory” in the 40 years since Brown and Kulik coined the term.We evaluate evidence for veridical, phenomenological, and metacognitive features that have been proposed to differentiate Flashbulb from ordinary autobiographical memories.We further consider the event conditions thought to be necessary to produce Flashbulb memories and discuss how post-event processing may distinguish Flashbulb memories.We conclude that a categorical dissociation between flashbulb and other autobiographical memories is untenable, but that Flashbulb memories still pose important, as yet unanswered, questions.

Brown and Kulik (1977) observed a phenomenon that captured the public’s attention – seemingly indelible memory for important, emotional events. They dubbed it “Flashbulb memory (FBM)” and conducted the first modern empirical study on the topic.The concept was equally effective in capturing the attention of memory researchers, and in the 40 years following their seminal publication, the topic has been investigated almost as often as the events that lead to such memories allow. During this time, the description of the phenomenon has undergone an interesting and important transformation. The initial “special mechanism” hypothesis was that FBMs were a permanent, veridical (though not necessarily complete) memory record that resulted from a unique memorial process involving automatic encoding of all aspects of an important (emotional) event as it happened. However, this strong hypothesis did not last long, as evidence of both errors of omission and commission in the recall of FBMs were soon identified (Christianson, 1989; Neisser & Harsch, 1992; Neisser et al., 1996). Consequently, the revised FBM hypothesis was more agnostic as to why they developed and how they were different from other autobiographical memories (AMs). Our task is to determine if FBMs are simply a subtype of autobiographical remembering or if they are a distinct category of memories. If there are special mechanisms involved in FBMs then there should be three factors that differentiate

74  Talarico and Rubin

them from everyday memories. First, with regards to memory properties, FBMs should be different from ordinary AMs in some way; they could be more accurate or more vivid, for example, than everyday memories.The second way to identify FBMs is that the conditions necessary to produce these memories should be different from ordinary events; for example, they could require strong emotions. Finally, how the individual processes the event (e.g., how one rehearses the memory) should differ for FBMs relative to AMs. For each of these, the claim of a special mechanism requires more than just a difference that could be seen as one extreme of a continuum; there should be some discontinuity between “ordinary” memories and “special” FBMs. In other words, there should be some a priori, objective threshold for defining whether a given memory meets the criteria for being considered an FBM. In order to compare FBMs to ordinary AMs we need to define “ordinary.” Here, we will define ordinary AMs as easy-to-access memories that are brought to mind by a request for a particular kind of memory (e.g., a memory from a particular time, of a particular type of event, or in response to a particular word). The results would likely be different if FBMs were compared to trivial or noteworthy memories, but “trivial” and “noteworthy” require defining along which dimension the events are trivial or noteworthy. Unless noted otherwise, the comparisons we report from the literature are between ordinary AMs (as defined above) and FBMs. We will proceed to review the various ways in which FBMs have been claimed to be different from AMs and the various mechanisms proposed to cause these differences. We will ask if FBMs have more of a given property and report whether there are consistent findings across studies showing that FBMs are more extreme. This is the minimal test of a special mechanism. If this test is met, we will ask if the differences are large enough to exclude a continuum on which FBMs are at one end and where there is little overlap in the distributions. For the proposed mechanisms, we will also ask whether they have been shown to be necessary (i.e., can FBMs exist when these mechanisms are not invoked), and whether they have been shown to be sufficient (i.e., can FBMs occur only when these mechanisms occur). As a summary, our conclusions are indicated in Tables 4.1 and 4.2. Table 4.1 has a list of the ways FBMs have been claimed to be unique.Table 4.2 describes the mechanisms proposed to account for these differences. TABLE 4.1  A summary of the differences between FBM and ordinary AM

FBM > AM Memory characteristics Longevity Accuracy Consistency Vividness Confidence

No X X X

Yes

Discontinuous

X X

No No No No No

Note: Discontinuous implies a large difference with little overlap in the distributions; no characteristics exhibit this. We would have the same results if we replaced “discontinuous” with “as compared to noteworthy memories that were not in response to a flashbulb event.”

FBMs as ordinary memory processes  75 TABLE 4.2 A summary of the evidence supporting the mechanisms proposed for enhancing

FBM relative to ordinary AM Sufficient

Necessary

Only in FBM

Event conditions Consequentiality Distinctiveness Negative affect

No Yes No

No No No

No No No

Memory processes Significance Surprise Emotional intensity Rehearsal

Yes No ? ?

Yes No No ?

No No No No

Characteristics of FBMs Longevity Surprisingly, there is a paucity of evidence in the FBM literature addressing the relative permanence of such memories. Anecdotally, FBMs are extremely long lasting. Empirically, studies that examine longevity typically obtain one retrospective report years after the event and evaluate it for vividness and completeness to determine whether it qualifies as an FBM. For example, using these criteria, between 50% and 99% of participants in any given study have FBMs for learning of historical events after delays of approximately ten years (Brown & Kulik, 1977; Hirst et al., 2015), approximately 30 years (Colegrove, 1899), and approximately 60 years (Berntsen & Thomsen, 2005). Berntsen and Thomsen (2005) asked participants to recall another event from the same time period (i.e., the most positive and most negative personal event during the occupation period) and found that most were able to do so. Contrary to the claims of Brown and Kulik (1977), there are other events that one can remember from equally long ago. Denver, Lane, and Cherry (2010) provide converging evidence for the objective longevity of FBMs without a relative superiority to ordinary AMs. When their participants were provided with a description of the FBM phenomenon and asked to freely recall public events for which they have this type of memory, their recall produced a standard reminiscence bump (Rubin & Schulkind, 1997). Denver, Lane, and Cherry (2010) further showed few reliable differences between older adults’ and younger adults’ FBM despite vast differences in the delay intervals among events recalled by both groups. In conclusion, FBMs are long-lasting, but they are not indelible, nor are they are more permanent than noteworthy everyday memories.

Accuracy Talarico and Rubin (2009) describe in detail the (surprisingly) few studies that investigate verifiable accuracy of FBMs. The two case studies report at least one

76  Talarico and Rubin

critical inaccuracy, though many accurate details (Greenberg, 2004; Neisser, 1982; Thompson & Cowan, 1986). Furthermore, those inaccuracies are systematically biased to enhance the features of the event that contribute to its personal significance. The day of Danish liberation is remembered as more sunny, less cloudy, less windy, less rainy, and/or warmer than it actually was by Berntsen and Thomsen’s (2005) participants, and Neisser (1982, 1986) misidentified a football game as a baseball game (i.e., America’s pastime) when remembering the attack on Pearl Harbor (Thompson & Cowan, 1986).The issue of objective accuracy is ripe for further investigation. Event features most likely to be recalled accurately, as well as the magnitude and direction of errors, should be examined whenever archival data are available to confirm self-reports (see Luminet & Spijkerman, in press, for a similar argument). A larger body of evidence has examined consistency between memory reports as a proxy for accuracy.

Consistency In order to obtain consistency data, two (or more) retrospective reports are ­collected. The report closer in time to the event is considered the standard. Later reports are then compared to that earlier report and inconsistencies are identified. Although two consistent reports are not necessarily accurate, an inconsistent report implies that at least one report is inaccurate. FBM consistency has also been shown to correlate with performance on the autobiographical Implicit Association Test (aIAT), itself a correlate of accuracy (Curci et al., 2014; Lanciano et al., 2013). Contrary to the arguments of Julian, Bohannon, and Aue (2009), we do not consider “wrong time slices” (Neisser & Harsch, 1992) to be accurate. Recalling an event that actually occurred (e.g., a 30th birthday party) but was not the event requested (e.g., “tell me about your 40th birthday party”) is inaccurate recall. Because changes across recalls do not imply a lack of accuracy, investigators may want to emphasize explicit inconsistencies when drawing conclusions about FBMs. The overwhelming evidence is that FBMs are incomplete (Brown & Kulik, 1977) and include inconsistencies (Christianson & Engelberg, 1999; Curci, 2005; Curci & Luminet, 2006; Curci et al., 2001; Greenberg, 2004; Larsen, 1992; Lee & Brown, 2003; Liu, Ying, & Luo, 2012; McCloskey, Wible, & Cohen, 1988; Nachson & Zelig, 2003; Neisser, 1982; Neisser & Harsch, 1992; Schmolck, ­ ­Buffalo, & Squire, 2000; Talarico & Rubin, 2003, 2006; Weaver, 1993; Weaver & Krug, 2004; Wright, 1993). Memory for the general gist of the event is better than memory for specific details (Bohannon & Symons, 1992; Pillemer, 1984; Schmidt, 2004, 2012; Schmidt & Bohannon, 1988; Thompson & Cowan, 1986) and ­peripheral details are more likely to be inconsistent than are central details (Christianson, 1989; Romeu, 2006;Tekcan et al., 2003). However, how one divides responses into central vs. peripheral information often coincides with whether the memory reports show consistency or inconsistency.Tekcan et al. (2003) considered “time” and “others present” to be peripheral details because those two questions were responsible for the majority of inconsistencies found in their participants’

FBMs as ordinary memory processes  77

memory reports. Which categories account for r­ eliable recall is also variable. For example, both Christianson (1989) and P ­ illemer (1984) found reliable recall for “informant,” and “location,” ­information. ­Pillemer (1984), but not Christianson (1989), found “ongoing activity” to be reliably recalled. I­mportantly, in none of these investigations was recall of central vs. p­ eripheral details of a non-FBM obtained. In Curci and Luminet’s (2009) study of French President Mitterrand’s death, French participants were more ­consistent in ­location and time details than were Belgian participants, even though both groups showed high overall ­consistency within these categories. Moreover, inconsistencies in FBMs, once introduced, are repeated over time and not corrected or further altered (as they are for semantic details of the event) (Coluccia, Bianco, & Brandimonte, 2006; Hirst et al., 2015; Tekcan et al., 2003; Weaver & Krug, 2004). Further, delayed recall is highly related to initial recall (Conway et al., 2009; Weaver et al., 2008). Longer delays between the event and the initial memory report often produce enhanced consistency scores (Coluccia, Bianco, & Brandimonte, 2006;Weaver et al., 2008;Winningham, Hyman, & Dinnel, 2000), though not always (Coluccia, Bianco, & Brandimonte, 2010; Kvavilashvili et al., 2009; Lee & Brown, 2003).Within AMs, the total number of consistent details has also been shown to increase with the number of rehearsals (Campbell et al., 2011; Nadel, Campbell, & Ryan, 2007). Lastly, FBMs do not include fewer inconsistencies than everyday memories (Talarico & Rubin, 2003, 2006). Therefore, in conjunction with the evidence described above, we must conclude that FBMs are not permanent, perfect copies of experienced events, nor are they so much more consistent than everyday memories that a special mechanism is required to differentiate the categories.

Vividness However, there are characteristics of FBMs that may still differentiate them from ordinary AMs. Vividness has been of interest to FBM research since Brown and Kulik (1977) described the “live quality that is almost perceptual” (p. 74). Julian et al. (2009) have suggested that elaboration (i.e., quantity of details recalled) may be of interest as well1. There are strong pragmatic advantages to assessing elaboration as it only requires obtaining a single memory report, irrespective of delay from the event. However, there are at least three methodological issues that undermine its utility. First, there is wide variability in both the total number of details and the nature of those details when defining FBM. Kizilöz and Tekcan (2013) asked a large sample of individuals to recall three separate FBMs and identified, without regard to accuracy or consistency, seven distinct categories that captured the majority of information provided in the narratives: informant, location, others present, own ongoing activity, own affect, own thoughts, and aftermath. Interestingly, although time had been asked of participants in more than half of the studies included in their review (as were location, ongoing activity, source, and others present), it was rarely spontaneously mentioned and therefore not included in their group of canonical categories.

78  Talarico and Rubin

More generally, these who, what, where, when, and how-type questions are the foundation of good story-telling and so it is not surprising that they constrain the structure of autobiographical event recall (Neisser, 1982). Second, longer delays between the event and the initial memory report often result in longer memory narratives (Lee & Brown, 2003). In part, this may be due to narrative conventions regarding shared information as Bohannon (1988) found that time information was much more likely to be included in reports obtained after three years than in those recorded after only one week.The frequency and timing of retrievals can also influence total narrative length (Nadel et al., 2007). Third, Marsh and Tversky (2004) have shown that the functional context of retrieval can influence what and how much individuals report. Similarly, reporting medium (e.g., verbal vs. typed) influences both absolute word count and the units of information provided in memory reports (Grysman & Denney, 2016). Therefore, elaboration is a less useful measure than one might expect. Instead, defining vividness as a phenomenological experience of remembering perceptual detail can be informative. Rubin and Kozin (1984) tried to reframe FBMs as “vivid memories”, as they thought that enhanced vividness was the defining feature of the phenomenon. In fact, FBMs often exhibit ceiling effects in vividness ratings regardless of the delay between event and memory report (Kvavilashvili et al., 2010; Kvavilashvili et al., 2003; Niedzwienska, 2003; Talarico & Rubin, 2003; Tinti et al., 2009;Weaver & Krug, 2004).Therefore, FBMs are more vivid than some ordinary memories (see Kvavilashvili et al., 2010 for direct comparison), but other types of AMs, such as those with high emotional intensity, are as vivid as FBMs (Talarico, LaBar, & Rubin, 2004). That is, the distributions of vividness overlap so there is no discontinuity that would require a special mechanism.

Confidence In contrast to objective evidence of memory inaccuracy, participants consistently report enhanced confidence in FBM accuracy. FBMs are usually recalled with a higher degree of confidence than other memories of equal age (Brown & Kulik, 1977; Kvavilashvili et al., 2010; Paradis, Solomon, Florer, & Thompson, 2004; Talarico & Rubin, 2003, 2006; Weaver, 1993; see Denver, Lane, and Cherry, 2010 for equally high confidence ratings in FBMs and ordinary AMs), even when individuals are confronted with evidence that the event in memory could not have occurred as it is remembered (Neisser & Harsch, 1992). Confidence is often at ceiling for FBMs (Christianson & Engelberg, 1999; Curci et al., 2014; Denver, Lane, and Cherry, 2010; Neisser et al., 1996; Niedzwienska, 2003; Talarico & Rubin, 2003, 2006; Weaver, 1993; Weaver & Krug, 2004) and often remains that high for at least months after the event (Christianson & Engelberg, 1999; Coluccia, Bianco, & Brandimonte, 2010; Conway et al., 2009; Denver, Lane, and Cherry, 2010; Hirst et al., 2015; Kraha & Boals, 2014; Kvavilashvili et al., 2009; Liu, Ying, & Luo, 2012; Niedzwienska, 2003; Weaver & Krug, 2004). It may be that confidence ratings

FBMs as ordinary memory processes  79

are based on equally reliably enhanced vividness ratings, as the two are correlated (Neisser & Harsch, 1992). Therefore, along with vividness, the second distinctive property of FBMs is a discrepancy between metacognitive perception and objective reality. In fact, this discrepancy may have led to the identification of the phenomenon in the first place and may well lead to the most interesting applications of the phenomenon to ordinary memory processing. Thus, it is the secondary, phenomenological characteristics like vividness and confidence that may serve to retain the utility of FBM as a distinct category. If FBMs are differentiated by phenomenological experience, then the mechanisms responsible for the phenomenon must account for these differences, not explain (non-existent) encoding or retrieval differences. The primary question in FBM research then becomes why we are more likely to maintain vivid, confidently held memories of these particular events. Enhancements to the subjective experience of remembering support Berntsen’s (2009) model of FBM formation being driven by social group identification and a subsequent feedback loop where the memory itself serves to perpetuate identification with the social group. This may be considered a special case of Fernández’s (2015) epistemic function of AMs more generally (e.g., an individual benefits from believing herself to be a patriot by relying on her vivid memory for learning about the September 11th attacks as evidence for that belief). However, we are still left to identify which event features are necessary to produce the defining memory characteristics.

Event conditions necessary to produce FBM The vast majority of research in this field has been done in the aftermath of a public tragedy.This has been because consequentiality, distinctiveness, and emotional affect have been the primary features of the event thought to influence the formation of FBM (i.e., a memory report which satisfies the criteria described above for vividness and confidence, longevity and/or consistency). Here, we will discuss objective characteristics of the event thought to produce FBMs. Subjective assessments of the events will be discussed later.

Consequentiality Consequential events most often studied include disasters with loss of life (e.g., earthquakes, terrorist attacks) or events with political implications (e.g., assassinations, resignations, invasions). FBM research is differentiated from traumatic memory research as the participants in the latter are directly affected by the events being studied. In the case of FBMs, participants are rarely so personally involved (see Pillemer, 2009 for a review of this distinction). However, the events being investigated are often on such a scale that the aftermath affects the lives of participants in other, more subtle ways.

80  Talarico and Rubin

Is this comprehensive consequentiality responsible for FBMs? In short, no. The best evidence for the importance of consequentiality is indirect – an association between physical proximity to the event location and FBM formation. Crossnational studies have found this effect (Conway et al., 1994; Curci et al., 2001; Curci & Luminet, 2006; Kvavilashvili et al., 2003) as have national studies with participants sampled from multiple locations (Er, 2003; Neisser et al., 1996; Sharot et al., 2007), although Luminet et al. (2004) found few differences in FBM specificity by nationality, and Pezdek (2003) found that those closest to the event were less likely to recall their personal circumstances than were those living far away. For consistency, some studies have found differences as a result of proximity (Conway et al., 1994; Er, 2003) whereas others have not (Curci et al., 2001; Curci & Luminet, 2006). Enhanced vividness, however, does seem to be associated with being physically closer to the event (Kvavilashvili et al., 2003; Sharot et al., 2007). Many of the studies examining distance conflate directly experiencing the event with closest physical proximity. In cases involving political events (Conway et al., 1994; Curci et al., 2001), simply living in the affected nation ought to be less consequential than experiencing a natural disaster (Er, 2003; Neisser et al., 1996) or terrorist attack (Sharot et al., 2007), for example. The bulk of the empirical evidence fails to support the claim that objective consequentiality is relevant for the formation (Er, 2003; Tekcan, 2001), accuracy (Berntsen & Thomsen, 2005), consistency (Niedzwienska, 2003; Weaver, 1993), or vividness (Berntsen & Thomsen, 2005; Rubin & Kozin, 1984) of FBMs. However, Koppel, Brown, Stone, Coman, and Hirst (2013) showed that different factors predicted consistency for consequential (i.e., the first inauguration of U.S. President Obama) and nonconsequential (i.e., the emergency landing of US Airways Flight 1549) public events; emotional intensity and significance predicted consistency of memories for a consequential event, whereas rehearsal predicted consistency for a nonconsequential event. So, although (inter)national events may retain value by generating large numbers of potential participants, exclusively examining such events simply because they are assumed to have a requisite degree of consequentiality seems unnecessary.

Distinctiveness The evidence in support of distinctiveness effects is much stronger than was found for consequentiality, as it has been correlated with the formation (EderyHalpern & Nachson, 2004; Larsen, 1992; Wright & Gaskell, 1992) and vividness (Bohn & Berntsen, 2007; Edery-Halpern & Nachson, 2004) of FBMs. Mahmood, Manier, and Hirst (2004) found no relationship between distinctiveness and the formation or vividness of FBMs, but, distinctiveness in their study was defined as the first event in a series of similar, emotional, personally significant events (i.e., the deaths of lovers, friends, and/or family members due to AIDS). An event may be distinctive for reasons other than that it is the first of its kind, however; Brown and Kulik (1977) studied memory for a series of assassinations of political figures in a

FBMs as ordinary memory processes  81

relatively brief period of time, yet each was a distinctive event. Edery-Halpern and Nachson (2004) found that the least distinctive event in their sample of terrorist attacks within a two-year period in Israel was also significantly less well remembered (i.e., fewer details were recalled and more responses were left blank for this memory compared to the other events). This is not surprising given that the episodic memory literature includes ample evidence of a distinctiveness advantage (i.e., von Restorff effect; see Schmidt, 2012 for a review within an FBM context). Brewer’s (1988) study of AM suggests that the lower the frequency of event occurrence, the greater the likelihood of later cued recall. Thus, any effects of distinctiveness that might be present may be similar for FBM and AM.

Emotional Affect Another event feature known to enhance ordinary memory and thought to influence FBMs is negative emotional affect. For example, negative stimuli “pop-out” in a neutral context to a greater extent than neutral stimuli in a fearful context (Ohman, Flykt, & Esteves, 2001). As they typically involve disasters, attacks, and assassinations, most FBM studies have included only negative events. Kraha, Talarico, and Boals (2014) examined the surprising announcement of Osama bin Laden’s death, which was interpreted positively by the U.S. students sampled, and found little evidence of FBM. However, this lack of phenomenological enhancement was attributed more to lack of emotional intensity (see below) than to positive affect per se. Other studies have shown both positive and negative events can lead to FBMs (Berntsen & Thomsen, 2005; Bohn & Berntsen, 2007; Demiray & Freund, 2015; Liu,Ying, & Luo, 2012; Scott & Ponsoda, 1996; Tekcan, 2001). Positive interpretations of an event are more likely to produce FBMs (Bohn & Bernsten, 2007) and increase accuracy, vividness, and rehearsal of that event relative to the memories of those who interpret the event negatively (Breslin & Safer, 2011; Talarico & Moore, 2012). Kensinger and Schacter (2006) and Holland and Kensinger (2012) both found that consistency was higher (although confidence was lower) for positive interpretations of a sporting event and an election outcome, respectively. (However, in their examination of a sporting event, Talarico and Moore [2012] found no differences in consistency or confidence among fans of the winning and losing teams.) Generally, these results are consistent with pleasantness biases in autobiographical recall (see Walker, Skowronski, & Thompson, 2003 for a review). Furthermore, there is evidence from collective memories that even profoundly negative events are more likely to persist in the culture if they evoke positive connotations. For example, Hirst and Meksin (2009) describe how the assassinations of Lincoln and Kennedy endure because each President was subsequently deified by popular culture. Similarly, there is often an emphasis on patriotism and heroism in the face of tragedy (e.g., the Pearl Harbor or September 11th attacks) in societal recollections of those events (see Stone & Jay, Chapter 8, this volume, for further discussion of FBM for positive events).

82  Talarico and Rubin

Summarizing the conditions necessary to produce FBMs, what could be a unique characteristic of FBMs (consequentiality) fails to predict the memory phenomenon, and well-characterized features of AM moderately account for the key characteristics of vividness and confidence, suggesting that subjective interpretations of events may be more responsible for the phenomena.

Individual processing of the subjective flashbulb event Characteristics of how an individual processes the event at encoding and during rehearsal/retrieval are thought to be important determinants of FBMs. Encoding factors are closely tied to event features: distinctiveness with surprise, emotional affect with emotional intensity, and consequentiality with significance. What differentiates them is the objective vs. subjective nature of assessment. How these characteristics contribute to the FBM phenomenon individually and interactively is the focus of much current work in this area.

Significance Significance, or personal importance, refers to the individual’s subjective assessment of the event. This assessment can be based on any number of felt criteria and is not necessarily related to any material changes in the individuals’ circumstances. As with consequentiality and emotion, group membership is frequently used as a proxy for significance.This provides a methodological advantage in identifying large numbers of participants for whom individual reactions to a common event are expected to vary. There are also theoretical reasons to justify this technique. Berntsen (2009) argues quite persuasively that it is an event’s importance to our social identity specifically that determines whether an event will produce an FBM. Because FBM research has emphasized recall of public events, it is not surprising that social identity is the most salient criteria for determining significance. Various traits have been used to demarcate social groups, including race (Brown & Kulik, 1977) gender, (Wright, Gaskell, & O’Muircheartaigh, 1998), language (Stone et al., 2013), religion (Curci et al., 2014; Lanciano, Curci, & Soleti, 2013; Tinti et al., 2009), and explicit membership in social movements (e.g., participation in resistance activity during military occupation by foreign forces (Berntsen & Thomsen, 2005). Across all operational definitions, social group membership (and therefore personal significance) was positively associated with FBM formation (Brown & Kulik, 1977; Wright et al., 1998), enhanced accuracy (Berntsen & Thomsen, 2005), greater consistency (Curci et al., 2014; Lanciano et al., 2013; Tinti et al., 2009), increased elaboration (Curci et al., 2014; Stone et al., 2013), and enhanced vividness (Berntsen  & Thomsen. 2005; Tinti et al., 2009), though it was unrelated to confidence (Curci et al., 2014). Lanciano et al. (2013) further showed that ratings of importance (a conflation of event consequentiality and personal significance) for the death of Pope John Paul II differed as a function of religious affiliation, consistent with Berntsen’s (2009)

FBMs as ordinary memory processes  83

social identity model. Similarly, Tinti et al. (2009) used national origin as an alternate avenue to identification with the Pontiff, allowing it to serve as a proxy for significance. After controlling for religiosity, they found that participants with no particular affiliation to Pope John Paul II (i.e., Swiss individuals) had less consistent and less vivid FBMs for his death than did participants who could identify with him (i.e., Italian and Polish individuals). Luminet et al. (2004) showed that U.S. versus non-U.S. citizenship status was more predictive of developing an FBM for the September 11th attacks than was nation of residence. In Stone et al.’s (2013) examination of linguistic groups within Belgium, both French and Dutch speakers thought the division of the University of Leuven was equally important, but Dutch speakers were more likely to be personally and/or politically involved in the division and, as a result, had more elaborate FBMs and rehearsed those memories more frequently than did French speakers. Not all FBM studies rely on group membership to examine this characteristic; individual ratings of personal importance have also been investigated (when sufficient variability exists to do so). Participant ratings of significance are positively correlated with FBM formation (Bohannon & Symons, 1992; Conway et al., 1994; Larsen, 1992; Mahmood, Manier, & Hirst, 2004; Wright & Gaskell, 1992; but see Wright et al., 1998). Paradis et al. (2004) found that their New York City participants rated both September 11th and 12th as personally important and their sample developed FBMs for both of those days, in terms of initial recall and later consistency. Niedzwienska (2003) also found significance to be correlated with consistency of the FBM report. Furthermore, personal significance has been reliably related to vividness ratings (Mahmood, Manier, & Hirst, 2004; Nachson & Zelig, 2003; Niedzwienska, 2003; Rubin & Kozin, 1984) and to the number of details reported (i.e., elaboration) in German,Turkish, British, and American samples, though not within a Chinese sample (Kulkofsky et al., 2011). The authors attributed this anomaly to collectivist cultural expectations to de-emphasize personal goals and activities that therefore potentially dampened significance ratings within this group. However, individual ratings of personal significance are not universally found to predict FBM phenomena. Otani et al. (2005) classified memory reports as FBMs or non-FBMs, yet found no difference in the significance ratings of participants in each group. Davidson and Glisky (2002) also found no differences in the significance ratings of two events, yet one event led to reliably more FBM reports than the other. As the study of irrelevant characteristics (e.g., consequentiality, negative valence) declines and is replaced by more systematic study of relevant characteristics, the nuanced nature of such effects can be determined. It is our belief that personal significance as assessed by social group membership will remain a determining feature of FBM formation, but that other criteria for personal significance may also lead to vivid, confidently held FBMs as well. As with distinctiveness, there is an abundance of data for a self-referential effect in memory performance, with personally relevant material enhancing memory (see Symons & Johnson, 1997 for a review).Thus, the influence of significance on FBMs can be predicted from more general features of AMs.

84  Talarico and Rubin

Surprise As significance is differentiated from consequentiality, so, too, is surprise different from distinctiveness. Surprise is a personal, emotional reaction to the event, not a property of the event. Note that although an event can be expected, and therefore not surprising, it can still be distinctive, as was seen in the case of several terrorist attacks in Israel, the sad inevitability of which does not prevent each attack from being distinct (Edery-Halpern & Nachson, 2004). Although surprise is a key component in many models of FBM formation (Brown & Kulik, 1977; Er, 2003; Finkenauer et al., 1998), there is little data to support its inclusion. Only when comparing FBM for two similar events (the deaths of Princess Diana and Mother Theresa), were higher ratings of surprise associated with greater initial recall and later consistency of FBM. Surprise only seems to be directly influential when tied to social identity, otherwise it plays no direct role in FBM formation (Curci & Luminet, 2009). FBMs have been found for expected events (Curci et al., 2001; Lanciano et al., 2013; Tekcan, 2001; Tinti et al., 2009; Winograd & Killinger, 1983). Additionally, Coluccia, Bianco, & Brandimonte (2010) found no differences between expected and unexpected events in the relationships between delay and consistency, confidence, or elaboration. Equal surprise ratings were provided by those who did and those who did not develop FBMs for the Kobe earthquake (Otani et al., 2005). There is even evidence that surprise and FBM phenomena are negatively correlated. Berntsen and Thomsen (2005) found that participants rated the invasion of Denmark as more surprising than its liberation, but were more likely to have FBMs for the liberation than for the invasion. It may be that surprise is a retrospective evaluation that is based on the emotional intensity of one’s reaction. If this is the case, we would expect surprise to play a role in FBM formation when it is present, but it would be neither a necessary nor a sufficient condition for the phenomenon. Interestingly, the effect of surprise in AM has not been thoroughly investigated. We do know that surprise tends to be a positive emotion in ordinary AMs (Talarico, LaBar, & Rubin, 2004). In other words, when cued to generate memories of surprise, participants are more likely to recall pleasant events (e.g., a surprise birthday party) than unpleasant events (e.g., an unexpectedly low score on an exam). Yet, when evaluating positive and negative events generally, positive events tend to be expected (e.g., weddings) whereas negative events are typically unexpected (e.g., divorce) (Berntsen, 2002; Rubin & Berntsen, 2003). The public events studied in FBM research are predominantly surprising and negative, therefore the scope of FBM research has been limited by what may be coincidental, not causal, event features. Anticipatable public events (e.g., elections) may be ripe for additional exploitation by FBM investigators (e.g., Boals, 2010; Holland & Kensinger, 2012).

Emotional Intensity Some of the most contradictory findings in the FBM literature are those involving emotional intensity2. The positive-negative valence dimension depends primarily

FBMs as ordinary memory processes  85

on the nature of the event itself (though some events may be interpreted differently among different groups). Depth of feeling, however, is necessarily a subjective reaction to a given event within a valence category. There are data supporting emotional intensity’s role in FBM formation using participant ratings (Berntsen & Thomsen, 2005; Bohannon, 1988; Bohannon & Symons, 1992; Davidson & Glisky, 2002; Paradis et al., 2004) and using culture as a proxy for emotion (Brown & Kulik, 1977; Curci et al., 2001), but nearly as many studies that fail to find a correlation with participant ratings (Otani et al., 2005; Smith, Bibi, & Sheard, 2003; Tekcan, 2001), or with culture as a proxy (Luminet et al., 2004).Wright et al. (1998) found that men, who rated the Hillsborough football disaster as less emotional than women, were more likely to develop FBM, not less. The same contradictory pattern emerges for intensity and consistency, with some studies finding a positive relationship between the two (Bohannon & Symons, 1992; Conway et al., 1994; Davidson & Glisky, 2002; Schmolck, Buffalo, & Squire, 2000), but more that fail to find such a relationship (Christianson & Engelberg, 1999; Kvavilashvili et al., 2009; Nachson & Zelig, 2003; Neisser et al., 1996; Neisser & Harsch, 1992; Schmidt, 2004; Talarico & Rubin, 2003).Vividness of the FBM is equally divergent. Rubin and Kozin (1984) failed to find a correlation between emotional intensity and vividness, whereas others have found the two to be related (Berntsen & Thomsen, 2005; Nachson & Zelig, 2003; Pillemer, 1984). Lastly, Hirst et al. (2015) failed to find a relationship between emotional intensity and confidence at either three-year or ten-year intervals. Even if the effects of emotional intensity were more reliable, the advantage of emotionally intense experiences over neutral events is well established in the episodic and AM literatures and so it cannot be seen as unique to FBM.

Rehearsal The final processing feature we will examine is the only non-encoding-specific mechanism discussed in this literature. The effects of rehearsal on FBM seem to dissociate based on the dependent variable of interest. Increased rehearsal has been correlated with the formation of FBMs (Bohannon, 1988; Bohannon & Symons, 1992; Curci et al., 2001; Davidson & Glisky, 2002; Otani et al., 2005; Tekcan & Peynircioglu, 2002; Tinti et al., 2014), although Hornstein, Brown, and Mulligan (2003) found that to be true only for covert, not overt, rehearsal. For vividness, there seems to be no relationship with rehearsal (Pillemer, 1984; Rubin & Kozin, 1984). Hirst et al. (2015) found that confidence was correlated with rehearsal after a three-year delay (as did Kvavilashvili et al., 2009), but not after a ten-year delay. For consistency, the pattern is quite variable. There are studies showing a positive correlation between rehearsal and consistency (Bohannon & Symons, 1992; Conway et al., 2009; Davidson & Glisky, 2002; Schmolck, Buffalo, & Squire, 2000), a negative correlation between the two (Kvavilashvili et al., 2009), and no relationship (Kvavilashvili et al., 2009; Pillemer, 1984; Schmolck, Buffalo, & Squire, 2000),

86  Talarico and Rubin

across delays ranging from six months to three years. There is no systematic relationship among delay, type of rehearsal (overt vs. covert), or consistency. Moreover, rehearsal via media exposure can increase errors of commission as is seen in the “crashing memories” phenomenon (Crombag, Wagenaar, & van Koppen, 1996) of reporting false memories of non-existent footage (Ost et al., 2008). Media exposure can also serve to increase omissions via retrievalinduced forgetting (Coman, Manier, & Hirst, 2009). These details are unlikely to be reintroduced into the memory narrative because there are few opportunities for correction. As with most AMs, there is little evidence of what actually occurred, therefore little can be presented to refute or corroborate one’s personal recollection. Yet, we know that rehearsal is a potent mechanism for sustaining memory. Repeated overt retrieval attempts lead to more consistent and more elaborated AM reports (Campbell et al., 2011; Nadel et al., 2007). In other words, individuals add details to their memory reports while also maintaining the originally provided information. Kvavilashvili et al. (2010) found a positive correlation between rehearsal and consistency for a staged autobiographical event in contrast to a negative correlation for FBM of the September 11th attacks. Specifying the differences between public and private, overt and covert rehearsal behaviors may help make sense of these disparate results.

Combining factors Most events are chosen as subjects of FBM research because they exhibit many of the features we have noted here. This has led some investigators to adopt statistical techniques such as latent variable modeling and structural equation modeling to determine the relationships among these features (see Luminet, Chapter 3, this volume) frequently within FBM and non-FBM of the same event. Each of these models defined FBM as recall of some number of canonical categories or some measure of completeness and specificity, not vividness or confidence, the more reliable characteristics to differentiate FBMs from ordinary AMs. Further, these models consider significance and distinctiveness to be indirect factors in FBM formation, though the data presented here suggest that they are among the predominant mechanisms responsible for FBM. Day and Ross (2014) are the only investigators to model confidence, one of the more reliable FBM characteristics. They found that the strongest predictor of confidence when remembering Michael Jackson’s death was a sense of attachment to the performer (which lead to greater initial surprise, increased emotional intensity, and enhanced rehearsal). Furthermore, they found that individuals’ beliefs about the persistence of FBM also correlated with their later confidence in their own memories, but not with the consistency of those memory reports. Models like this suggest a high degree of interrelatedness among metacognitive features (i.e., beliefs about memory generally and beliefs about one’s own memory) and between the ways in which events are processed and their resulting phenomenology. In contrast,

FBMs as ordinary memory processes  87

there seems to be little predictive value of event features on resulting mnemonic characteristics.

Summary and future directions FBMs are distinguished from ordinary memories by their vividness and the confidence with which they are held. There is little evidence that they are reliably different from ordinary AMs in longevity, accuracy, or consistency. Features of the event and processing characteristics of the individual identifiable in ordinary AM explain enhancements to each of these memory properties. Curci and Lanciano (2009), Lanciano and Curci, (2012; see also chapter 1, this volume) have argued that FBMs are better viewed as a category separate from ordinary memories rather than along a continuum with them. To support this claim they conducted two studies, each with a large number of participants who were asked to recall details of a single public event on two occasions. The consistency of their responses to canonical FBM questions were rated and analyzed statistically. In both studies, the authors found a better fit of the empirical data to a categorical (taxonomic) model than to a continuous (dimensional) model of memory. Further, guidance for interpreting the claim of distinct categories is based on a limited theory of episodic/autobiographical memory. Specifically, Conway (1995) has argued that FBMs are stable clusters of sensory-perceptual details that are highly integrated and therefore accessed holistically. They are contrasted with ordinary AMs which are dynamic reconstructions of information drawn from event-specific knowledge and generalized autobiographical themes. Brewin (2014) reinforces this idea when suggesting that FBMs are defined by “a more detailed perceptual record of experience than ordinary memories” (p. 73). He goes on to argue that FBMs are evidence of a distinct perceptual memory system, separate from episodic memory, and particularly linked with emotional reactivity. It is this specific, empirically unsupported difference between FBM and AM that is discussed in both papers, and thus is not an ideal assumption for testing the nature of categories. A more common and better supported view of AM organization is based on a more flexible view of the construction of autobiographical memories at all levels (e.g., Bartlett, 1932; Hirst & Phelps, 2016; Neisser, 1982; Rubin, 2012; Rubin & Umanath, 2015).This view allows for more flexible use of non-hierarchical knowledge bases and can adequately account for the findings reviewed in our chapter for both FBM and ordinary AM. For the event conditions, consequentiality seems to be irrelevant to FBM, distinctiveness is the most predictive, and emotional affect is as yet understudied. Of the processes discussed, significance is the most promising determinant of FBM, especially as related to social group membership in the context of public events. It is correlated with formation, consistency, vividness, and confidence of FBMs. Surprise has an unreliable influence on the formation and consistency of FBMs (the least reliable features of FBMs overall) and its effects on vividness and confidence (the most reliable features) have not been systematically investigated. Therefore, this seems to be one of the more promising areas for future

88  Talarico and Rubin

investigation. Emotional intensity is an unreliable predictor of FBM, yet this is most likely due to inconsistencies in defining FBM. At least some of the enhanced confidence and/or vividness associated with FBMs are probably due to enhanced emotionality; however, the exact nature and scope of that influence has yet to be determined. Lastly, rehearsal tends to be correlated with the formation of FBMs, but the relationships between rehearsal and consistency, vividness, and confidence are quite variable. Because rehearsals can take multiple forms (public vs. private, overt vs. covert) and can simultaneously enhance recall of some details while introducing erroneous details, disentangling the specific effects of rehearsal on particular features of FBM is an important goal for future investigators. We have no doubt that FBM will remain a topic of frequent investigation. After 40 years of wandering, we may be finally approaching the promised land of understanding this fascinating and complex phenomenon. Berntsen’s (2009) resurrection and expansion of the social aspects of Brown and Kulik’s (1977) hypothesis may be the most significant contribution to the FBM literature in recent years, in part because it did not rely on any claims to special mechanisms; these factors are important to all AMs. By reclaiming the public nature of events from mere methodological convenience to an important theoretical feature, Berntsen’s (2009) model helps to sustain FBM as a distinctive phenomenon and as a fruitful area for inter- and multi-disciplinary work on remembering. Listening to news is not an inherently memorable activity (Larsen, 1992). The counterintuitive enhancement of memory for otherwise mundane activities was what spurred initial interest in FBM. Neisser’s (1982) suggestion that “we remember the details of a flashbulb occasion because those details are the links between our own histories and History” (p. 48) was prescient, but perhaps too broad. Evidence of FBM phenomena from events that are public, but on a relatively smaller scale than is typical of natural disasters or political events, are proof of this. For example, memories of sporting events for fans of a given team underscore the role of personal significance (and the irrelevance of objective consequentiality) in Flashbulb memory. These memories are likely only among those who consider the event to be important, and therefore emotionally intense. Groups of fans observing subsequent games serve as a natural context for rehearsal and sharing one’s memory serves as a potent symbol of one’s identity as a “true” fan. Rehearsing one’s personal relationship to an event rather than the factual details of the event itself is consistent with the role of the feedback loop in creating and maintaining FBMs. Telling one’s personal story of learning about a shared event serves narrative conventions to introduce novel details to a conversation and serves larger social functions by building group cohesion.The expectation of remembering, of having that story to share, is an underappreciated characteristic of the FBM phenomenon. Both individuals and groups have that expectation. There is presumed judgment of one’s failure to remember a given event and what that failure implies about the individual’s understanding or interpretation of a given event. One important direction for future research ought to be moving away from indirect assessment of social identity (e.g., cross-cultural studies) and toward a

FBMs as ordinary memory processes  89

priori measures of group identification and selection of samples who vary significantly on those measures. One should predict that FBM phenomenology is likely to be greatest when membership in a social group is important (not just incidental) to an individual and when that membership is vulnerable. Because every individual is simultaneously a member of multiple social groups, each of those groups is not of equal importance. Feelings of insecurity about one’s group membership should enhance the perceived significance of both the event itself and of the FBM for its symbolic function. The pressure to prove one’s status should correlate with phenomenological features of the memory such as vividness and confidence in its accuracy. For groups wherein one’s status is more established or more assured, the FBM phenomenology may still be at the higher end of a continuum within AM, but they may not be at the extreme ceiling seen for other events. In his book, Extraordinary memories for exceptional events, Schmidt (2012) defines FBM methodologically (i.e., studies asking individuals “how did you first learn the news of public event X”), phenomenologically (i.e., as a detailed, long-lasting, vivid, and confidently-held AM), and theoretically (i.e., accurate and detailed memories that are the result of a special remembering mechanism). We argue here that the theoretical definition is unsustainable. There is insufficient evidence to support a categorical dissociation between FBMs and other AMs. The phenomenological definition is nearing refinement but there are still some unanswered questions. We believe this to be the most fruitful and therefore most important direction for FBM research in the coming years. Most importantly, we believe that a clear, a priori definition of what constitutes an FBM or, better, a set of alternative definitions that can be contrasted to see which provides a better understanding of the data, is necessary in future empirical work. Clearly stating what data are required to identify a given memory as an FBM or a non-FBM will increase the utility of new studies in addressing these important questions. Without such definitions, future studies risk simply repeating past studies on new events and merely describing the phenomenon rather than explaining it. We have also argued that the methodological definition may be more than a simple convenience.The personal memory of learning about a public event remains a unique domain within AM. This characteristic ties event features to memory features and therefore provides clues to potential mechanism(s) for the phenomenon. It also captures why the phenomenon is of such enduring interest. Long-lasting, detailed, vivid, confidently-held memories of directly-experienced, emotional, significant, and well-rehearsed events is not unexpected, but the idea that simply receiving news can change a mundane experience into a noteworthy memory remains surprising and is why FBMs deserve to remain a distinct phenomenon.

Notes 1 The stronger claim by Julian et al. (2009) that elaboration may be used as a proxy for accuracy is supported by some evidence within semantic memory for event details, not within AM.

90  Talarico and Rubin

2 Note that most of the AM literature relies on self-reports of emotional intensity and are therefore best characterized as a measure of depth of feeling. Physiological arousal is a related, but independent, construct not addressed here.

References Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology. New York: Cambridge University Press. Berntsen, D. (2002). Tunnel memories for autobiographical events: Central details are remembered more frequently from shocking than from happy experiences. Memory & Cognition, 30, 1010–1020. Berntsen, D. (2009). Flashbulb memory and social identity. In O. Luminet and A. Curci (Eds.), Flashbulb memories: New issues and new perspectives (pp. 187–205). New York: Psychology Press. Berntsen, D., & Thomsen, D. K. (2005). Personal memories for remote historical events: Accuracy and clarity of flashbulb memories related to World War II. Journal of Experimental Psychology: General, 134, 242–257. Boals, A. (2010). Events that have become central to identity: Gender differences in the centrality of events scale for positive and negative events. Applied Cognitive Psychology, 24, 107–121. Bohannon, J. N. III. (1988). Flashbulb memories for the space shuttle disaster: A tale of two theories. Cognition, 29, 179–196. Bohannon, J. N. III., & Symons, V. L. (1992). Flashbulb memories: Confidence, consistency, and quantity. In E. Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb” memories (pp. 65–91). New York: Cambridge University Press. Bohn, A., & Berntsen, D. (2007). Pleasantness bias in flashbulb memories: Positive and negative flashbulb memories of the fall of the Berlin Wall among East and West Germans. Memory & Cognition, 35, 565–577. Breslin, C. W. & Safer, M. A. (2011). Effects of event valence on long-term memory for two baseball championships games. Psychological Science, 22, 1408–1412. Brewer, W. F. (1988). Memory for randomly sampled autobiographical events. In U. Neisser & E.Winograd (Eds.), Remembering reconsidered: Ecological and traditional approaches to the study of memory (pp. 21–90). New York: Cambridge University Press. Brewin, C. R. (2014). Episodic memory, perceptual memory, and their interaction: Foundations for a theory of posttraumatic stress disorder. Psychological Bulletin, 140, 69–97. Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. Campbell, J., Nadel, L., Duke, D., & Ryan, L. (2011). Remembering all that and then some: Recollection of autobiographical memories after a 1-year delay. Memory, 19, 406–415. Christianson, S. (1989). Flashbulb memories: Special, but not so special. Memory & Cognition, 17, 435–443. Christianson, S., & Engelberg, E. (1999). Memory and emotional consistency: The MS Estonia ferry disaster. Memory, 7, 471–482. Colegrove, F. W. (1899). Individual memories. American Journal of Psychology, 10, 228–255. Coluccia, E., Bianco, C., & Brandimonte, M. A. (2006). Dissociating veridicality, consistency, and confidence in autobiographical and event memories for the Columbia shuttle disaster. Memory, 14, 452–470. Coluccia, E., Bianco, C., & Brandimonte, M. A. (2010). Autobiographical and event memories for surprising and unsurprising events. Applied Cognitive Psychology, 24, 177–199.

FBMs as ordinary memory processes  91

Coman, A., Manier, D., & Hirst, W. (2009). Forgetting the unforgettable through conversation: Socially shared retrieval-induced forgetting of September 11 memories. Psychological Science, 20, 627–633. Conway, A. R. A., Skitka, L. J., Hemmerich, J. A., & Kershaw, T. C. (2009). Flashbulb memories for 11 September 2001. Applied Cognitive Psychology, 23, 605–623. Conway, M. A. (1995). Flashbulb memories. Hove, UK: Lawrence Erlbaum. Conway, M. A., Anderson, S. J., Larsen, S. F., Donnelly, C. M., McDaniel, M. A., McClelland, A. G. R., . . . Logie, R. H. (1994).The formation of flashbulb memories. Memory & Cognition, 22, 326–343. Crombag, H. F. M., Wagenaar, W. A., & van Koppen, P. J. (1996). Crashing memories and the problem of ‘source monitoring.’ Applied Cognitive Psychology, 10, 95–104. Curci, A. (2005). Latent variable models for the measurement of flashbulb memories: A comparative approach. Applied Cognitive Psychology, 19, 3–22. Curci, A., & Lanciano, T. (2009). Features of autobiographical memory: Theoretical and empirical issues in the measurement of flashbulb memory. The Journal of General Psychology, 136, 129–152. Curci, A., Lanciano,T., Maddalena, C., Mastandrea, S., & Sartori, G. (2015). Flashbulb memories of the Pope’s resignation: Explicit and implicit measures across different religious groups. Memory, 23, 529–544. Curci, A., & Luminet, O. (2006). Follow-up of a cross-national comparison on flashbulb and event memory for the September 11th attacks. Memory, 14, 329–344. Curci, A., & Luminet, O. (2009). Flashbulb memories for expected events: A test of the Emotional-Integrative Model. Applied Cognitive Psychology, 23, 98–114. Curci, A., Luminet, O., Finkenauer, C., & Gisle, L. (2001). Flashbulb memories in social groups: A comparative test-retest study of the memory of French President Mitterrand’s death in a French and a Belgian group. Memory, 9, 81–101. Davidson, P. S. R., & Glisky, E. L. (2002). Is flashbulb memory a special instance of source memory? Evidence from older adults. Memory, 10, 99–111. Day, M. V., & Ross, M. (2014). Predicting confidence in flashbulb memories. Memory, 22, 232–242. Demiray, B., & Freund, A. M. (2015). Michael Jackson, Bin Laden and I: Functions of positive and negative, public and private flashbulb memories. Memory, 23, 487–506. Denver, J.Y., Lane, S. M., & Cherry, K. E. (2010). Recent versus remote: Flashbulb memory for 9/11 and self-selected events from the reminiscence bump. The International Journal of Aging and Human Development, 70, 275–297. Edery-Halpern, G., & Nachson, I. (2004). Distinctiveness in flashbulb memory: Comparative analysis of five terrorist attacks. Memory, 12, 147–157. Er, N. (2003). A new flashbulb memory model applied to the Marmara earthquake. Applied Cognitive Psychology, 17, 503–517. Fernández, J. (2015). What are the benefits of memory distortion? Consciousness and Cognition, 33, 536–547. Finkenauer, C., Luminet, O., Gisle, L., El-Ahmadi, A., van der Linden, M., & Philippot, P. (1998). Flashbulb memories and the underlying mechanisms of their formation: Toward an emotional-integrative model. Memory & Cognition, 26, 516–531. Greenberg, D. L. (2004). President Bush’s false ‘flashbulb’ memory of 9/11/01. Applied Cognitive Psychology, 18, 363–370. Grysman, A., & Denney, A. (2016). Gender, experimenter gender and medium of report influence the content of autobiographical memory report. Memory, 25, 132–145.

92  Talarico and Rubin

Hirst, W., & Meksin, R. (2009). A social-interactional approach to the retention of collective memories of flashbulb events. In O. Luminet and A. Curci (Eds.), Flashbulb memories: new issues and new perspectives (pp. 207–225). New York: Psychology Press. Hirst, W., Phelps, E. A., Meksin, R.,Vaidya, C. J., Johnson, M. K., Mitchell, K. J., . . . Olsson, A. (2015). A ten-year follow-up of a study of memory for the attack of September 11, 2001: Flashbulb memories and memories for flashbulb events. Journal of Experimental Psychology: General, 144, 604–623. Hirst, W., & Phelps, E. A. (2016). Flashbulb memories. Current Directions in Psychological Science, 25, 36–41. Holland, A. C., & Kensinger, E. A. (2012).Younger, middle-aged, and older adults’ memories for the 2008 U. S. Presidential Election. Journal of Applied Research in Memory & Cognition, 1, 163–170. Hornstein, S. L., Brown, A. S., & Mulligan, N. W. (2003). Long-term flashbulb memory for learning of Princess Diana’s death. Memory, 11, 293–306. Julian, M., Bohannon, J. N., & Aue, W. (2009). Measures of flashbulb memory: Are elaborate memories consistently accurate? In O. Luminet and A. Curci (Eds.), Flashbulb memories: New issues and new perspectives (pp. 99–122). New York: Psychology Press. Kensinger, E. A., & Schacter, D. L. (2006).When the Red Sox shocked the Yankees: Comparing negative and positive memories. Psychonomic Bulletin & Review, 13, 757–763. Koppel, J., Brown, A. D., Stone, C. B., Coman, A., & Hirst, W. (2013). Remembering President Barack Obama’s inauguration and the landing of US Airways Flight 1549: A comparison of the predictors of autobiographical and event memory. Memory, 21, 798–806. Kraha, A., & Boals, A. (2014). Why so negative? Positive flashbulb memories for a personal event. Memory, 22, 442–449. Kraha, A., Talarico, J. M., & Boals, A. (2014). Unexpected positive events do not result in flashbulb memories. Applied Cognitive Psychology, 28, 579–589. Kulkofsky, S., Wang, Q., Conway, M. A., Hou,Y., Aydin, C., Mueller-Johnson, K., & Williams, H. (2011). Cultural variation in the correlates of flashbulb memories: An investigation in five countries. Memory, 19, 233–240. Kvavilashvili, L., Mirani, J., Schlagman, S., Erskine, J. A. K., & Kornbrot, D. E. (2010). Effects of age on phenomenology and consistency of flashbulb memories of September 11 and a staged control event. Psychology and Aging, 25, 391–404. Kvavilashvili, L., Mirani, J., Schlagman, S., Foley, K., & Kornbrot, D. E. (2009). Consistency of flashbulb memories of September 11 over long delays: Implications for consolidation and wrong time slice hypotheses. Journal of Memory and Language, 61, 556–572. Kvavilashvili, L., Mirani, J., Schlagman, S., & Kornbrot, D. E. (2003). Comparing flashbulb memories of September 11 and the death of Princess Diana: Effects of time delays and nationality. Applied Cognitive Psychology, 17, 1017–1031. Lanciano, T., & Curci, A. (2012). Type or dimension? A taxometric investigation of flashbulb memories. Memory, 20, 177–188. Lanciano,T., Curci, A., Mastandrea, S., & Sartori, G. (2013). Do automatic mental associations detect a flashbulb memory? Memory, 21, 482–493. Lanciano, T., Curci, A., & Soleti, E. (2013). ‘I knew it would happen . . . and I remember it!’: The flashbulb memory for the death of Pope John Paul II. Europe’s Journal of Psychology, 9, 220–230. Larsen, S. F. (1992). Potential flashbulbs: Memories of ordinary news as the baseline. In E. Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb” memories (pp. 32–64). New York: Cambridge University Press. Lee, P. J., & Brown, N. R. (2003). Delay-related changes in personal memories for September 11, 2001. Applied Cognitive Psychology, 17, 1007–1015.

FBMs as ordinary memory processes  93

Liu, A.,Ying, X., & Luo, J. (2012). The flashbulb memory of the positive and negative events: Wenchuan earthquake and acceptance to college. Proceedings of World Academy of Science, Engineering and Technology, 65, 49–54. Luminet, O., Curci, A., Marsh, E. J., Wessel, I., Constantin, T., Gencoz, F., & Yogo, M. (2004). The cognitive, emotional, and social impacts of the September 11 attacks: Group differences in memory for the reception context and the determinants of flashbulb memory. The Journal of General Psychology, 131, 197–224. Luminet, O., & Spijkerman, R. (in press). ‘11 November 1918, an exceptional day!’ Flashbulb memories of the First World War Armistice in Belgium from a psychological and a historical perspective. Memory Studies. Mahmood, D., Manier, D., & Hirst, W. (2004). Memory for how one learned of multiple deaths from AIDS: Repeated exposure and distinctiveness. Memory & Cognition, 32, 125–134. Marsh, E. J., & Tversky, B. (2004). Spinning the stories of our lives. Applied Cognitive Psychology, 18, 491–503. McCloskey, M., Wible, C. G., & Cohen, N. J. (1988). Is there a special flashbulb-memory mechanism? Journal of Experimental Psychology: General, 117, 171–181. Nachson, I., & Zelig, A. (2003). Flashbulb and factual memories: The case of Rabin’s assassination. Applied Cognitive Psychology, 17, 519–531. Nadel, L., Campbell, J., & Ryan, L. (2007). Autobiographical memory retrieval and hippocampal activation as a function of repetition and the passage of time. Neural Plasticity, 2007, 1–14. Neisser, U. (1982). Snapshots or benchmarks? In U. Neisser & I. E. Hyman (Eds.), Memory observed: Remembering in natural contexts (pp. 68–74). San Francisco: Worth Publishers. Neisser, U. (1986). Remembering Pearl Harbor: Reply to Thompson and Cowan. Cognition, 23, 285–286. Neisser, U., & Harsch, N. (1992). Phantom flashbulbs: False recollections of hearing the news about Challenger. In E.Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb” memories (pp. 9–31). New York: Cambridge University Press. Neisser, U., Winograd, E., Bergman, E. T., Schreiber, C. A., Palmer, S. E., & Weldon, M. S. (1996). Remembering the earthquake: Direct experience vs. hearing the news. Memory, 4, 337–357. Niedzwienska, A. (2003). Misleading postevent information and flashbulb memories. Memory, 11, 49–558. Ohman, A., Flykt, A., & Esteves, F. (2001). Emotion drives attention: Detecting the snake in the grass. Journal of Experimental Psychology: General, 130, 466–478. Ost, J., Granhag, P. A., Udell, J., & Roos af Hjelmsäter, E. (2008). Familiarity breeds distortion: The effects of media exposure on false reports concerning media coverage of the terrorist attacks in London on 7 July 2005. Memory, 16, 76–85. Otani, H., Kusumi, T., Kato, K., Matsuda, K., Kern, R. P., Widner, R., & Ohta, N. (2005). Remembering a nuclear accident in Japan: Did it trigger flashbulb memories? Memory, 13, 6–20. Paradis, C., Solomon, L. Z., Florer, F., & Thompson, T. (2004). Flashbulb memories of personal events of 9/11 and the day after for a sample of New York City residents. Psychological Reports, 95, 304–310. Pezdek, K. (2003). Event memory and autobiographical memory for the events of September 11, 2001. Applied Cognitive Psychology, 17, 1033–1045. Pillemer, D. B. (1984). Flashbulb memories of the assassination attempt on President Reagan. Cognition, 16, 63–80.

94  Talarico and Rubin

Pillemer, D. B. (2009). “Hearing the news” versus “being there”: Comparing flashbulb memories and recall of first-hand experiences. In O. Luminet and A. Curci (Eds.), Flashbulb memories: New issues and new perspectives (pp. 125–140). New York: Psychology Press. Romeu, P. F. (2006). Memories of the terrorist attacks of September 11, 2001: A study of the consistency and phenomenal characteristics of flashbulb memories. The Spanish Journal of Psychology, 9, 52–60. Rubin, D. C. (2012). The basic systems model of autobiographical memory. In D. Berntsen & D. C. Rubin (Eds.), Understanding autobiographical memory: Theories and approaches (pp. 11–32). New York: Cambridge University Press. Rubin, D. C., & Berntsen, D. (2003). Life scripts help to maintain autobiographical memories of highly positive, but not highly negative, events. Memory & Cognition, 31, 1–14. Rubin, D. C., & Kozin, M. (1984).Vivid memories. Cognition, 16, 81–95. Rubin, D. C., & Schulkind, M. D. (1997). Distribution of important and word-cued autobiographical memories in 20-, 35-, and 70-year-old adults. Psychology and Aging, 12, 524–535. Rubin, D. C., & Umanath, S. (2015). Event memory: A theory of memory for laboratory, autobiographical, and fictional events. Psychological Review, 122, 1–23. Schmidt, S. R. (2004). Autobiographical memories for the September 11th attacks: Reconstructive errors and emotional impairment of memory. Memory & Cognition, 32, 443–454. Schmidt, S. R. (2012). Extraordinary memories for exceptional events. New York: Psychology Press. Schmidt, S. R., & Bohannon, J. N. (1988). In defense of the flashbulb memory hypothesis: A comment on McCloskey, Wible, and Cohen (1988). Journal of Experimental Psychology: General, 117, 332–335. Schmolck, H., Buffalo, E. A., & Squire, L. R. (2000). Memory distortions develop over time: Recollections of the O.J. Simpson trial verdict after 15 and 32 months. Psychological Science, 11, 39–45. Scott, D., & Ponsoda,V. (1996). The role of positive and negative affect in flashbulb memory. Psychological Reports, 79, 467–473. Sharot,T., Martorella, E. A., Delgado, M. R., & Phelps, E. A. (2007). How personal experience modulates the neural circuitry of memories of September 11. Proceedings of the National Academy of Science, 104, 389–394. Smith, M. C., Bibi, U., & Sheard, D. (2003). Evidence for the differential impact of time and emotion on personal and event memories for September 11, 2001. Applied Cognitive Psychology, 17, 1047–1055. Stone, C. B., Mercy, A., Licata, L., Klein, O., & Luminet, O. (2013). Mnemonic differences and similarities across opposing social groups: The linguistic conflict at the University of Leuven as a case study. Journal of Applied Research in Memory and Cognition, 2, 166–172. Symons, C. S., & Johnson, B. T. (1997). The self-reference effect in memory: A meta-analysis. Psychological Bulletin, 121, 371–394. Talarico, J. M., LaBar, K. S., & Rubin, D. C. (2004). Emotional intensity predicts autobiographical memory experience. Memory & Cognition, 32, 1118–1132. Talarico, J. M., & Moore, K. M. (2012). Memories of ‘The Rivalry’: Differences in how fans of the winning and losing teams remember the same game. Applied Cognitive Psychology, 26, 746–756. Talarico, J. M., & Rubin, D. C. (2003). Confidence, not consistency, characterizes flashbulb memories. Psychological Science, 14, 455–461. Talarico, J. M., & Rubin, D. C. (2006). Flashbulb memories are special after all; in phenomenology, not accuracy. Applied Cognitive Psychology, 21, 557–578.

FBMs as ordinary memory processes  95

Talarico, J. M., & Rubin, D. C. (2009). Flashbulb memories result from ordinary memory processes and extraordinary event characteristics. In O. Luminet and A. Curci (Eds.), Flashbulb memories: new issues and new perspectives (pp. 79–97). New York: Psychology Press. Tekcan, A. I. (2001). Flashbulb memories for a negative and a positive event: News of Desert Storm and acceptance to college. Psychological Reports, 88, 323–331. Tekcan, A. I., Ece, B., Gulgoz, S., & Er, N. (2003). Autobiographical and event memory for 9/11: Changes across one year. Applied Cognitive Psychology, 17, 1057–1066. Tekcan, A. I., & Peynircioglu, Z. F. (2002). Effects of age on flashbulb memories. Psychology & Aging, 17, 416–422. Thompson, C. P., & Cowan, T. (1986). Flashbulb memories: A nicer interpretation of a Neisser recollection. Cognition, 22, 199–200. Tinti, C., Schmidt, S., Sotgiu, I., Testa, S., & Curci, A. (2009). The role of importance/ consequentiality appraisal in flashbulb memory formation: The case of the death of Pope John Paul II. Applied Cognitive Psychology, 23, 236–253. Tinti, C., Schmidt, S., Testa, S., & Levine, L. J. (2014). Distinct processes shape flashbulb and event memories. Memory & Cognition, 42, 539–551. Walker, W. R., Skowronski, J. J., & Thompson, C. P. (2003). Life is pleasant – and memory helps to keep it that way! Review of General Psychology, 7, 203–210. Weaver, C. A., III. (1993). Do you need a “flash” to form a flashbulb memory? Journal of Experimental Psychology: General, 122, 39–46. Weaver, C. A., III, & Krug, K. S. (2004). Consolidation-like effects in flashbulb memories: Evidence from September 11, 2001. American Journal of Psychology, 117, 517–530. Weaver, C. A., III, Terrell, J. T., Krug, K. S., & Kelemen, W. L. (2008). The delayed JOL effect with very long delays: Evidence from flashbulb memories. In J. Dunlosky & R. A. Bjork (Eds.), Handbook of metamemory and memory (pp. 155–172). New York: Psychology Press. Winningham, R. G., Hyman, I. E., & Dinnel, D. L. (2000). Flashbulb memories? The effects of when the initial memory report was obtained. Memory, 8, 209–216. Winograd, E., & Killinger,W. A. (1983). Relating age at encoding in early childhood to adult recall: Development of flashbulb memories. Journal of Experimental Psychology: General, 112, 413–422. Wright, D. B. (1993). Recall of the Hillsborough disaster over time: Systematic biases of “flashbulb” memories. Applied Cognitive Psychology, 7, 129–138. Wright, D. B., & Gaskell, G. D. (1992). The construction and function of vivid memories. In M. A. Conway, D. C. Rubin, H. Spinnler, & W. A. Wagenaar (Eds.), Theoretical perspectives on autobiographical memory (pp. 241–261). Dordrecht, Netherlands: Kluwer Academic Publishers. Wright, D. B., Gaskell, G. D., & O’Muircheartaigh, C. A. (1998). Flashbulb memory assumptions: Using national surveys to explore cognitive phenomena. British Journal of Psychology, 89, 103–118.

5 THE CONSEQUENCES OF CONSEQUENTIALITY Jasmen Rice, Karina Hamamouche, and John N. Bohannon III

FBMs result from consequential events that disrupt personal or cultural plans. This simple position, however, fails to address several factors inherent to FBMs. This chapter explores a three-dimensional model of consequentiality, organized as interactions between opposing event factors: self vs. other, social vs. physical, and pleasant vs. aversive events. We argue that these conditions continually select event memories for retention making the “special mechanism” argument usually employed in the FBM debates simply wrong-headed. A study of people’s recall on discovering the election of America’s first black president is presented to test the valence dichotomy.

Preamble One score and 19 years ago, Brown and Kulik (1977) brought forth a new memory, conceived as a flashbulb of source recall and dedicated to the proposition that, otherwise, it was created equal to other memories. Now we are still engaged in a great debate whether any memory labeled “special” or produced by a “special” mechanism can long endure. We are met here in the second volume of that debate. The brave researchers who struggled within this arena have illuminated its boundaries far beyond the poor power of this chapter to add or detract. The world will little note what we write here, but will long remember their Flashbulb memories. It is rather for us here to be dedicated to the task remaining before us – that we highly resolve that flashbulb phenomena conceived as source recall, selected by consequentiality, supported by rehearsal, and encoded by arousal shall not perish from the journals.1 Particularly poignant is that the first published flashbulb accounts were from 19th century subjects recalling their discoveries of Lincoln’s assassination (Colegrove,

The consequences of consequentiality  97

1899). After the modern assessment of discovering horrific news began in 1970s, the topic of Flashbulb memory (FBM) has suffered the slings and arrows of outrageous assertions.2 McCloskey, Wible, and Cohen (1988) insisted that FBMs should be 100% complete, 100% immutable, and 100% resistant to forgetting in 100% of the participants. Only then would those authors consider FBMs “special.” Granted, if any psychological phenomenon displayed those distinctions, we would not just consider it special, but unique in our discipline. Predictably, FBMs failed to live up to those demands even in the 60 odd participants tested by McCloskey et al. (1988) recalling the Challenger disaster. Neisser and Harsch (1992) sought similar evidence from Challenger discovery memories in 44 participants who had their discovery memories retested after three years.They concluded some of the follow-up memories were “wildly inaccurate” (i.e. inconsistent; for a description of the “accuracy” estimate see Julian, Bohannon, & Aue, 2009). Two studies attempted to compare an FBM with more mundane event memories. Weaver (1993) used college moving-in day memories in 22 participants to compare to the opening of the first Iraq war. Talarico and Rubin (2003) asked 54 participants to select an event from the prior weekend to compare to 9/11. Although Weaver (1993) found no reliable differences, Talarico and Rubin (2003) found all the rating scales showed significant differences (e.g. FBMs were more vivid, confident and arousing whereas the scored recall data was in the predicted direction, but not significant). Please note the small sample sizes of these oft-cited studies. Wasserman and Lazar (2016) were particularly critical of small sample studies in a recent paper outlining the position of the American Statistical Association for two reasons. First, small sample studies have the highest rate of spurious significant findings that often fail to be replicated. Small sample investigations frequently occur in the population of research projects, requiring less time, money, and effort, and thus run the risk of unusual results just by chance. Secondly and worse yet, small sample studies in arenas of noisy data are also more likely to see differences, but fail to show significance. Public event FBMs, such as the Challenger explosion or 9/11, are exactly that sort of noisy study when scoring memory narratives (Bohannon, 1988). Indeed, when submitting reports to the prestigious journal, Psychological Science, authors are now required to justify why they stopped collecting data and argue their sample size was sufficient to answer the question. It is likely that many earlier FBM studies would not have satisfied this criterion. It is therefore somewhat disconcerting to see all FBM studies in this volume evaluated as equally veridical without considering the above basic assumption. Does a study of 22 students (Weaver, 1993) showing no differences in memory measures really balance the likes of Er (1997) who assessed 655 participants and did? Aue et al. (2004) attempted to replicate Talarico and Rubin (2003) using the Columbia disaster with over 300 subjects and found both significant rating differences and narrative recall differences. Thus, let us resolve to apply our current, stringent standards to the flashbulb literature. Those studies which have no reason for a limited sample, unlike Mahmood, Manier, and Hirst’s (2004) study

98 Rice et al.

of discovering deaths by AIDS (N = 80), should be weighed as less reliable than reports from more sizable samples. At the very least small samples should be noted with every citation to let the reader make a more considered evaluation. Unless we identify the studies with unreliable results, we suffer the consequence of allowing questionable results to influence our ideas of consequentiality. Thus, this has to be the first consideration prior to constructing a new model of FBM mechanisms. This chapter will lay out several solid assumptions about FBMs within the greater perspective of neurological compromises adopted to allow experiences to be remembered. FBMs possess qualities that exemplify four of the seven memory “faults” presented in Schacter’s (2002) book, The seven sins of memory. These four “sins” are particularly relevant to FBM theory and the investigations and conclusions that result. Then, we will explore the nature of consequentiality in the creation of FBMs using a proposed taxonomy of consequentiality. Lastly, we will present data on people’s recall of discovering the election of Barack Obama, the first black president of the United States. This particular event is potentially useful in comparing different values of arousal and valence within the new 3-D model of consequentiality.

Assumption ambulation Before exploring the nature of consequentiality, some basic assumptions beyond insisting on adequate methodology, should be established. We will borrow heavily from Schacter’s (2002) book, The seven sins of memory. He offered up seven common memory “faults” or imperfections and explained them as design compromises needed to deal with the organic limitations of rememberers. We will extract for your perusal some of the most relevant compromises related to FBM. First, memory has to be selective (Cahill & McGaugh, 1998). Remembering everything is simply impossible. Because each memory consolidation involves rearranging connections in neural real estate, recalling every instant of even a single day could not occur. We don’t, even with our large brains, have that sort of gargantuan capacity. Animals with smaller brains like tree shrews and monkeys, from whom we evolved, had this problem in spades. It is unlikely that systems/mechanisms designed to help select experience worth precious neural real estate disappeared as organisms ascended the evolutionary tree. We humans are likely to employ those same encoding selection mechanisms (Cahill & McGaugh, 1998). Further, consolidation of memories takes time (Gold, 1986). It is quite likely contiguous details are lost whilst attempting to record the last experience. Given the lack of room and the time restrictions, selecting which memories to record is a critical process for all animals, even the big brained ones like humans. FBMs are just one product of the selection process. Second, selective memory cannot be random. A process that is as likely to invest as much neural tissue in recording irrelevant experiences as adaptive experience

The consequences of consequentiality  99

(e.g. avoiding social or physical damage and pursuing reproductive opportunities) is antithetical to evolutionary theory (Cahill & Alkire, 2003). The arousal-based FBM mechanism and rehearsal are two means by which experiences can be selected for memorial encoding. Events (threat and sex) that trigger amygdaloid activity that agonistically affects hippocampal function will enhance encoding (Cahill & Alkire, 2003). Further, if something is socially and personally important, such that recollections are frequently rehearsed and overtly shared with others, then the memory also remains available. Most autobiographical recollections fit the nonrandom prediction, being social triumphs, failures or involving threat. Public FBMs are borderline events within this assumption depending upon other factors often labeled consequentiality. Third, recorded experiences tend to be consistent. Memory evolved to aid adaptation from past experience (Schacter, 2002). If recalled details are as likely to be wrong as veridical, then memory cannot be adaptive. If we rely on consistency to estimate accuracy then there is no evidence that FBMs are inherently more inconsistent than any other kind of recollection. Bohannon and Symons (1992) (N = 116) found only 33% of the FBM details of the Challenger disaster changed over the course of three years. No longitudinal FBM study (see the six longitudinal studies described above) has ever found more than 35% inconsistency. Indeed, even the small sample studies like McCloskey et al. (1988) and Neisser and Harsch (1992) reported no more than 35% inconsistent recall. Please note, forgetting over time is not inconsistency. If a recalled detail included in one assessment of FBM is not recalled, then forgetting occurred (see Schacter, 2002). Importantly, the reasons one forgets are different than the mechanism whereby one answer is replaced by another (inconsistency). Indeed, when Tyson et al. (2003) assessed consistency in their N = 700+ sample of 9/11 memories, they had to include instances of forgetting as “inconsistent” just to get the consistency rate for 9/11 FBMs below 93%. If so, from whence comes this assumption that FBMs are inherently inconsistent? We suspect it was from Ulric Neisser himself. The most famous and errant FBM was Ulric Neisser’s (1982) (N = 1) report of his own memory discovering Pearl Harbor (Thompson & Cowan, 1986). Neisser grew up in New York City as an avid baseball fan. He recalled ten details from his Pearl Harbor FBM (see Julian et al., 2009, for a table of details), nine of which were accurate but unfortunately strongly associated with his hobby, baseball. Those details were also true for the veridical football contest playing on his radio on the afternoon of December 7th, 1941. In a self-generated DRM (Deese-RoedigerMcDermott, see Roediger & McDermott, 1995) procedure where strong semantic associations trigger a false memory, Neisser recalled “baseball” (instead of the accurate “football”) as the contest, which could not have occurred in December after the World Series which ended in October. It was remarkable that his DRM error could not have occurred unless he accurately recalled the other nine details. Thus, it is monumentally tragic that the single flaw in his own FBM was over generalized to the rest of his FBM recall (Neisser, 1982) and then to all FBMs in all people (Neisser & Harsch, 1992).

100 Rice et al.

Fourth and lastly, memory cannot be immutable (Schacter, 2002). Otherwise, errorful memories could never be corrected and we would all still recall how our favorite TV cartoon characters were real and lived in our television sets. Julian et al. (2009) used this logic to suggest that inconsistent memories are unlikely to be uniformly false after mutation. Sometimes changes led to error correction. This was particularly true of early news reports of the 9/11 attacks, which reported as many as 50,000 dead in the World Trade Center collapse (Julian et al., 2009). These factual errors about 9/11 were corrected over the next week in the media. Kiburz, McClure, and Bohannon (2009) asked 86 participants to view a slide show and have their memories for the show tested immediately, one week, and two weeks later. The initial test examined the participants’ memories individually. At a oneweek delay, 48 participants formed 24 dyads and remembered in pairs, writing down their answers, discussing them and selecting a single answer upon which they both agreed. After a three-week delay, all participants again recalled the slide show individually. The results were striking. Initial errors were twice as likely to be replaced by correct answers. Further, the now corrected answers were more likely to appear in week three than the initial error. Lastly, participants’ confidence in their answers predicted several results of note. First, confident answers were significantly more likely to be correct. Second, correct answers initially rated as strongly confident were almost never surrendered during discussion a week later. This may be the role of confidence in FBM maintenance. Although not a complete assurance of veridicality, high confidence in an FBM detail predicts veridicality and maintains its integrity over time in being resistant to change from other sources. However, Neisser (1982) showed that internal mechanisms of change, such as DRM semantic associations, may include the rare error about which the rememberer is also quite confident, albeit wrongly so. Whereas The seven sins of memory laid out key features of FBMs, one feature commonly included in FBM models is consequentiality (e.g. Conway, 1995; Er, 2003). As such, we have chosen to explore the nature of consequentiality in FBMs throughout the remainder of this chapter.

A taxonomy of consequentiality Conway (1995) used the single factor of consequentiality early and centrally in his complex structural equation model of FBM formation. To be included in the FBM model, a consequential event must be disruptive of personal or cultural plans (Conway, 1995). Only after determining the consequentiality of an event do the mechanisms of affect and rehearsal begin to function. This idea is quite friendly with reconstructive memory theory. As a person attempts to recall past events, particularly the where and when of events, major changes in the path of life allow easy retracing of the course of life and recall of the point of turning. Brown et al. (2009) labeled the time of life after the drastic change a life epoch, and argued memories are organized by epochal boundaries. However, this seems too simple.

The consequences of consequentiality  101

Not all consequences for the rememberer are created equal, nor are they necessarily governed by the same memory mechanisms. We humbly offer a taxonomy of consequences. The first dichotomy might be whether or not the event occurs to the rememberer. Whatever consequences occur, they must occur directly to that person. This distinguishes FBM victims vs. observers vs. distance learners. In the forensic realm, it separates the assault victim from the witness and later, the jury. The second dimension categorizes the nature of the event, physical or social (figure 5.1). Physical events commonly involve some form of injury to the self or others, although sex and kisses are also physical events. Social events involve some change in social relation or in social position, often for the worse. The last dimension is determined by valence, pleasant or aversive. Many pleasant events like reproductive opportunity (kisses and sex) are just as memorable as are aversive events (divorce and car accidents). How the valence affects autobiographical recall has yet to be studied extensively. We outline the six alternatives from the combination of the three dimensions below. We may well be questioned for not including another dimension, private vs. public (e.g. recalling falling out of a tree and breaking a bone vs. discovering 9/11); however, we omitted this dimension for two excellent reasons. First, the distinction is artificial and only allows researchers to narrow the time of encoding for a population (i.e. 9/11). The private vs. public distinction does not

Pleasant

Self

Social

Physical

Other FIGURE 5.1 Hypothetical

Aversive

taxonomy of FBM-inspiring events and their consequentiality. Horizontal is the nature of the event.Vertical is the target of the event or to whom the event occurs. Diagonal represents the valence of the event. FBMs such as assassinations typically fall in the physical, aversive, and other volume of the FBM space.

102 Rice et al.

vary the putative memory mechanism one tittle. Exactly the same arousal, rehearsal, and consequentiality variables apply to both public and private events that inspire vivid, long-term recall. The second reason is that a fourth dimension would not easily fit into Figure 5.1 and would make it confusing.

Self vs. other Events that happened to the rememberer are, in general, better recalled than events the rememberer learned happened to another, when all other dimensional values were equal. Our personal rewards and disasters will always be more consequential and enhance our autobiographical recall more than the same events happening to others. Contrarily, recalling events that happen to others is the basis of observational learning (Bandura, 1986). It is much better to watch terrible consequences accrue to others than have to suffer them yourself. With Bandura’s (1986) model of observational learning, consequences to others triggered the acquisition of the observed event. In the original Bandura, Ross, and Ross (1963) Bobo studies, children in the no-consequence condition were the only group that recalled significantly fewer novel aggressive acts than the two consequence groups. To highlight this distinction, Hilllman et al. (2010) assessed adults (N = 182) who had been injured as children without losing consciousness but whose injury severity required a hospital visit. Clearly the injury happened to the rememberer and their recall of details such as weather, location, others present was quite extensive. In an attempt to verify these long-delayed recollections, the parents (N = 109) were contacted to check their memories of the same event. Luckily for FBM theory, over 40% of the parents were present and saw their child’s injury. Evolutionary theory would hold that injury to one’s child holds tremendous evolutionary consequences for the parent even without the parent directly benefitting (memorial-wise) from the physical trauma. Indeed, viewing parents’ narrative recall was extensive and equal to the child’s own narrative. However, the self vs. other advantage prevailed as predicted in that child victims recalled more precise, probed details.This is expected as direct pain mechanisms work in the child victim (Gold, 1986), but were absent in the parental observer. Thus, victims of injury (self) had the most detailed recall, whereas their viewing parents (proximal other) had almost as good recollections and the distant learning parents (distal other) had decent recall but not as extensive as the other groups. Note that the overwhelming majority of FBM studies examined the memory of distant learners whose recall was likely the worst of the three above conditions. Henry, Bohannon and Miller (2000) also captured a similar distinction by examining adults (N = 102), some of whom discovered the death of their own parent whereas the remainder discovered the death of a parent of a close friend. Many of the deaths were expected due to the results of long-term, fatal diseases and some were miserable surprises due to accidents. There was a main effect of self vs. other where the death of your own parent was more memorable than the death of your friend’s parent, highlighting the effect of personal involvement in FBM creation.

The consequences of consequentiality  103

However, not all personally experienced events may result in vivid and extensive recall. Christianson (1992) suggested that the arousal effect is not monotonic but an inverted “U” (e.g. Yerkes-Dodson Law). Too much arousal results in panic and several effects antagonistic to good memory (e.g. narrowing of perception to only the most threatening of stimulus elements). Further, Diamond, Park, Campbell, and Woodson (2005) found that extended arousal caused a buildup of cortisol in the hippocampus inhibiting further learning for some time after the arousing stressor terminates. Rats placed in plexiglass boxes to protect them and then placed in a cage next to a cat for 15 minutes were unable to learn to solve a maze a half hour later. Just being in the plexiglass box with no cat allowed rapid maze learning with the same delay to test. Leider, Sauer, and Bohannon (2012) asked car accident participants (N = 66) to recall the events at the time and at half hour intervals following the accident. Diamond et al. (2005) predicted what was found in humans. There was a peak of memory details a half hour after the collision and lack of details an hour after the accident. Clearly, the arousal mechanism is more complex than initially envisioned by Livingston’s “Now Print” theory (Livingston, 1967). It is rare that a self vs. other distinction results in public FBM events. The victims of the Challenger disaster all perished obviating such comparisons for that public event. However, the astronaut Sally Ride (private communication, 1994) was scheduled to be onboard the shuttle for that fatal mission. She had contracted, as it turned out, a life-saving flu. Her best friend, in whose wedding she served as matron of honor, Judith Resnick took her place. Sally Ride’s job that January morning was to escort the crowd of crew member’s families and loved ones to a viewing stand for the shuttle launch. Seventy-three seconds post launch, the Challenger’s main fuel tank exploded and the intact, solid rocket boosters flew off at tangents. The mass of close relations on the viewing stand cheered, thinking that was a normal booster separation. Sally Ride knew differently and as she turned to the loved ones, she knew she was the one to tell them that their beloved crew members were all dead. Sally Ride observed directly but did not directly suffer the pain of the explosion, however her memory of the moment is vivid and she related it with extreme confidence.

Social vs. physical Social membership Humans are social animals. Therefore, threats to our social membership in groups, agonistic social standing, and even individual romantic pair bonds are treated seriously and are often recalled like FBMs. For example, Eisenberger, Lieberman, and Williams (2003) told participants to play catch in simple computerized three-way catch simulations, in which they believed to be playing with two other humans. The participants tossed around a simulated ball with two (computer controlled) avatars and after a while the two (computer controlled) avatars played catch with each other, excluding the human. Human participants got so angry they often

104 Rice et al.

terminated the experiment early. Thus, exclusion from a group even as small as a triad can lead to strong emotion, and within the arousal mechanism stronger, more elaborate memory for the event. The list of social exclusions in everyday life is long indeed: getting fired, divorce announcements, breakups of long-term nonmarriage relations and even being excluded before a social group is formed, as in being rejected by a potential romantic partner (asking another for a date and being told no).

The effects of in-group membership on FBM Pokorny and de Waal (2009) observed that monkeys often change family groups as they grow. Group membership determines both altruism and food sharing and also out-group membership elicits hostility and aggression. This occurred toward a departed monkey even when the former family is the group with whom the monkey had spent the majority of its life. Thus, even as we move from group to group, our current group membership is important to primates and, by extension, determine FBMs. Conway et al. (1994) found Danish people did not have extensive memories for Margaret Thatcher’s resignation like resident British folks. Given Pokorny and de Waal’s (2009) perspective, had some Danes moved to Britain for their residence, they may have recalled Thatcher’s resignation like multi-generation British people. Similarly, a popular college president, Dr. Bobby Fong, unexpectedly resigned from Butler University in 2011. Rice, Jennings, Lange, and Bohannon (2012) tested 100 faculty and students for their recall of discovering the news. Students had absurdly impoverished FBMs for the presidential resignation event whereas the faculty and staff possessed significantly more extensive recall. Long-term employees of the University had more group membership, which resulted in the departure of the president being more disturbing. In addition, faculty and students who rated themselves as more upset at the news had more extensive FBMs. Butler University’s successful and much beloved basketball coach, Brad Stevens, departed to coach the professional NBA team, the Boston Celtics. Puga, Asuncion, White, Lauth, and Bohannon (2014) studied 92 Butler students who varied in their interest and familiarity with the school’s basketball program. Students who identified with Butler basketball had significantly greater FBMs (50% more extensive) for discovering the coach’s departure. One oft-changing social group may be romantic partners. In the sequence of romance formation, maintenance, and termination, there are conceivably a plethora of occasions to feel threat or euphoria (Baumeister & Leary, 1995). The first possible memory-inducing injury comes before the formation of the romantic dyad. Someone must express romantic interest and ask the romantic target for a date. Pajkos, Sauer, Bohannon, Jordan, and Thompson (2011) ran a dating simulation where participants (N = 52) all viewed the same video introducing a young man seeking a date. The video fellow narrated his career goals and values along with details about his life preferences. All participants were invited to type out a dating

The consequences of consequentiality  105

request to be submitted to the handsome stranger on the video. After some ratings on the video person’s attributes, the participants are told they had the reply from the video person. Although all participants were rejected, half were rejected because of happenstance (the person already had a date that night) and half were rejected harshly due to the video person’s assessment of the participant (“You are out of my league. Try to find someone more tolerant of your obvious flaws.”). The second rejection was a socially fundamental threat. It questioned the participants’ evolutionary/ romantic adequacy. All participants completed a memory assessment of the video person’s information both immediately and after one week. The harsh rejection group showed significant retrograde memory enhancement of the details of the target at both times of testing, again demonstrating the importance of social threat in FBM creation.

The effects of out-group membership on FBM The observed events described above may be hugely consequential for both the observer and actor. For example, the extent that the observer is familiar with the actor in the event or a member of a social group containing the actor might further determine the extent of the learner’s recall.When the target of the FBM is not the rememberer, the relation of the observer to the target is a complex problem. Brown and Kulik (1977) [N = 80] found that black participants had extensive FBMs for discovering the death of Dr. Martin Luther King whereas white participants did not. Danes had trouble recalling the resignation of Margaret Thatcher, whereas British participants had extensive FBMs (Conway et al., 1994, N = 569). Curci et al. (2015) examined two aspects of group membership when studying the resignation of the Pope (N = 61). The study compared Catholics (both practicing and not) with non-Catholics and found FBM details were most strongly associated with true recall according to the implicit association test, and group membership (practicing Catholics) predicted more extensive FBMs. However, American students (N = 504) (Bohannon, Gratz, & Cross, 2007) had excellent and extensive FBMs for Princess Diana’s death despite not being British royalty or British citizens. Clearly, group membership or identification with victims may be difficult to discern. An example of this wavering group membership problem was revealed some years ago whilst one of the current authors was having lunch with Jack Valenti, the President of the Motion Picture Association of America and former President Lyndon Johnson’s speech writer.When he was told about FBMs, he asked for some examples. The explosion of the Challenger (no memory) and the death of Princess Diana (no memory) yielded only puzzled scowls. However, when cued with John Kennedy’s assassination,Valenti’s face split into a huge grin as he replied, “Of course I remember, I was in the third car!” Indeed,Valenti was in the Kennedy motorcade just 60 feet away when the president was shot. Although a single correspondence cannot completely speak to predictors of FBMs, this story suggests that maybe simple proximity to the event might be part of the mechanism determining an event’s

106 Rice et al.

relationship to a rememberer. Although, this suggestion of proximity is merely anecdotal, evidence exists that proximity may play a strong role in the formation of FBMs (e.g. Conway et al., 1994; Pezdek, 2003).

Group membership via physical proximity Pezdek (2003) (N= 569) examined groups of students from New York, California, and Hawaii to determine whether proximity might predict the extent of 9/11 FBMs. She found New Yorkers recalled the most of the series of 9/11 events (i.e., point of impact on North Tower, time elapsed between North Tower impact and South Tower impact, time elapsed between North Tower impact and South Tower collapse) whereas those further away had more traditional FBMs simply detailing their discoveries. The trouble with Pezdek (2003) was intrinsic to the 2.5-hour duration of the 9/11 event sequence: three well known plane strikes, two in NYC, and two tower collapses also in NYC. Pezdek’s protocols relied heavily on the participants correctly sequencing these events. In fact, New York City residents may have as many as four or five FBMs as they were likely awake at 8:45 AM EST and exposed to media and personal communication throughout the sequence. Some California participants may have been awake at 5:45 AM local time when the 9/11 attacks started, but Hawaian residents (2:45 AM) only awoke when the 9/11 sequence was over. The most distant participants were more likely to have gotten a traditional, single announcement of the attacks, much like Americans on the east coast got a single announcement of Pearl Harbor. To test the time zone hypothesis, Barga, Bohannon, Bluck, Laney, and Yarmey (2005) assessed 744 participants from California to Canada to Miami, with some participants within 50 miles of NYC. The sample varied in both time zone and distance from the east coast 9/11 events. Similar to Pezdek (2003), the most distant California participants had different FBMs than those closer to NYC. However, when only those participants from the same time zone as the 9/11 events were examined, the distance effect disappeared. Clearly, being awake for the entire 2.5-hour sequence of the 9/11 attacks made for a sequential FBM unlike other FBM announcements, which occurred in a single communication (e.g. “The space shuttle Challenger has exploded. . .”).

Physical injury The likelihood of injury or the delivery of pain is a consequence that animals have been trying to avoid since they evolved locomotion. After a painful event, the victim’s body is flooded with hormones and a delayed surge of plasma glucose (Gold, 1986). He found that if the glucose surge was blocked, memory for the painful event was not enhanced. However, if a glucose surge was artificially induced, then an event that was mildly painful was remembered just as well as an event involving immense pain. In human development, the age of childhood amnesia is thought to occur just before the age of six in controlled memory experiments (Multhaup, Johnson, & Tetirick, 2005) and FBMs starting at age six are little different from

The consequences of consequentiality  107

adult accounts (Winograd & Killenger, 1983). However, if you ask adults to recall a childhood personal injury, there is no significant developmental difference across the ages of under six years to early teens (Hillman et al., 2010, N = 182). These childhood recollections were compared to the parents’ recollections of the injury (N = 103). When a child victim and their parent recalled the same detail (e.g. day of the week, location, etc.), their agreement was 93% between adult and child. This study suggests that just as in animal experiments (e.g. Gold, 1986), humans seem to treat the threat of injury to the self (in circumstances not involving unconsciousness) as a memory enhancing mechanism. Everyday instances of carelessness that result in closing drawers or doors on one’s fingers or stubbing one’s toes in the dark, can generate significant recall despite the fact that one’s life rarely changes drastically. Thus, from an evolutionary perspective, avoiding injurious events and their painful consequences may be the most archaic of selective, memory/learning mechanisms. Of the typical public FBMs studied, the majority ask about discovering the deaths of others like 9/11 or assassinations like President Kennedy.

Valence: Aversive vs. pleasant All FBMs have some degree of valence. While the effects of event valence, pleasant or aversive, seem to have little immediate effect on encoding (see Versace, Bradley, & Lang, 2010), a plethora of work has suggested that valence may be an important feature in FBM maintenance.The work of Rubin (Rubin & Kozin, 1984; Rubin & Schulkind, 1997) indicated that many people reminisce more about pleasant events than unpleasant. Thus, it may be that instead of having a direct effect on the initial memory consolidation, the valence of an event may cause a rememberer to recall the nicer events from their lives and thus generate more recounts and more rehearsal support for those memories. Indeed after the passage of more than five years, sexual encounters are remembered as well as accidental encounters between cars (Sauer et al., 2012). After all, memories for sex, kisses, and car accidents are produced and maintained by the same memory mechanisms. Unfortunately, nature does not counterbalance valence with the other factors/dimensions. Physical injury is almost never pleasant and winning races almost always are glorious. Contrasting the valences of event is a tricky issue due to their inherent confounding but some comparisons shed light on the effect.

Pleasant events Once a romantic dyad is formed, elements of the romantic progress of the dyad often result in FBM-like recall for the event sequence. Reed, Bohannon, Gooding, and Stehman (2000) assessed 100 couples and had each, in isolation, recount the story of their first kiss with the promise that the experimenters would never reveal the stories to their partners. The amount of details between partners varied with the arousal of the kiss and consistency between kissers varied with the number of recounts. Vaclavik, Sauer, Jennings, and Bohannon (2011) examined

108 Rice et al.

memories for first sexual encounters and found that encounters whose valence was rated as pleasant and arousing had more extensive and detailed narratives than those whose sex valence was rated as unpleasant and not very arousing. Sauer, Vaclavik, and Bohannon (2012) asked participants for their recollection of marriage proposals. Proposals that resulted in high rated valence and whose marriage was ongoing had the most detailed recollections. The failure of arousal to generally predict proposal recall was most likely due to a ceiling effect as all proposals resulted in high arousal.

Aversive events Hinton, Heckman, Agypt, and Bohannon (2005) asked participants (N = 123) to recount elements of a past, failed romance and describe three events in the sequence: first meeting, first kiss, and breakup. As expected, both kisses and breakups were rated as significantly more arousing than first meetings. However, first meeting and kisses were rated as much more pleasant than the breakup. When the recollective narratives were scored, only arousal predicted the extent of the memories and not valence. If the event was arousing, either pleasantly through kisses or miserably through breakups, an extensive engram of the event resulted. It seemed that breaking up was hard to do, but easy to remember. Together, these studies demonstrate the importance of arousal, rather than valence, on memory encoding. One of the worst forms of breakups, which induce major life changes, is divorce. Vaclavik, Sauer, Modesti, and Bohannon (2012) examined the recollections of 65 divorcees and found the extent of recall was determined by the arousal at the time of the announcement. However, in order to maintain that level of detail, the divorcees needed to recount their story many times. That interaction of rehearsal and initial arousal was initially reported for FBMs of the Challenger disaster (Bohannon, 1988). This supports our contention within the model that person event memory mirrors recall of public FBMs in the memory mechanisms employed. However, the adults who proceed with their divorce affect others as well. Children of divorced parents also receive an announcement of the termination of their nuclear family. Lifestyle changes from that point forward are often unavoidable for the children, including residence and school changes (Hetherington, Cox, & Cox, 1979). Agypt and Bohannon (2006) examined 50 adults’ recollection of their parent’s divorce announcements and created an estimate of life change to test whether change alone determined consequentiality and memory. For example, divorce sometimes resulted in immediate relocation, school changes and more. For other children and families, nothing changed except the absence of the father. The number of changes reported by the, now adult, children did not predict the quantity or elaborate nature of their parental divorce narratives. However, affect and recounts jointly determined the extent of recall in a fashion similar to the divorcing adults and outlined above. Further, the children received the news at differing ages. Only the older children (aged 12 or older) were upset by the news and recalled more. It was possible that the younger children did not understand the repercussions of the new family arrangement of single parenting.

The consequences of consequentiality  109

In the course of a romantic dyad there are other sorts of aversive threats to dyad survivability, which may serve as FBM-inducing events. In monogamous species, members of the mated dyad become territorial and defend their mated status against interlopers (Buunk, Angleitner, Oubaid, & Buss, 1996). Explicitly, how memorable would a discovery of one’s partner’s actual infidelity be? Fisher, Bable, and Bohannon (2013) asked 149 participants who had been the victims of infidelity from their partners about their discovery of the situation. A strong gender by infidelity type interaction predicted the extent of their discovery recollections. Males recalled more details from their discoveries if the infidelity included sexual behavior whereas females recalled more details from discovering their male partners were romancing another female. Given the varied mechanisms inherent in the concept of consequentiality (see Er, 2003; Luminet & Curci, 2009), it seems unlikely that a single measure such as envisioned by Conway (1995) will supply sufficient accuracy to be useful. On the other hand, Linear Structural Equation models can easily accommodate variables that are the resultant predicted vector from a host of other sources. At the very least, the idea of consequentiality needs to be elaborated into a more complex function reflective of the multiple influences outlined above. For example, Fisher et al. (2013) found that social threat from the participants’ discovery of their partner’s infidelity was a threat to the rememberer. A gender by infidelity type interaction resulted in their discovery narratives. However, Lyons Fisher, Charron, and Bohannon (2015) tested 112 participants about their discovery that a friend was a victim of infidelity. When the infidelity event did not threaten the rememberer directly, the recall narratives failed to discriminate between gender and infidelity type. That result could only be predicted from a confluence of the social vs. physical and the self vs. other dimensions. A single consequentiality factor, as used by Conway (1995), would be unlikely to predict that complex result. The mechanisms that produce public FBMs are also continually selecting important life events for encoding. We argued that accidents with cars or household implements, kisses, sex, and breakups are all personally consequential, vividly memorable over long periods of time, and held with confidence. Just because such memories did not all happen together across the world (e.g. on 9/11), allowing FBM scientists a known point of encoding, is not a reason to classify them as normal autobiographical events in contrast to public FBMs. It is an artificial distinction. Lastly, since the same encoding selection process operates continually throughout life, and FBMs are just one product of that process, continuing to argue about the “special” nature of FBMs or their causal mechanisms is a bankrupt endeavor. No, the selection mechanisms are continually screening experiences to discard the majority of day-to-day events. Since these mechanisms also produce the vivid longterm recall labeled FBM, we cannot call them “special.” Since there are a plethora of personal events that are also emotional and vividly recalled with confidence over the life span we cannot call public FBMs “special” because they are but a small example of many remembered events produced by the same selective encoding mechanisms. In fact, a method that allowed participants to self-select a personal

110 Rice et al.

event to recall as a comparison to a public FBM (e.g. Talarico & Rubin, 2003) would classify a kiss or a car accident as a mundane memory. In the view of our model, kisses and car accidents are far from mundane and are produced by the same selective, memory-enhancing functions as FBMs from 9/11.

The main event: Obama’s election We lament that arousal, social group membership, and valence are unlikely to be counter-balanced in real life. The death of Princess Diana and the Challenger explosion were both uniformly aversive, albeit varied in rememberer’s arousal. Marriage proposals were uniformly pleasant and highly arousing. Wouldn’t it be nice if there were a public event that caused a varied emotional reaction in both amplitude and valence? Is there a public event where some fraction of the population might be very happy whereas others may be very frightened and still others not care one whit about the event? The election of America’s first black president seems to fit those criteria. Using this event, the effects of valence can be tested without employing different events of differing inherent valence and differing dates of encoding. In some Americans of conservative attitudes, the election can be considered a disaster with unforeseen catastrophes ahead with a nonwhite president. Fortunately, the emotions associated with aversive autobiographical events tend to attenuate over time, called the fading affect bias (Walker & Skowronski, 2009). As a contrasting effect, the flourishing affect bias happens when an event that was initially only mildly positive or negative may become more intense when the event is later recalled. This might occur once the consequences of the event are fully realized (Ritchie, Skowronski, Hartnett, Wells, & Walker, 2009). In our three dimensional taxonomy, Obama’s election was public, like most classic FBM studies, social, in that no physical damage accrued to anyone due to the election and, lastly, divided in valence depending to which groups the participants belonged. The authors had four predictions. First, nonwhites and political liberals were expected to have more extensive memories of discovering the election results, as they had greater social group membership with a black president (i.e. members of a liberal political outlook and a racial group broadly construed as nonwhite). Second, participants reporting higher levels of affect during the election announcement were predicted to have more extensive memories than those who reported experiencing low affect (i.e. regardless of initial valence). Third, participants processing event facts from the media were expected to remember more facts about the event (i.e. the date of the event), whereas those who received their discoveries from another person were predicted to recall more personal details surrounding their discovery (i.e. where they were during their discovery). The selection mechanism that aids memory at encoding will enhance recall of the current contents of processing (Bohannon et al., 2007). If you are processing the facts of an event from the media, that is what will be enhanced. If talking with another person who provided the discovery information (Obama’s been elected!), one’s personal circumstances will gain the memorial advantage. Although not predicted by the consequentiality model

The consequences of consequentiality  111

offered above, Bohannon et al. (2007) found that the interaction between source and information type to be the largest effect across five different FBM events. Theoretically, researchers have argued whether FBMs are largely the result of differences in selection and encoding (Bohannon & Symons, 1992; and many others) or due to post hoc reconstruction (Neisser & Harch, 1992). Determining what is recalled at an FBM discovery dependent on the conditions of encoding (person or media) would further support the importance of encoding as a core FBM mechanism. Lastly, if we did not account for this source effect then that variance would simply be dumped into statistical estimates of error, making our other tests more insensitive. Fourth, the authors predicted that those with strongly negative affect at the time of the election will show a fading affect bias over time as the world did not end with the election of a black president (i.e. valence). Those participants (nonwhites and liberals) whose initial reaction was very positive will maintain or increase their arousal over time. We expect participants with positive valence to talk more and recount their discoveries to others in line with Rubin and Schulkind’s (1997) prediction that pleasant events were recounted more often than unpleasant events.This specific event did not include the physical vs. social comparison as no participant was injured discovering Obama’s election. However, looking at the double consequentiality comparison of different in- vs. out-group social membership (racial and politics) vs. both pleasant and aversive valence should contribute support to at least two dimensions of our model. There were 656 participants in this study, including undergraduates from Butler University and Winston Salem State University,3 as well as residents of Indianapolis, IN and Baltimore, MD. Ages ranged from 17 to 78 with a mean age of 22. 55% of the participants were white, whereas 44.2% reported being nonwhite (two participants did not report their race). Regarding political affiliation, 34.5% reported being liberal, 39.2% reported being moderate, and 22.3% reported being conservative (26 participants did not list their political affiliation). Participants were also grouped regarding their reported affect (high affect vs. low affect) and how many times they recounted their discovery stories of President Obama’s 2008 election (high recount vs. low recount). Participants completed a memory survey similar to many other FBM studies: a free recall narrative of their discovery of the election results, a detailed set of memory probes and confidence in their answers, ratings of their memory’s vividness, arousal at the time of encoding, an estimate of how many times they recounted their discovery to another, and lastly, a narrative and probes about the facts of the election. The survey took 20–30 minutes to complete and three waves of participants were gathered. 203 participants were gathered within three weeks of the 2008 presidential election. 391 different participants were assessed after four months and 62 more after four years. Each new wave contained groups of white and nonwhite participants as well as groups differing on their politics. The differences across race are detailed in Table 5.1. Differences across delays are detailed in Table 5.2.

112 Rice et al. TABLE 5.1  Means across dependent variables for race

Variables

White N=364

Nonwhite N=290

Age (17–78)

19.11 (4.24)

24.10 (12.5)

Affect rating1 (1–5)

3.04 (1.06)

Arousal rating2 (1–5)

2.91 (1.20)

Excitement rating3 (1–5)

3.16 (1.45)

Est. # of recounts4 (0–99)

2.07 (3.18)

Mean memory score5 (0–3)

.957 (.71)

F (1,631)

P

Notes

49.68

< .0001 Nonwhites tended to be older than whites 4.36 (0.82) 129.23 < .0001 Nonwhites reported greater affect than whites 4.21 (0.94) 218.74 < .0001 Nonwhites were more aroused than whites 4.51 (0.89) 183.11 < .0001 Nonwhites were more excited than whites 6.54 (12.57) 41.97 < .0001 Nonwhites recounted more than whites .776 (.632) 6.01 F(1,584) < .05 Whites recalled marginally more than nonwhites

Notes 1 We used arousal ratings as a measure of affect changeability over time. Robinson (1980) asserted that the intensity of the affect associated with an arousing flashbulb event is most important in FBM, and such intensity can be observed through fluctuations in arousal ratings. Affect was then operationally defined as low (responses less than 3), medium (responses equal to 3), or high (responses greater than 3) for use in other analyses. 2 Arousal was measured by asking participants to indicate how excited or agitated they were on a scale of 1–5, with 1 being “couldn’t have cared less” and 5 being “extremely excited/agitated”. 3 Excitement was measured by asking the individual to rate how excited they were upon learning the election results on a scale from 1–5, with 1 being “very little” and 5 being “completely devastated/ ecstatic”. 4 Recounts were measured by asking the individual how many times they talked about or shared their memory since the election (i.e. “give a number”). 5 Mean memory score was defined as memory elaboration and measured by taking the average of both narrative and probed recall scores. The higher the average, the greater the recall.

The discovery and fact narratives were scored for the presence of flashbulb canonical features (see Julian et al., 2009) and the presence or absence of answers to our detailed probe questions. All of which were scored on a scale ranging from 0–3. As the narratives for each canonical feature increase on the scale, the more elaborate they were. For example, a score of 0 indicated a lack of response, a score of 1 was given for an implied response, and higher numbers were given for more elaboration. The fact narrative portion of the questionnaire was scored in a similar

The consequences of consequentiality  113 TABLE 5.2  Means across dependent variables for delay

Variables

2–3 weeks N=203

Age (17–78)

20.53 (5.6) 20.98 (8.1) 25.57 (19.1)

Affect rating (1–5)

3.30 (1.03) 3.82 (1.17)

3.43 (1.27)

Arousal rating (1–5)

3.04 (1.24)

3.72 (1.2)

3.33 (1.37) 20.36

Excitement 3.57 (1.37) rating (1–5)

3.90 (1.4)

3.40 (1.45)

Est. # of recounts (0–99) Mean memory score (0–3)

10 months N=391

4 years N=62

F (2,631) p 7.38

.221

5.55

2.58 (4.2) 4.48 (8.36) 5.72 (18.43) 54.21 .464 (.21)

.435 (.21)

.324 (.26) 12.77

Notes

Pos.

General results

TABLE 8.1 A comprehensive list of the studies examining positive FBMs in terms of the authors, the valences compared, the presence of a control group,

6) Bohn & Berntsen (2007)

Yes (Neutral condition)

Yes: (participants not alive during WWII)

Control group

Yes (Neutral Positive condition) (Fall of the Berlin Wall good) Negative (Fall of the Berlin Wall bad) Neutral (Fall of the Berlin Wall neither good or bad)

Positive (Liberation/ Personal memory) Negative (Occupation/ Personal memory) Positive (Red Sox fans) Negative (Yankees fans) Neutral (Fans of neither team)

4) Berntsen & Thomsen (2005)

5) Kensinger & Schacter (2006)

Valences compared

Researchers

Sporting event

Historical Yes: Pos. > Neg. > Neu. Important predictor of FBM characteristics Pos. > Neg. > Neu.

Yes: Pos. = Neg.

Yes: Now and Then: Historical Pos. = Neg. Important predictor of FBM characteristics

Is emotional intensity Type of events measured? compared

18 years

23–27 weeks

60 & 65 years

Temporal distance from event at recall

Confidence Accuracy Vividness

Detailed Confidence Consistency Vividness

Detailed Accuracy Vividness

Operationalization of FBMs

(Continued)

Detailed: Pos. = Neg. > Neu. Confidence: Pos. > Neg. & Neu. Consistency: Neg. > Pos. >Neu. Vividness: Pos. = Neg. = Neu. Confidence: Pos. = Neg. > Neu. Accuracy: Neg. > Pos. & Neu. Vividness: Pos. = Neg. > Neu.

Detailed: Pos. > Neg. Accuracy: Pos. = Neg. Vividness: Pos. = Neg.

General results

No Positive (Fans of the winning team) Negative (Fans of the losing team)

9) Talarico  & Moore (2012)

8) Liu,Ying & Luo (2012)

No

Yes (Neutral condition)

Positive (Fans of the winning team) Negative (Fans of the losing team) Neutral (Fans of neither team) Positive (Acceptance to college) Negative (Wenchuan Earthquake)

7) Breslin & Safer (2011)

Control group

Valences compared

Researchers

TABLE 8.1 (Continued)

Yes: Pos. > Neg.

Yes: Pos. = Neg.

No

Sporting event

Historical & personal

Sporting event

Is emotional intensity Type of events measured? compared

2–7 days (Time 1) & 10–11 weeks (Time 2)

1 year (Time 1) & 3 years (Time 2)

4–5 years

Temporal distance from event at recall

Confidence Consistency Vividness

Confidence Consistency Vividness

Vividness

Operationalization of FBMs

Confidence: Pos. = Neg. (Time 1 & 2) Consistency: Pos. = Neg. Vividness: Pos. > Neg. (Time 2)

Confidence: Pos. = Neg. Consistency: Pos. = Neg. Vividness: Pos. increased over time; Pos. = Neg. (at Session 2)

Pos. > Neg. > Neu.

General results

Valences compared

Control group

12) Kraha, Talarico & Boals (2014)

Positive (Assassination of Osama Bin Laden) Negative (See Talarico  & Rubin, 2003, 2007) Neutral (Self-nominated “everyday” event)

Yes (Selfnominated, “everyday” event)

No 10) Koppel, Positive Brown, (Inauguration Stone, of President Coman & Hirst Obama) Neutral (2013) (Landing of Flight 1549 off the coast of Manhattan) No 11) Kraha & Positive Boals (2014) (Acceptance into Greek Life) Negative (See Talarico  & Rubin, 2003)

Researchers

Historical & personal

No

Yes: Historical Pos. = Neg. = Neu.

Historical

Yes Pos. > Neu. Predicted FBM charactersitc

Is emotional intensity Type of events measured? compared

1–3 days (Time 1) 10, 45 or 227 days (Time 2) 368 days (Time 3)

Within 1 week (Time 1) 7 months (Time 2) & 1 year (Time 3)

2.5 weeks (Time 1) & 4 months (Time 2)

Temporal distance from event at recall

(Continued)

Confidence: Pos. > Neg. Consistency: Pos. (Time 2 & 3) > Neg. Vividness: Pos. > Neg. Confidence: Neg. > Pos. = Neu. Consistency: Neg. > Pos. = Neu. Vividness: Neg. > Pos. =Neu.

Confidence Consistency Vividness

Confidence Consistency Vividness

Pos. > Neu.

General results

Consistency

Operationalization of FBMs

Valences compared

Control group Yes: Sporting event Pos. > Neg. Partially explains FBM charactericsts

Is emotional intensity Type of events measured? compared 18 months

Temporal distance from event at recall Detailed Confidence Vividness

Operationalization of FBMs

Importance -> Intensity of emotion -> Rehearsal (personal circumstances) = +Detail +Vividness +Confidence

General results

Notes 1 The studies listed in the table all examine valence in different way (i.e. “objectively” positive, negative, or neutral or “subjectively” positive, negative, or neutral). For simplicity’s sake, all “valenced” events are simply termed positive, negative or neutral events. 2 While the work by Demiray and Freund (2014) examines positive FBMs, they do so primarily in terms of their function.Therefore, it is not included in the table but is discussed in detail below.

No 13) Tinti, Schmidt, Positive Testa & Levine (Ratings according to (2014) positive words, e.g., happiness) Negative (Ratings according to negative words, e.g., sadness)

Researchers

TABLE 8.1 (Continued)

FBMs for positive and negative events  169 TABLE 8.2 Comparing FBMs for positive events to those for negative and neutral events in

terms of vividness, confidence, consistency, and detail FBMs for positive events: Results

Vividness

Confidence

Consistency

Detail

Significantly more. . . No significant difference. . . Significantly less. . .

3 (2)

2 (2)

1 (2)

1 (1)

4 (2)

3 (1)

2 (1)

2 (0)

1 (0)

2 (0)

2 (0)

1 (0)

Note: Non-parenthetical numbers represent comparisons to FBMs for negative events; parenthetical numbers represent comparisons to FBMs for neutral events.

found no difference between the two, and only one study (Kraha et al., 2014) found memories for a positive event to be less vivid relative to memories of a negative event (see Tables 8.1 & 8.2). When comparing the vividness of positive events to neutral events, two studies found that the positive events elicited more vivid memories (Breslin & Safer, 2011; Bohn & Berntsen, 2007) compared to memories for neutral events. Two studies, however, found no difference in terms of vividness between memories for positive and negative events relative to neutral events (Kensinger & Schacter, 2006; Kraha et al., 2014).

Consistency Again, with their “photographic” nature, FBMs were thought to be highly accurate. However, since FBMs are idiosyncratic and not easy to verify, many researchers have relied upon consistency over time as a proxy for accuracy. Out of the five studies comparing the consistency of memories for positive events and negative events, one (Kraha & Boals, 2014) found positive events to elicit more consistent memories, two (Liu,Ying, & Luo, 2012;Talarico & Moore, 2012) found no difference, and two (Kensinger & Schacter, 2006; Kraha et al., 2014) found positive events to elicit less consistent memories compared to memories for negative events. Memories for positive events relative to neutral events, though, follow the same patterns outlined above; two studies found more consistent memories for positive events (Kensinger & Schacter, 2006; Koppel, Brown, Stone, Coman, & Hirst, 2013) while one study showed no differences in consistency (Kraha et al.). (See Tables 8.1 & 8.2.)

Detail Similar to vividness and consistency, FBMs are also commonly believed to include many details. Four* studies compared the extent to which positive and negative events elicited detailed memories. Of those, only one study (Berntsen & Thomsen, 2005) found positive events to elicit more detailed memories compared to negative

170  Stone and Jay

events, two studies (Kensinger & Schacter, 2006;Tekcan, 2001) found no difference, and one study (Ruiz-Vargas, 1993) found less detailed memories for positive events compared to negative events. Only one study compared level of detail in memories for positive events relative to neutral events and found positive events to elicit more detailed recollections of the two (Kensinger & Schacter, 2006). (See Tables 8.1 & 8.2.)

Summary While only preliminary conclusions can be made, the extant research examining FBMs for positive events suggest that, on the one hand, negative events may elicit more confident and consistent FBMs relative to positive events; on the other hand, positive events may elicit more vivid and equally detailed memories relative to negative events. However, given the small sample size of studies, any definitive conclusions surrounding the characteristics of FBMs for positive events and how they differ from negative and neutral events remain elusive. Despite these drawbacks, it does appear that positive events elicit, at least, very vivid, FBMs, and this, perhaps, should not have come as a surprise for a number of reasons. First, positive information has been shown to be highly memorable as it is less threatening and reinforces wellbeing, but also because positive information is simpler and easier to understand (Unkelbach, Fiedler, Bayer, Stegmüller, & Danner, 2008). Second, there is a tendency for people to recall, long-term, a greater proportion of positive events than negative events, and to reinterpret negative events so that they become less negative overtime (Taylor, 1991). Third, there is also a superiority of positivity in the way we communicate. This is reflected in our language and its structure; positive words are used with greater frequency than negative ones (at least in Western societies), and the former are more primary and more representative of the totality of the concept or dimension (that is, happy is more frequently used, simpler, and represents the whole dimension of positive valence, while unhappy is secondary, more complex, and only represents a portion of its negative dimension (Matlin & Stang, 1978). Moreover, negative information is generally minimized or reconstructed in interpersonal communication (Taylor, 1991). These psychological tendencies tend to be stronger in rituals, monuments, and other cultural products; what is negative for the social group is rarely commemorated, and what is recalled are events or facts that glorify the group as heroes or martyrs (Páez & Liu, 2012; see also Páez et al., Chapter 11, this volume; Pennebaker & Basanik, 1997; Pennebaker, Páez, & Rimé, 1997). Given the evidence demonstrating FBMs for positive events, the ubiquity of these events, and the general mnemonic benefits associated with positive information and events, one pertinent question naturally arises: Why, then, has the research on FBMs focused, almost exclusively, on negative events?

”Why so serious?” A negative bias in FBM research? In answering this question, we will examine two (not mutually exclusive) possibilities for this bias: 1) negative events provide a context to examine traumatic

FBMs for positive and negative events  171

memories and, in turn, post-traumatic stress disorder (PTSD) and 2) negative events, in and of themselves may, in general, be more distinctive and, therefore, garner more attention. We will briefly discuss each of these possibilities.

Flashbulb memories as traumatic memories? Some researchers have argued that negative events provide an opportunity to better understand the formation of traumatic memories and, in turn, PTSD (Galea et al., 2002; Hirst & Phelps, 2016; Sierra & Berrios, 1999; see also Tat, Gold, & Budson, Chapter 7, this volume). Indeed, on the surface, there are similarities between personally traumatic memories and FBMs as a result of negative, public events (Conway, 1996). For example, both types of memories tend to be vivid, persistent, and can elicit some form of anxiety (e.g. fear). However, if we take the terrorist attacks on 9/11 as an example, this comparison appears to be misguided. Although there is evidence suggesting that 9/11 increased the levels of stress for Americans outside of New York, Pennsylvania, and Washington D.C., the extent to which this “stress” relates to post-traumatic stress disorder is unclear (Shuster et al., 2001). For example, a study by Silver, Holman, McIntosh, Poulin, and Gil-Rivas (2002) found that 17% of the US population outside of New York City exhibited post-traumatic stress two months after the attacks, but only 5.8% did six months later. Schlenger et al. (2002), alternatively, only found PTSD symptoms in 4% of the population one to two months after the attacks. This led the authors to argue that the level of distress as a result of the 9/11 attacks, for those Americans outside of New York City, was “within expected ranges for a general community sample” (Schlenger et al., 2002, p. 585). Even within Manhattan, one to two months after the attacks, Galea et al. (2002) found that only 7.5% of local respondents, and only 3.5% of respondents from across the rest of the country, exhibited PTSD symptoms. Thus, even for an arguably quintessential FBM event like 9/11, the negative psychological repercussions, outside directly experiencing the event, appear to be minimal. And while this refrain connecting FBMs to traumatic memories persists as an aside in much of the FBM literature, most FBM researchers do not couch their results in terms of traumatic memories or PTSD (see Hirst & Phelps, 2016 for a review).

Negative events are distinctive? We contend that the focus on negative events in examining FBMs stems from two, interconnected biases: 1) the pleasantness bias and 2) the selection bias. In terms of the former, humans, by and large, are happy creatures. In a study by Chwalisz, Diener, and Gallagher (1988), the general moods of healthy and disabled students were reported. Both populations reported being happy 50% of the time while only unhappy 22% of the time. This positivity extends to the recall of memories as well (Waldfogel, 1948), though this is less surprising considering the robust mood dependency effects (i.e. memories matching one’s mood are more accessible). Positive memories even maintain their affect longer than negative memories, in what is known as the fading affect bias (see, e.g. Walker, Skowronski, & Thompson, 2003).

172  Stone and Jay

The implications of this pleasantness bias in terms of FBM research, we contend, are two-fold. First, this pleasantness bias makes negative events more distinctive or unique (Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001). Or, to put it another way, negative events have a greater likelihood of “bring[ing] about a fundamental change in the fabric of daily life” (Svob, Brown, Reddon, Uzer, & Lee, 2014, p. 249). As a result, negative events tend to draw individuals’ attention relative to positive events (see Baumeister et al., 2001 for a review). This, in turn, leads to the selection bias; given negative events’ distinctiveness and penchant for disrupting our daily lives by grabbing our attention, they are more likely to be selected as an area of inquiry. That is, researchers and laymen alike may have a lower threshold for when negative events capture their attention and this, in turn, leads to greater efforts on the part of society and researchers to study their mnemonic consequences at the expense of interest in, and inquiry of, the mnemonic consequences of positive events.

Summary In general, the copious amount of research focusing on FBMs for negative events appears to stem from a bias towards the “distinctiveness” of said negative events. While some researchers may claim that FBMs serve a unique proxy for traumatic memories, the extant research does not bear this out. “FBMs as traumatic memories” is a false equivalency. However, if positive events elicit FBMs, and if the burgeoning research examining FBMs for negative events is simply the bi-product of a distinctiveness bias, what, if any, unique biopsychosocial functions might positive and negative FBMs serve? We will begin to answer this question next.

Do positive and negative Flashbulb memories serve unique biopsychosocial functions? To begin to answer this question, given that FBMs are a type of autobiographical memory (e.g. “where were YOU when 9/11 occurred?”), it may be best to begin with the posited functions of autobiographical memories. Autobiographical memories are identity forming and personally relevant memories, typically episodic in nature (though the case could be made for some semantic memories, e.g. one’s birthday), and are argued to serve three primary functions; the self (e.g. Bluck & Alea, 2008; Conway, 1996; McAdams, 1995; Pillemer, 1992; Ross & Wilson, 2003; see also Bluck & Levine, 1998), social (e.g. Alea & Bluck, 2003; Marsh, 2007; Nelson, 2003), and directive (Kuwabara & Pillemer, 2010; Pillemer, 2003). First, the self function of autobiographical memories stems from their role as creating and maintaining an enduring sense of self or autobiographical identity (Bluck & Alea, 2008; Conway, 1996; McAdams, 1995; Pillemer, 1992; Ross & Wilson, 2003; see also Bluck & Levine, 1998). Autobiographical memories provide the scaffold through which individuals are able to maintain a coherent and corresponding identity (Conway, Singer, & Tagini, 2004). Second, the social function allows

FBMs for positive and negative events  173

individuals to foster bonds and develop relationships with other individuals (Alea & Bluck, 2003; Marsh, 2007; Nelson, 2003). By sharing both individual and group experiences, individuals and groups may begin to develop shared representations of the past, and these shared representations may have important implications for both individual and collective identity as well as self-esteem (Alea & Bluck, 2003; Fivush, Haden, & Reese, 1996, 2006; Marsh, 2007; Nelson, 2003). Last, the directive function ensures that individuals learn from their previous experiences. Autobiographical memories serve as examples of how to solve problems, avoid particular situations, and shape how individuals behave in the future (Kuwabara & Pillemer, 2010; Pillemer, 2003; see also Baumeister et al., 2001). In other words, autobiographical memories help maintain a sense of self over time, foster and maintain social relationships, and help ensure that we make the best possible decisions now and in the future. Critically, however, research suggests that how autobiographical memories and, in turn, FBM serve these functions may depend on their valence. For example, Rasmussen and Berntsen (2009) compared five different types of memories; participants’ most positive and most negative memories, their most frequent involuntary and most vivid FBMs, and a control (i.e. ordinary) memory. These memories were compared in terms of response time, function, memory characteristics, and the centrality of the event to the participant’s identity and life story. Positive memories were rated higher in terms of self and social functions as well as higher in terms of ratings of phenomenological characteristics. These results dovetail nicely with research by Bilbao, Páez, Da Costa, and Martinez-Zelaya (2013) and Ryff (1989), where they found that positive events were more strongly associated with Ryff ’s (1989) psychological well-being. Research has also shown how positive events are more likely to enhance positive beliefs about the self and the world relative to the ability of negative events to erode these beliefs. Indeed, “the existence of positive emotions matters much more than the absence of negative ones” (Helliwell, Huang, & Wang, 2015, p. 21).Thus, under some circumstances, good is stronger than bad (Bilbao et al., 2013; Catlin & Epstein, 1992). Alternatively, Rasmussen and Berntsen (2009) found that negative memories were rated higher in terms of directive functions (see Echterhoff & Hirst, 2006, for similar results). Interestingly, FBMs were rated significantly higher on the social function relative to the participants’ most positive, most negative, and most involuntary memories, as well as their memories from the previous week. The importance of FBMs in terms of social functions has also been captured in other research. Indeed, the mnemonic consequences of flashbulb events are socially circumscribed. For example, while the French may have FBMs for the death of former President of France, Francois Mitterrand, French-speaking Belgians do not (Curci et al., 2001). And this social identification boundary can be seen throughout the FBM literature: Africans Americans vs. European Americans for the assassination of Malcom X (Brown & Kulik, 1977); Americans vs. Germans for the terrorist attacks on 9/11 (Echterhoff & Hirst, 2006); UK residents vs. non-UK residents for the resignation of Margaret Thatcher (Conway, Anderson, Larsen, & Donelly, 1994); those with ties to the Danish resistance movement vs. those without ties for the German invasion,

174  Stone and Jay

occupation, and liberation of Denmark (see Berntsen, Chapter 9, this volume, for a complete review). Clearly, then, FBMs are intimately intertwined with social identity and vice versa (Hirst & Phelps, 2016). It is unclear, however, the extent to which this social function associated with FBMs is a result of valence, as Rasmussen and Bernsten did not distinguish FBMs in terms of their valence. Regardless, based on the extant literature, one could speculate that positive FBMs may be quantitatively different in terms of their social function relative to other ordinary autobiographical memories. However, recent research by Demiray and Freund (2014) suggests that this may not be case. Demiray and Freund (2014) examined participants’ recall for four different events. These events systematically differed in terms of valence (positive, negative) and “privacy-ness” (private event, public event). The positive, private event was news of one’s own or a loved one’s pregnancy, the negative, private event was the news of a loved one’s sickness or death, the positive, public event was the news of Osama Bin Laden’s death, and, finally, the negative, public event was the news of Michael Jackson’s death. When the researchers compared the participants’ FBMs for these events in terms of their perceived functions, self (self-continuity), social (social bonding), and directive (providing guidance), they found that private memories were perceived as more functional overall, regardless of emotional valence. However, perceived social function was highest for the positive, private memory (i.e. positive autobiographical memories; Alea & Bluck, 2003; Marsh, 2007; Nelson, 2003). Alternatively, the directive function was highest for the private memories, regardless of emotional valence. In general, though, more significant memories (i.e. high in personal importance, emotional intensity, rehearsal and recollective experience) were perceived as more functional (i.e. in terms of self-continuity, social, and directive). Thus, it would appear that FBMs (operationalized as public events) serve less of a biopsychosocial function relative to private events (though, one could argue that the private events in the present study are also public, see, e.g. Mahmood, Manier, & Hirst, 2004). What then becomes clear from this literature is that, while positive FBMs may serve some form of social function, there is no strong evidence to suggest that a) the functions served by positive and negative FBMs differ and, perhaps more importantly, FBMs serve any unique functions above and beyond ordinary autobiographical memories. Clearly, though, this area of inquiry requires more research. We will come back to this issue below.

An evolutionary function? Others, however, have posited that the function of negative FBMs serve a more basic survival function. That is, surprising events may signify a possible survival threat (e.g. a predator). Indeed, when asked to rate personally relevant, positive and negative memories, individuals are more likely to rate positive memories as less surprising than negative memories (e.g. Berntsen, 2002; Rubin & Kozin, 1984;Tromp, Koss, Figueredo, & Tharan, 1995). Therefore, when a “surprising” event occurs, this may lead to enhanced memory processes and better encoding of the surrounding

FBMs for positive and negative events  175

environment (see also Brown & Kulik, 1977). To this end, Kock, Chatelain-Jardon, and Carmona (2008) examined whether a surprising event would facilitate the encoding of the context in which a (negative) surprise took place. In Kock et al.’s study, all participants were asked to review online learning materials in six learning modules. For participants in the “surprise” condition, a picture of a snake in an attack position along with hissing noises suddenly appeared for ten seconds between modules 3 and 4. After all modules were completed, participants took a memory test that asked three multiple-choice questions about each module. Participants in the treatment condition performed significantly better than those in the control condition, but only in terms of the items surrounding the “surprise zone” ’ That is, exposure to the snake image and hissing noise improved performance for the test items from module 3 and module 4. In contrast, the two groups did not differ in memory performance before or after the surprise zone (i.e. in modules 1, 2, 5, or 6). From these results, Kock et al. (2008) concluded that natural selection designed our brains to better encode contextual information surrounding surprising events. However, there are a number of issues with their premise and their interpretation of their results. First, FBMs occur even when the event is not surprising (see Curci & Luminet, 2009 for a review). Second, surprising events can also be positive (e.g. awards, proposals, elections, etc.). It is not clear how such instances would fit within their framework. Third, “surprise” may not have caused the “enhanced” encoding. Rather, the “surprising” event may simply have provided a mnemonic landmark by which to better recall the information both before and after the “surprising” event (Shum, 1998). Regardless, there are reasons to believe that, at the very least, negative FBMs may stem from some evolutionary process. Negative events may draw our attentional resources because the cost of losing opportunities or resources is less than failing to perceive threats. Thus, humans may be drawn to negative events because it has been selected through evolution as an adaptive trait (Baumeister et al., 2001). Thus, though difficult to test empirically, the possibility that FBMs serve some evolutionary function remains an open debate.

Summary In general, the exact and/or unique biopsychosocial function of negative or positive FBMs serve remains unclear. However, there is not enough evidence to suggest that negative FBMs serve a more important or unique function than positive FBMs. Additionally, while some positive FBMs may serve social functions and negative FBMs may serve directive or survival functions, the extent to which their functions differ from autobiographical memories also remains an open question. However, in extrapolating biopsychosocial functions to FBMs, it may be worth noting that FBMs are, as mentioned before, forged from public, collective events. Thus, a major difference between autobiographical memories and FBMs may stem from the latter’s inherent collectiveness (Luminet & Spijkerman, in press; Neisser, 1982). In other words, the function of FBMs may include the aforementioned functions, but

176  Stone and Jay

also serve more collective functions (e.g. collective identity, Hirst & Phelps, 2016; see also Hirst & Meksin, Chapter 10, this volume). The extent to which this is true may necessitate a novel approach to understanding the functions of both positive and negative FBMs. In particular, the “social” and “directive” function generally examined in terms of autobiographical memories may not properly capture the totality of the more collective aspects associated with the formation and maintenance of FBMs. We discuss this possibility below in terms of avenues for future research.

Concluding thoughts and avenues for future research What can we distill in terms of positive events, negative events, and the formation of FBMs and their biopsychosocial functions? Given the dearth of research directly comparing FBMs for positive and negative events, any definitive conclusions remain elusive. Despite this, some general points can be made. First, positive events and negative events may both elicit FBMs, but the characteristics of these FBMs differ. Positive events tend to elicit more vivid memories while negative events tend to elicit more confident and consistent memories. Alternatively, there appears to be no difference in terms of valence in the number of details recalled (again, see Tables 8.1 and 8.2). Second, using negative FBMs as a proxy for traumatic memories is a false equivalency. Third, negative events may be more distinctive and this may be evolutionarily advantageous (Baumeister et al., 2001), but moving forward, the corpus of research suggests that FBMs for both positive and negative events is needed to better understand the mechanisms driving them and their phenomenological characteristics. Last, the perceived biopsychosocial functions of both negative and positive FBMs is still poorly understood. At the very least, they seem to serve the same functions as ordinary autobiographical memories. With these points in mind, the question moving forward should not be whether positive events elicit FBMs but, rather, what are the systematic differences across positive and negative memories in terms of their phenomenological characteristics and their functions. All of these aspects should be examined at both the autobiographical and collective level (Neisser, 1982; see, also, Hirst & Phelps, 2016). In particular, we believe the following research avenues are ripe for future exploration: 1) Before any definitive claims can be made in terms of the similarities and differences across positive and negative FBMs, more systematically rigorous research is needed whereby the characteristics of the positive and negative events are better controlled. For example, in the present body of research examining FBMs for positive events, six studies compared memories of historical events,3 three compared historical events to personal events4 and four compared memories of sporting events from opposing perspectives. Additionally, the temporal distances across studies range from one day to 65 years (see Table 8.1). Of the six studies examining memories for historical events, two measured recall

FBMs for positive and negative events  177

within three years of the events (Koppel et al., 2013; Kraha et al., 2014), three measured recall between (roughly) ten and 20 years after the events (Bohn & Berntsen, 2007; Ruiz-Vargas, 1993; Scott & Ponsoda, 1996) and one study measured recall more than 20 years after the events (Berntsen & Thomsen, 2005). By systematically controlling for these issues both within and across studies, we believe a clearer picture will emerge surrounding mechanisms driving positive and negative FBMs (see, e.g. Berntsen & Thomsen, 2005; Bohn & Berntsen, 2007) and their subsequent, (tentative) diverging characteristics (i.e. vivid positive FBMs vs. confident negative FBMs). Until then, any direct and systematic comparison across valence and/or studies becomes near impossible. 2) Across the studies described here, intensity fluctuated both in terms of valence and its role in forging FBMs (see Table 8.1). However, in addressing 1), the role of emotional intensity may also become clearer. For now, further research is needed. 3) A more nuanced approach to valence is needed. Much of the research presented here couched their studies in terms of positive, negative and neutral. However, not all positive and negative events are equal. For example, a sad event (e.g. 9/11) compared to a shameful event (e.g. My Lai Massacre) may have different mnemonic trajectories and elicit different types of FBMs. Similarly, a joyful event (e.g. President Obama being elected President) and a prideful event (e.g. legalization of same-sex marriages) may also have different mnemonic trajectories and elicit different types of FBMs. Further research is needed to better understand these possibilities. 4) What biopsychosocial functions do FBMs for positive and negative events serve at a collective level? Might they serve a more collective function not currently captured by the functions commonly examined in autobiographical memory research, or do they simply serve the same functions as ordinary autobiographical memories? Researchers should examine this possibility by crafting questionnaires that may better measure collective functions of memories. In doing so, a clearer picture of the difference between FBMs and autobiographical memories in terms of valence and their function will emerge. While the research examining FBMs is vast (as the present volume attests) the answers to these questions will push FBM research to new and exciting realms and provide a better understanding of the ways in which FBMs, in particularly, positive FBMs, differ and/or are similar to ordinary autobiographical memories, both in terms of their phenomenological and biopsychosocial functions. More importantly, it will provide a more nuanced understanding of how both positive and negative events shape individual and collective memories and, in turn, may shape individual and collective identities, respectively.

Notes 1 The positivity of any given event is largely dependent on any given individuals’ perspective (e.g. ideology, values, social identity, etc.) and the size of the group impacted by these positive events also varies; from an individual learning of their acceptance to college (e.g.

178  Stone and Jay

Tekcan, 2001) and a city’s sports team winning a national championship (e.g. Breslin & Safer, 2011), to citizens of a nation learning about the inauguration of a new leader (e.g. Koppel et al., 2013). 2 The * here and elsewhere highlights the fact that one of the studies that examined this characteristic (Tinti et al., 2014), did not include a valenced event as a point of comparison. Thus, while this study is included in Table 8.1, it is not included in Table 8.2 for clarity’s sake. 3 In some of these studies, the researchers also include personal events. We included them under the heading of “historical events” as their primary research question involved the historical events themselves. The personal events merely represented the control items. This is compared to the studies we have included in our table as “historical and personal” event types where the direct comparison is between a historical event and a personal event. 4 It is worth noting that in each study comparing across a historical and personal event, the historical event is always negative and the personal always positive.

References Alea, N., & Bluck, S. (2003). Why are you telling me that? A conceptual model of the social function of autobiographical memory. Memory, 11, 165–178. doi: 10.1080/741938207 Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than good. Review of General Psychology, 5, 323–370. doi:10.1037/1089–2680.5.4.323 Berntsen, D. (2002). Tunnel memories for autobiographical events: Central details are remembered more frequently from shocking than from happy experiences. Memory & Cognition, 30, 1010–1020. doi: 10.3758/BF03194319 Berntsen, D., & Thomsen, D. K. (2005). Personal memories for remote historical events: Accuracy and clarity of flashbulb memories related to World War II. Journal of Experimental Psychology: General, 134, 242–257. doi: 10.1037/0096–3445.134.2.242 Bilbao, M. A., Páez, D., Da Costa, S., & Martinez-Zelaya, G. (2013). Extreme life-changing events, impact on basic beliefs and post-stress growth:Why positive events reinforce eudemonic wellbeing more strongly than negative events undermine it. Monograph Positive Psychology. Terapia Psicologica, 31, 127–139. doi: 10.4067/S0718–48082013000100012 Bluck, S., & Alea, N. (2008). Remembering being me:The self continuity function of autobiographical memory in younger and older adults. In F. Sani (Ed.), Self continuity: Individual and collective perspectives (pp. 55–70). New York: Psychology Press. Bluck, S., & Levine, L. J. (1998). Reminiscence as autobiographical memory: A catalyst for reminiscence theory development. Ageing and Society, 18, 185–208. doi: 10.1017/ S0144686X98006862 Bohn, A., & Berntsen, D. (2007). Pleasantness bias in flashbulb memories: Positive and negative flashbulb memories of the fall of the Berlin Wall among East and West Germans. Memory & Cognition, 35, 565–577. doi: 10.3758/BF03193295 Breslin, C. W., & Safer, M. A. (2011). Effects of event valence on long-term memory for two baseball championship games. Psychological Science, 22, 1408–1412. doi: 10.1177/0956797611419171 Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. doi: 10.1016/0010–0277(77)90018-X Catlin, G., & Epstein, S. (1992). Unforgettable experience: The relations of life events to basic beliefs about self and world. Social Cognition, 10, 189–209. doi: 10.1521/ soco.1992.10.2.189

FBMs for positive and negative events  179

Chwalisz, K., Diener, E., & Gallagher, D. (1988). Autonomic arousal feedback and emotional experience: Evidence from the spinal cord injured. Journal of Personality and Social Psychology, 54, 820–828. doi: 10.1037/0022–3514.54.5.820 Conway, M. A. (1996). Autobiographical memories. In E. Bjork & R. Bjork (Eds.), Handbook of perception and cognition:Vol. 10. Memory (pp. 165–194). Conway, M. A., Anderson, S. J., Larsen, S. F., & Donelly, C. M. (1994).The formation of flashbulb memories. Memory & Cognition, 22, 326–343. doi:10.3758/BF03200860 Conway, M. A., Singer, J. A., & Tagini, A. (2004). The self and autobiographical memory: Correspondence and coherence. Social Cognition, 22, 491–529. doi: 10.1521/ soco.22.5.491.50768 Curci, A., & Luminet, O. (2009). Flashbulb memories for expected events: A test of the emotional-integrative model. Applied Cognitive Psychology, 23, 98–114. doi: 10.1002/acp.1444 Curci,A., Luminet, O., Finkenauer, C., & Gisle, L. (2001). Flashbulb memories in social groups: A comparative test-retest study of the memory of French President Mitterrand’s death in a French and a Belgian group. Memory, 9, 81–101. doi: 10.1080/09658210042000120 Demiray, B., & Freund, A. M. (2014). Michael Jackson, Bin Laden and I: Functions of positive and negative, public and private flashbulb memories. Memory, 23, 487–506. doi: 10.1080/09658211.2014.907428 Echterhoff, G., & Hirst, W. (2006). Thinking about memories for everyday and shocking events: Do people use ease-of-retrieval cues in memory judgments? Memory & Cognition, 34, 763–775. doi: 10.3758/BF03193424 Fivush, R., Haden, C. A., & Reese, E. (1996). Remembering, recounting, and reminiscing: The development of autobiographical memory in social context. In D. Rubin (Ed.), Remembering our past: Studies in autobiographical memory (pp. 341–359). Cambridge, UK: Cambridge University Press. Fivush, R., Haden, C. A., & Reese, E. (2006). Elaborating on elaborations: Role of maternal reminiscing style in cognitive and socioemotional development. Child Development, 77, 1568–1588. doi: 10.1111/j.1467–8624.2006.00960.x Galea, S., Ahern, J., Resnick, H., Kilpatrick, D., Bucuvalas, M., Gold, J., & Vlahov, D. (2002). Psychological sequelae of the September 11 terrorist attacks in New York City. New England Journal of Medicine, 346, 982–987. doi: 10.1056/NEJMsa013404 Helliwell, J. F., Huang, H., & Wang, S. (2015). The geography of world happiness. World happiness report, 12–41. Retrieved from http://worldhappiness.report/wp-content/uploads/ sites/2/2015/04/WHR15.pdf Hirst,W., & Phelps, E.A. (2016). Flashbulb memories. Current Directions in Psychological Science, 25, 36–41. doi: 10.1177/0963721415622487 Kensinger, E. A., & Schacter, D. L. (2006). When the Red Sox shocked the Yankees: Comparing negative and positive memories. Psychonomic Bulletin & Review, 13, 757–763. doi: 10.3758/BF03193993 Kock, N., Chatelain-Jardon, R., & Carmona, J. (2008). An experimental study of simulated web-based threats and their impact on knowledge communication effectiveness. IEEE Transactions on Professional Communication, 51, 183–197. doi: 10.1109/TPC.2008.2000345 Koppel, J., Brown, A. D., Stone, C. B., Coman, A., & Hirst, W. (2013). Remembering President Barack Obama’s inauguration and the landing of US Airways Flight 1549: A comparison of the predictors of autobiographical and event memory. Memory, 21, 798–806. doi: 10.1080/09658211.2012.756040 Kraha, A., & Boals, A. (2014). Why so negative? Positive flashbulb memories for a personal event. Memory, 22, 442–449. doi: 10.1080/09658211.2013.798121

180  Stone and Jay

Kraha, A., Talarico, J. M., & Boals, A. (2014). Unexpected positive events do not result in flashbulb memories. Applied Cognitive Psychology, 28, 579–589. doi: 10.1002/acp.3039 Kuwabara, K. J., & Pillemer, D. B. (2010). Memories of past episodes shape current intentions and decisions. Memory, 18, 365–374. doi: 10.1080/09658211003670857 Liu, A.,Ying, X., & Luo, J. (2012). The flashbulb memory of the positive and negative events: Wenchuan earthquake and acceptance to college. International Journal of Social, Behavioral, Educational, Economic, Business, and Industrial Engineering, 6, 738–743. doi: 1999.10/6656 Luminet, O., & Spijkerman, R. (in press). ‘11 November 1918, an exceptional day!’ Flashbulb memories of the First World War Armistice in Belgium from a psychological and a historical perspective. Memory Studies, Special issue on The Social Cognition of History. Mahmood, D., Manier, D., & Hirst, W. (2004). Memory for how one learned of multiple deaths from AIDS: Repeated exposure and distinctiveness. Memory & Cognition, 32, 125–134. doi: 10.3758/BF03195825 Marsh, E. J. (2007). Retelling is not the same as recalling: Implications for memory. Current Directions in Psychological Science, 16, 16–20. doi: 10.1111/j.1467–8721.2007.00467 Matlin, M. W., & Stang, D. J. (1978). The Pollyana principle. Cambridge, UK: Schenkman. McAdams, D. P. (1995).What do we know when we know a person? Journal of Personality, 63, 365–396. doi: 10.1111/j.1467–6494.1995.tb00500.x Neisser, U. (1982). Snapshots or benchmarks? In U. Neisser (Ed.), Memory observed: Remembering in natural contexts (pp. 43–48). San Francisco, CA: Freeman. Nelson, K. (2003). Self and social functions: Individual autobiographical memory and collective narrative. Memory, 11, 125–136. doi: 10.1080/741938203 Páez, D., & Liu, J. H. (2012). Collective memory of conflicts. In D. Bar-Tal (Ed.), Intergroup conflicts and their resolution: A social psychological perspective (pp. 105–124). New York: Psychology Press. Pennebaker J.W., & Basanik, B. L. (1997). On the creation and maintenance of collective memories: History as social psychology. In J. W. Pennebaker, D. Páez, & B. Rimé (Eds.), Collective memory of political events: Social psychological perspectives (pp. 3–20). Mahwah, NJ: Lawrence Erlbaum. Pennebaker, J., Páez, D., & Rimé, B. (1997). Collective memory of Political Events: Social psychological perspectives. Mahwah, NJ: Lawrence Erlbaum. Pillemer, D. B. (1992). Remembering personal circumstances: A functional analysis. In E. Eugene & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb” memories (pp. 236– 264). New York: Cambridge University Press. doi: 10.1017/CBO9780511664069.013 Pillemer, D. B. (2003). Directive functions of autobiographical memory: The guiding power of the specific episode. Memory, 11, 193–202. doi: 10.1080/741938208 Rasmussen, A. S., & Berntsen, D. (2009). Emotional valence and the functions of autobiographical memories: Positive and negative memories serve different functions. Memory & Cognition, 37, 477–492. doi: 10.3758/MC.37.4.477 Ross, M., & Wilson, A. E. (2003). Autobiographical memory and conceptions of self getting better all the time. Current Directions in Psychological Science, 12, 66–69. doi: 10.1111/1467–8721.01228 Rubin, D. C., & Kozin, M. (1984). Vivid memories. Cognition, 16, 81–95. doi: 10.1016/0010–0277(84)90037–4 Ruiz-Vargas, J. M. (1993). ¿Cómo recuerda usted la noticia del 23-F? Naturaleza y mecanismos de los “recuerdos-destello” [How do you remember the 23-F news? Nature and mechanisms of “flashbulb memories”]. Revista de Psicologia Social, 8, 17–32. doi: 10.1080/02134748.1993.10821667 Ryff, C. D. (1989). Happiness is everything, or is it? Explorations on the meaning of psychological well-being. Journal of Personality and Social Psychology, 57, 1069–1081. doi: 10.1037/0022–3514.57.6.1069

FBMs for positive and negative events  181

Schlenger, W. E., Caddell, J. M., Ebert, L., Jordan, B. K., Rourke, K. M., Wilson, D., . . . Kukla, R. A. (2002). Psychological reactions to terrorist attacks: Findings from the national study of Americans’ reactions to September 11. Jama, 288, 581–588. doi:10.1001/jama.288.5.581 Scott, D., & Ponsoda,V. (1996). The role of positive and negative affect in flashbulb memory. Psychological Reports, 79, 467–473. doi: 10.2466/pr0.1996.79.2.467 Shum, M. S. (1998). The role of temporal landmarks in autobiographical memory processes. Psychological Bulletin, 124, 423–442. doi: 10.1037/0033–2909.124.3.423 Shuster, M. A., Stein, B. D., Jaycox, L. H., Collins, R. L., Marshall, G. N., Elliot, M. N, . . . Berry, S. H. (2001). A national survey of stress reactions after the September 11, 2001, terrorist attacks. The New England Journal of Medicine, 245, 1507–1512. doi: 10.1056/ NEJM200111153452024 Sierra, M., & Berrios, G. E. (1999). Flashbulb memories and other repetitive images: A psychiatric perspective. Comprehensive Psychiatry, 40, 115–125. doi: 10.1016/ S0010–440X(99)90115–3 Silver, R. C., Holman, E. A., McIntosh, D. N., Poulin, M., & Gil-Rivas,V. (2002). Nationwide longitudinal study of psychological responses to September 11. The Journal of the American Medical Association, 288, 1235–1244. doi: 10.1001/jama.288.10.1235 Svob, C., Brown, N. R., Reddon, J. R., Uzer, T., & Lee, P. J. (2014). The transitional impact scale: Assessing the material and psychological impact of life transitions. Behavior Research Methods, 46, 448–455. doi: 10.3758/s13428-013-0378-2 Talarico, J. M., & Moore, K. M. (2012). Memories of ‘the rivalry’: Differences in how fans of the winning and losing teams remember the same game. Applied Cognitive Psychology, 26, 746–756. doi: 10.1002/acp.2855 Talarico, J. M., & Rubin, D. C. (2003). Confidence, not consistency, characterizes flashbulb memories. Psychological Science, 14, 455–461. doi: 10.1111/1467–9280.02453 Talarico, J. M., & Rubin, D. C. (2007). Flashbulb memories are special after all; in phenomenology, not accuracy. Applied Cognitive Psychology, 21, 557–578. doi: 10.1002/acp.1293 Taylor, S. (1991). Asymmetrical effects of positive and negative events: The mobilizationminimization hypothesis. Psychological Bulletin, 110, 67–85. doi: 10.1037/0033– 2909.110.1.67 Tekcan, A. I. (2001). Flashbulb memories for a negative and a positive event: News of Desert Storm and acceptance to college. Psychological Reports, 88, 323–331. doi: 10.2466/ pr0.2001.88.2.323 Tinti, C., Schmidt, S., Tsta, S., & Levine, L. J. (2014). Distinct processes shape flashbulb and event memories. Memory & Cognition, 42, 539–551. doi: 10.3758/s13421-013-0383-9 Tromp, S., Koss, M. P., Figueredo, A. J., & Tharan, M. (1995). Are rape memories different? A comparison of rape, other unpleasant, and pleasant memories among employed women. Journal of Traumatic Stress, 8, 607–627. doi: 10.1016/0010–0277(84)90037–4 Unkelbach, C., Fiedler, K., Bayer, M., Stegmüller, M., & Danner, D. (2008). Why positive information is processed faster: The density hypothesis. Journal of Personality and Social Psychology, 95, 36–49. doi: 10.1037/0022–3514.95.1.36 Waldfogel, S. (1948). The frequency and affective character of childhood memories. Psychological Monographs: General and Applied, 62, 1–39. doi: 10.1037/h0093581 Walker, W. R., Skowronski, J. J., & Thompson, C. P. (2003). Life is pleasant – and memory helps to keep it that way! Review of General Psychology, 71, 203–210. doi: 10.1037/1089–2680.7.2.203

9 FLASHBULB MEMORIES AND SOCIAL IDENTITY1 Dorthe Berntsen

Most theories of the formation of FBMs have examined mechanisms at the level of the individual. For example, Brown and Kulik (1977) claimed that a certain level of subjective surprise and consequentiality was indispensable for the development of FBMs. However, conceptualizing causal mechanisms at the level of individual cognition and emotion has produced contradictory findings. In this chapter, I propose that a person’s social identity (i.e. his or her group identifications) is a key factor for the development and maintenance of FBMs. I review evidence that membership in social groups is a robust predictor of FBMs.

In January 1972, my parents had recently bought their first television set. At that time in Denmark, it was not unusual for programs to be cancelled with a “Technical problems, we apologize” statement appearing on the black and white television screen. However, on Friday the 14th of January, the normal evening entertainments were for no obvious reason replaced by three people playing chamber music. The chamber music went on and on with no explanation as to why. It was a nuisance for a child at my age. I had been looking forward to the usual Friday night entertainments. After a while, the chamber music was finally interrupted by the announcement that the Danish King Frederik IX had died, after which the music resumed. I remember that even after this clarification, my brothers and I were still grumpy: How could the death of the king be so important that chamber music had to replace The Persuaders with Tony Curtis and Roger Moore? The next day, Crown Princess Margrethe officially took over the throne in a television transmitted ceremony. She appeared on the balcony at Amalienborg Palace in Copenhagen to receive the greetings from the thousands of gathered people. She entered the balcony immediately after the Prime Minister from the same balcony had yelled the news of the king’s death in all four directions of the compass as the tradition

Flashbulb memories and social identity  183

prescribed. I remember watching this on TV together with my mother while we were having our afternoon coffee and sharing our sympathy for this young princess, now queen, who came out on the balcony, pale and moved, to take up her new duties less than 24 hours after her father had passed away. More recently, on the 26th of December 2003, at least 15,000 people died in an earthquake in the ancient city Bam in Iran. I have no memory of where I was and what I was doing when I first heard this horrible news. Does this mean that the latter, to me personally, was less emotional, less surprising, less important and less consequential than the news of the king’s death when I was nine years old? Following the literature on FBMs, this seems to be what we should assume. However, I think this assumption is wrong. To me personally, both events were quite insignificant. None of them did in any direct way affect me or people close to me. At a personal level, I knew the king and his family just as little as I knew the thousands of victims of the earthquake in Iran. While personally irrelevant, the death of the king was nonetheless central to my social identity as a Dane. In contrast, Iran was geographically and culturally too distant to me for the tragic news of the devastating earthquake to have similar effects. Social identity, rather than personal relevance, is central for the formation and maintenance of FBMs for public events. I will develop and pursue this view in the present chapter. I first review some of the inconsistent findings that have characterized FBM research. I then suggest that many of these inconsistencies may be solved by basing our understanding of FBMs on mechanisms related to social identity.

Conflicting findings in research on FBMs The notion of FBMs was introduced by Brown and Kulik (1977). They defined this phenomenon as “. . . memories for the circumstances in which one first learned about a very surprising and consequential (or emotionally arousing) event” (p. 73). Hearing the news of the assassination of John F. Kennedy was described as “the prototypical case” (p. 73). Thus, even though Brown and Kulik (1977) mentioned FBMs in relation to personal events, their descriptions and prototypical example of FBMs suggest that vivid memories for personal events are nonessential for this phenomenon. I follow their lead here and focus on FBMs for public events. The key issue related to FBMs, according to Brown and Kulik, is not memory for the news event itself, but why we remember our personal circumstances for receiving this news, since “there is no obvious utility in such memories” (p. 74). In addition to dealing with personal circumstances for the reception of public news, FBMs had three outstanding characteristics, according to Brown and Kulik (1977).They were characterized by “an almost perceptual clarity” (p. 73).They were assumed to be relatively accurate, though at the same time “very far from complete” (p. 75). And finally, they were assumed to be highly durable, indeed indelible. As Brown and Kulik wrote: “It is our assumption that the FB memory is always there, unchanging as the slumbering Rhinegold” (p. 83).

184 Berntsen

Operationalizations of FBMs vary greatly across different studies (see Hirst & Phelps, 2016; Thomsen & Berntsen, 2003; Talarico & Rubin, Chapter 4, this volume, for a review). One reason seems to be that Brown and Kulik’s (1977) anecdotal description of the phenomenon allows several possible operationalizations (e.g. focusing on subjective clarity vs. objective accuracy). In their original work, an FBM was present if a person confirmed that he or she remembered the personal circumstances for receiving the news and was able to provide at least one canonical category (such as information on ongoing activity). Other studies have found these operational criteria too lenient and alternatively tested the claims of accuracy and stability by measuring consistency – that is, how well a memory report given immediately after the event agrees with a memory report given at a later time (e.g. Conway et al., 1994; Hirst et al., 2009, 2015; Neisser & Harsch, 1992). Yet, another class of studies has measured only the phenomenological qualities of the memories, such as vividness and details at the time of recall (e.g. Kvavilashvili et al., 2003). A few studies have examined claims of long term durability by measuring memories after very long delays, using a variety of methods (e.g. Berntsen & Rubin, 2006; Berntsen & Thomsen, 2005; Hirst et al., 2015; Tekcan & Peynircioglu, 2002; Winograd & Killinger, 1983). These approaches tap different memory aspects that need not be parallel. For example, people can have vivid FBMs that may be highly inaccurate (e.g. Neisser & Harsch, 1992). Talarico and Rubin (2003, 2007) demonstrated that FBMs differ from memories for a mundane control event only with regard to confidence, not with regard to consistency. Bohn and Berntsen (2007) found that people with a positive vs. negative emotional attitude to the same event showed a different pattern with regard to accuracy and subjective clarity.The positive groups scored higher on clarity, the negative group scored higher on accuracy, consistent with what has been found in laboratory studies using other emotional material (e.g. Bless et al., 1996; Storbeck & Clore, 2005; more details on this study is offered later in this chapter). Such inconsistencies regarding operationalizations call into question whether FBMs are a unified and distinct phenomenon. It further challenges the validity of the FBM notion that not only do researchers use different operationalizations, they also present conflicting findings regarding causal mechanisms for the formation of FBMs, even when the same operationalization has been used (see Talarico and Rubin, Chapter 4, this volume, for a review). As illustrated by Figure 9.1, in Brown and Kulik’s (1977) account certain (non-specified) critical levels of surprise and consequentiality triggered the formation of FBMs through the activation of a special neural mechanism, labeled “Now Print!” – a purported biological mechanism, ensuring an extraordinary detailed encoding of the surrounding circumstances. A narrative (and measurable) correlate of the memory representation is further developed through repeated rehearsal of the event. However, a prerequisite for this rehearsal to take place is the immediate formation of the memory through the Now Print mechanism.Thus, the emotional reaction that triggers the Now Print mechanism cannot be bypassed. As argued by Brown and Kulik (1977) “ . . . the registration of surprise and unexpectedness

Flashbulb memories and social identity  185

in the central nervous system is the first step and the sine qua non of all else” (p. 84). Brown and Kulik (1977) presented only indirect evidence in support of their purported biological mechanisms. Their main evidence was observations of group differences, which is in complete agreement with the central point to be developed in this chapter. Brown and Kulik showed markedly higher frequencies of FBMs among African Americans as compared to white Americans for the news of the assassination of four of political leaders engaged with questions of civil rights.These differences were matched by generally higher ratings of consequentiality among African Americans for the same four events. Surprise was not measured, but nevertheless claimed to be decisive. Brown and Kulik’s (1977) original model has been the target for much critique. One of the strongest and earliest critics was Neisser (1982) who argued against the idea of a special brain mechanism forming enduring, accurate memories at the time of encoding. Alternatively, he suggested that flashbulb-like memories were formed through frequent rehearsal in thoughts and conversations. This rehearsal was motivated by a need of the individual to experience him or herself as part of history: “We remember the details of a flashbulb occasion because those details are the links between our own histories and ‘History’ ” (p. 48). It is not clear from Neisser’s (1982) account how some events get to become “History” in a particular culture and others not. Neisser (1982) did not present empirical evidence for this alternative view, but other studies have shown that rehearsal can be more decisive for the formation of FBMs than emotional involvement at the time of the event (Otani et al., 2005) and also that FBMs may evolve in response to expected, non-surprising events (Berntsen & Thomsen, 2005) and/or events rated low on measures of emotion and consequentiality (Neisser et al., 1996). Because other studies have shown strong effects of emotion and little importance of rehearsal (e.g. Pillemer, 1984), one may assume that the relative importance of emotion vs. rehearsal varies as a function of characteristics of the target event. Several alternatives to Brown and Kulik’s (1977) original model have been presented (e.g. Conway et al., 1994; Er, 2003; Finkenauer et al., 1998; see Luminet & Curci, 2009). Even though these models are all generated from studies in which FBMs have been operationalized in terms of a certain level of consistency, the models do not agree with one another with regard to causal mechanisms. For Conway

Surprise Consequentiality

Now Print

FLASHBULB MEMORY *Life long

Rehearsal

FIGURE 9.1 The

*Vivid and detailed *Relatively accurate

original model of FBM formation introduced by Brown and Kulik (1977).

186 Berntsen

et al. (1994), prior interest and knowledge is the critical variable that instigates the relevant emotional reactions which eventually lead to FBM formation. In contrast, Finkenauer et al. (1998) found that novelty and surprise formed a direct pathway to explain FBMs. These factors were able to lead directly to the formation of FBMs, whereas importance and emotional feeling state supported the formation of FBMs indirectly through recurrent rehearsal. Er (2003) presented an alternative to both of these models by proposing that the initial appraisal of importance or consequentiality is crucial for prompting the emotional reactions that are involved in the formation of FBMs. Luminet and Curci (2009) tested all four models against FBMs for 9/11 in US and non-US participants.They found most support for a revision of the model proposed by Finkenauer et al. (1998), but also that the correlational patterns differed between the two social groups. Faced with such conceptual and empirical inconsistencies, some researchers have argued that FBMs are not different from other autobiographical memories of emotional events (Mahmood, Manier, & Hirst, 2004; Sierra & Berrios, 1999) or from other vivid autobiographical memories (Rubin & Kozin, 1984; see also Talarico and Rubin, Chapter 4, this volume). In the following section I will introduce an alternative understanding that may be able to unify different approaches to the study of FBMs and resolve some of the conflicting findings.

Reconceptualizing FBMs Following theories of categorization (e.g. Rosch, 1978), I suggest that a reconceptualization of FBMs should take its outset in Brown and Kulik’s (1977) prototypical FBM case rather than in their debatable attempts at developing a formal definition. The prototypical case is remembering the personal circumstances for receiving the news of the Kennedy assassination. This case helps identify some important differences between FBMs and other emotional memories. First, an FBM is related to the reception of a public event – thus, an event that is shared within a given community. Second, instead of focusing on the central details (that is, the news event itself) an FBM maintains the surrounding irrelevant details related to the personal context for the reception of this news. Several studies have found that the personal reception context is remembered better over time than information about the news event itself (Bohannon, 1988; Shapiro, 2006; Smith, Bibi, & Sheard, 2003; Tekcan et al., 2003; but see Nachson & Zelig, 2003). Such focus is the opposite of the tunnel memory effect that is often found for autobiographical memories of intensely negative events (e.g. Berntsen, 2002; see Christianson, 1992, for a review). The notion of tunnel memory refers to a tendency to focus exclusively on the most central, emotion related details, which clearly would be information about the news event itself in the case of FBMs. The tunnel memory phenomenon has been found to apply to personal memories for emotional events, thus, an increased focus on peripheral details relative to the central information seems to distinguish FBMs from other emotional autobiographical memories.

Flashbulb memories and social identity  187

Consistent with Brown and Kulik’s (1977) prototypical case, I suggest that the presence of FBMs varies as a function of the social and cultural relevance of the public news event, whereas personal relevance plays a minor role. More specifically, I suggest that the activation of a person’s social identity is a crucial causal factor for the development of a FBM of a given news event.The person has to appreciate that the news has relevance for a social group that he or she identifies with (see Hirst & Meksin, Chapter 10, this volume, for related perspectives; also see Luminet & Curci, 2009). According to social identity theory (Tajfel, 1982; Tajfel & Turner, 1979) and self-categorization theory (Turner, 1982), our self-concepts consist of two subsystems. One is our personal identity dealing with our personal achievements, bodily attributes, competences and so forth. The other is our social identity, consisting of each individual’s identifications with the social groups that he or she perceives themselves as a member of (e.g. Turner, 1982). Perceptions of group membership are shaped by such factors as similarity, common fate, proximity, and shared threat (Turner, 1982, p. 27), for which reason social groups are often segmented in terms of nationality, political affiliation, religion, gender and the like. According to selfcategorization theory, the relative dominance of these two components of the selfsystem will vary as a function of characteristics of the situation (Turner, 1982; see Haslam, Jetten, Postmes, & Haslam, 2009, for a review). Thus, at times, our personal identity may be almost completely dominant in determining our behavior, whereas at other times our social identity determines our behavior, such as when a person is jeopardizing personal interests to support central values of his or her group, or – less favorably – when a sense of personal responsibility is overshadowed by group behavior, such as when fans of opposing sport teams get involved in brutal fights. I base what follows on this standard version of social identity theory, although alternative views can be found (e.g. see Worchel, Morales, Páez, & Deschamps 1998; see Postmes & Branscombe, 2010, for an overview). It is possible to classify oneself as belonging to a certain social class (e.g. the group of Europeans) without a strong sense of identification with that group for the moment.This means that the strength of our identification with a certain group may vary from situation to situation depending on characteristics of the context (Deschamps & Devos, 1998; Haslam et al., 2009). One possibility that I will pursue here is that public events that are perceived to strongly affect a social group in which the individual perceives him- or herself as a member will make identification with this group particularly salient. As a result of this process, the person may develop a detailed and vivid memory of his or her personal circumstances for hearing the news of the event. This memory serves as an illustration of how the news of this event was able to interrupt his or her ordinary (and usually selfserving) activities, and is an expression of group commitment or “belongingness” (Maslow, 1943). Because, like individuals, groups are striving to maintain a positive self-image (Baumeister & Hastings, 1997), long term durability of FBMs to some extent depends on whether the target event is seen to support or to challenge positive values of the group.

188 Berntsen

A graphic illustration of this model is shown in Figure 9.2. The activation of social identity may lead to an emotional reaction formed by either the individual’s subjective appraisal of the relevance of the event or by contagion when observing the reaction of others (such as children observing the reaction of adults). The emotional salience as well as the social relevance of the FBM may lead to increased rehearsal, both individually and in public. However, the rehearsal and commemoration processes may also bypass subjective emotional reactions, as illustrated by the direct arrow from the social identity box to the rehearsal box. This is because some rehearsal and commemoration processes are likely to be instigated at the level of the community – such as television programs of the event, whose frequency will be independent of the initial emotional reaction of the individual (Frijda, 1997; Hirst et al., 2009). The outcome may be memories with the three main characteristics assumed by Brown and Kulik (1977) – that is, memories that are life-long, subjectively vivid and relatively accurate (“though very far from complete”, p. 75). Such memories are important for supporting and maintaining the individual’s identification with the social group concerned, as illustrated by the dotted arrow from the FBM box to the social identity box. It should also be emphasized that while accurately maintained details are a possible property of FBMs in the current model, they are not necessary.Vividness and expressiveness as symbolic representations of group identification are more central characteristics. Details with such symbolic connotations are therefore better maintained than less telling parts of the personal circumstances, which may lead to errors and biases in the memory reports – a point I will illustrate later. Another likely and important aspect is confidence in the memory, which shows high levels of stability across time (Talarico & Rubin, 2003) and is associated with social bonding (Day & Ross, 2014).

Activation of social identity

Emotion -- Appraisal -- Contagion Rehearsal -- Private -- Public

Flashbulb memory *Life long *Vivid and detailed *Relatively accurate

FIGURE 9.2 An alternative model of FBM formation taking activation of social identity

as a causal mechanism.

Flashbulb memories and social identity  189

Consistencies in FBM Research: Group differences The model presented in Figure 9.2 is not necessarily in conflict with previously suggested FBM models (e.g. Conway et al., 1994; Er, 2003; Finkenauer et al., 1998). It does not deny important cognitive and emotional processes at the level of the individual. However, it brings these processes into a social and cultural context that helps to resolve some of the inconsistencies that previous theoretical and empirical work has left unresolved. In order to test the model, I review evidence that membership in social groups shows a consistent effect across a variety of studies that involve different populations and different public events and use different operationalizations and measures of FBMs (such as subjective clarity, consistency and accuracy of the news event). Table 9.1 shows studies that have compared measures of FBMs for the same event between social groups that would be most likely to differ regarding their attitude and level of involvement in the target event. Only published studies satisfying the following criteria were included. First, the table includes only studies that clearly have measured effects of social groups. It therefore does not include studies that have varied mere geographical proximity towards an event on a relatively small scale (e.g. Edery-Halpern & Nachson, 2004; Sharot et al., 2007) because it is doubtful that such variation corresponds to differences in group membership. Second, studies need to have examined FBM for the same public event in two or more social groups. Third, consistent with the broad definitions of FBMs in the literature (see above) studies can include data on either self-reported FBMs, subjective clarity, consistency across two or more measurements, or other measures of accuracy. Fourth, data from the same study are presented only once; reanalyses based on already published group differences therefore are left out. With only a few exceptions, the studies in Table 9.1 have found differences consistent with what one would expect based on the assumption that social identity plays a central role for the formation of FBMs.These group differences do not show uniform associations with measures of cognitive and/or emotional mechanisms (such as self-reported measures of rehearsal and emotion).Thus, they cannot simply be reduced to such standard purported FBM mechanisms. The following is a brief review of the studies in Table 9.1.

Studies examining group differences Brown and Kulik (1977) found substantially more FBMs among African Americans as compared to white Americans for the news of the assassination of four political leaders engaged with questions of civil rights. These group differences on FBM prevalence were matched by higher ratings of consequentiality among African Americans for the same four events. Even though surprise was claimed to be decisive for the formation of FBMs, it was not measured, nor was emotional arousal or amount of rehearsal.

Year

1977 1978

1993 1994 1996 1998 2001 2003 2003

2003

2004

2005

2005 2006

2009

2009

2015

2015

First author

Brown Yarmey

Morse Conway Neisser Wright Curci Er Kvavilashvili

Pezdek

Luminet

Berntsen

Otani Curci

Tinti

Hirst

Hirst

Curci

US vs. non-US

September 11, 2001

New York vs. others

September 11, 2001

Catholic churchgoer vs. Catholic non-churchgoer vs. Evangelical

New York vs. others

September 11, 2001

Pope Benedict XVI resignation

Polish vs. Italian vs. Swiss

Residents vs. others US vs. non-US

Death of Pope John Paul II

Nuclear accident, Japan September 11, 2001

Resistance vs. others

New York vs. others

September 11, 2001

WWII invasion, liberation

Female vs. male UK vs. non-UK Resid. vs. fam. vs. others Female vs male French vs Belgian Residents vs. others UK vs. Italian

Black vs. White US vs. Canadians

Groups

Clarence Thomas hearings Thatcher’s resignation Earthquake, California Hillsborough football disaster President Mitterand’s death Earthquake, Turkey Princess Diana’s death

Death of civil right leaders Kennedy assassination

Event

TABLE 9.1 Studies examining group differences in relation to the formation of FBMs

Self-reported FBM Consistency Consistency Subjective clarity Consistency Consistency Subjective clarity Details and clarity Accuracy for event Details and clarity Accuracy for event Details and clarity Accuracy for event Subjective clarity Consistency Accuracy for event Consistency Accuracy for event Consistency Accuracy for event Consistency Accuracy for event Consistency Accuracy for event Consistency Details

Self-reported FBM Self-reported FBM Details and clarity

Central dep. var.

Yes Yes, restricted to young No Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes No Yes Yes No No No No Yes Yes Yes

Expected effect?

Flashbulb memories and social identity  191

Morse, Woodward, and Zweigenhaft (1993) found that twice as many female than male participants reported FBMs for the US Senate hearings for confirmation of Clarence Thomas as a Supreme Court Justice.These hearings involved an examination of allegations of sexual harassment, for which reason the event was expected to instigate more FBMs among women than men, consistent with the findings.The female participants also showed more knowledge and experience of sexual abuse, but the relation between such knowledge and the presence of a FBM was not analyzed. Wright, Gaskell, and O’Muircheartaigh (1998) found gender differences in relation to the Hillsborough football disaster. Male participants had more vivid memories for the event, in spite of the fact that females rated it as more emotional and important. The fact that the findings of Morse et al. (1993) and Wright et al. (1998) show differential effects of gender, depending on the nature of the FBM event, combined with the fact that most other FBM studies report no gender differences (e.g. see Edery-Halpern & Nachson, 2004), suggests that the findings of the two studies may reflect the activation of a social identity as women and men, in response to the events. Three studies in Table 9.1 examined FBMs for disasters among people who lived in the disaster area vs. others (Er, 2003; Neisser et al., 1996; Otani et al., 2005). All three studies found more FBMs (operationalized in terms of consistency) among residents as compared to others. In addition, Neisser et al. (1996) found no differences regarding frequency of FBMs among residents vs. non-residents with relatives and friends in the affected area, whereas both groups had more consistent memories than non-residents with no relatives or friends in the disaster area.This suggests that factors other than geographical proximity were decisive for these effects. The three disaster studies reported different findings with regard to cognitive and emotion variables mediating the group effects. Three studies have examined group differences in FBMs for the September 11th attack among US participants. Pezdek (2003) compared memory for September 11th between participants from New York vs. California and Hawaii. She found that the sample from New York had better memory for facts related to the event, but gave slightly less detailed descriptions of the reception context as compared to the other two groups. Consistency over time was not measured. Hirst et al. (2009) compared FBMs of September 11th between USA participants from New York and other locations in the USA and found more accurate event memories in the New Yorker group, but no effects on consistency. In a ten year follow up involving the same groups, Hirst et al. (2015) reported no effects of group on any of the included measures. Of the studies included in Table 9.1, these three studies are the ones with the weakest group effects. One likely reason is that the September 11th attack had a strong effect on all Americans, and also took part in other places than New York (i.e. Washington) and thus may have evoked a social identity of being an American more so than being a New Yorker. This interpretation is supported by the fact that studies comparing FBMs for this event across nationalities have shown consistent effects (Curci & Luminet, 2006, Luminet et al., 2004, to be described below).

192 Berntsen

A number of studies in Table 9.1 have compared FBMs for the same event among different nationalities. Luminet et al. (2004) examined immediate memory for facts and ratings of subjective reactions in response to the news of the September 11th terror attack among participants from the US and seven other nationalities. The US sample showed better memory for fact, more background knowledge, and higher ratings of emotions as compared to the non-US participants. Within the non-US nationalities, the Turkish sample scored particularly low on these measures. In a follow-up study, Curci and Luminet (2006) examined memory consistency among the US participants versus five of the other nationalities (the Turkish sample was not included).Very high consistency scores were observed in all groups.The US participants had more accurate memories with regard to facts related to the news event itself and they had more specific memories of the personal context for receiving the news. A non-significant difference regarding consistency was observed in the same direction. A follow-up structural equation model analysis (Luminet & Curci, 2009) showed a different pattern of predictors in these different national groups, with surprise playing a larger role on the US group. Yarmey and Bull (1978) found that US participants tended to have more FBMs than Canadian participants for the assassination of President Kennedy measured via self-reports at 12 years delay, but this difference was only significant for the younger participants (18–22 years) in the two nationalities. No differences between the nationalities were observed on self-reported vividness and details. Using a consistency-criterion for FBMs, Conway et al. (1994) found substantially more FBMs among their UK participants than non-UK participants for the British Prime Minister Margaret Thatcher’s resignation. Curci et al. (2001) found more FBMs for the French President Mitterand’s death among French vs. Belgian participants. In a follow-up, Curci and Luminet (2009) found that the emotional-integrative model predicted FBMs in both groups. Kvavilashvili et al. (2003) found that UK participants had more vivid and detailed FBMs for the death of Princess Diana as compared with an Italian group. In a study of FBMs in relation to the death of (Polish) Pope John Paul II, Tinti, Schmidt, Sotgiu, Testa, and Curci (2009) showed that Polish participants had more consistent memories and more accurate event memory compared to Italian participants, who in turn had more accurate and consistent memories than Swiss participants. These findings held when controlling for religious involvement. Curci et al. (2015) examined FBMs in relation to the resignation of Pope Benedict XVI in three Italian groups: Churchgoer Catholics, non-churchgoer ­ Catholics and Evangelical participants. They found that the Catholic participants had higher levels of accuracy than the Evangelical participants.They also found that the churchgoer Catholic group had higher consistency and reported more FBM details than the other two groups. Berntsen and Thomsen (2005) studied FBMs in relation to the WWII German occupation of Denmark. They found that participants who had reported ties to the Danish resistance had more accurate memories and rated their memories higher on measures of subjective clarity. Their memories were also more rehearsed and associated with stronger affect.

Flashbulb memories and social identity  193

In summary, with a few meaningful exceptions, FBM studies that have systematically examined memory characteristics as a function of membership in social groups have found effects in the expected direction. These group effects do not in any straightforward fashion correspond to the same underlying emotion and cognitive variables (e.g. surprise, rehearsal and affect) measured at the level of the individual. Taken together, the group differences reviewed here therefore support the claim that social identity interacts with the formation of FBMs in ways that extend beyond the individualistic mechanisms with which previous FBM research has been mostly concerned.

Social identity and long-term maintenance of FBMS The effects of social identity on FBMs are likely to be more pronounced with the passage of time due to the selection and biases that are imposed upon the individual memories via processes of rehearsal (such as media coverage) within the social group. It is likely that events that support central values of the group get more media attention and are more commemorated than events that challenge the positive self-image of the group. A couple of studies of FBMs for remote events can be seen to support this view. Hirst et al. (2015) found that event accuracy of the September 11th attacks correlated with self-assessed media exposure. Bohn and Berntsen (2007) compared FBMs between participants with a positive vs. negative emotional attitude towards the fall of the Berlin Wall in 1989. The negative group consisted of former East German communists, who were still members of the socialist party when the study was conducted in 2002, whereas the positive group was a mixture of former West German liberals and East and West Germans with no party affiliations. Despite the fact that the fall of the Berlin wall was found to have substantially more serious consequences for the socialist group, both according to their subjective ratings and according to historical facts (many lost their jobs and were socially degraded, see Bohn & Berntsen, 2007), and in spite of the fact that they answered a number of factual questions more correctly, their personal memories for the reception of the news were rated significantly less subjectively clear and involved less reliving than memories in the liberal/positive group. One likely explanation for these lower ratings of clarity was that the former East German communists also had talked less about the event and considered the event as less central to their personal and national identity as compared with the liberal/positive group. Consistent with this assumption, measures of personal and national identity were highly correlated with subjective memory clarity. Because the fall of the Berlin Wall clearly challenged central values of the socialist group by demonstrating that many of their values were not shared by the majority of East Germans, there may have been a need to downplay the importance of this event in memory in order to maintain the group. This need should have been especially pronounced among individuals who did not give up their group membership but chose to stay in the party under these unfavorable circumstances (Ellemers, Spears, & Doosje, 2002). Thus, the findings make sense under the assumption that the fall of the wall challenged central values

194 Berntsen

of the socialist group, for which reason FBMs for this event were reduced through less rehearsal, and vice versa in the liberal/positive group (Baumeister & Hastings, 1997). In a related study, Berntsen and Thomsen (2005) asked 145 older Danes if they remembered their personal context for receiving the news of the German occupation of Denmark in April 1940 and the announcement of the German capitulation in May 1945. In addition, we asked them for their memory of their most positive and most negative personal event from the time of the occupation (Berntsen & Thomsen, 2005). They were asked to offer written descriptions of these four categories of memories, as well as to rate their characteristics on a number of scales. When we read through these memory descriptions, it soon became obvious that a number of them dealt with the importance of group membership (Danes vs. Germans) and national identity (Berntsen, 2005). For example, many memory reports illustrated that apparent trivial situations could be remembered for decades, if the situation in question had addressed dilemmas of group commitment (e.g. showing the appropriate Danish attitude) vs. kindness to all fellow human beings (including German soldiers). A 73 year old woman reported the following situation as her most negative memory from the period of the war (my translation): I was on the train from Hellerup together with my father sometime in 1940– 1941. A German soldier was sitting opposite to us, and with all my childish attitude I looked at him rather grumpily. He bent forward and said: “Can’t you smile at all?” I could not, but the episode remains so clear in my memory, contemplated through the years – he presumably had children at home in Germany? (Berntsen, 2005) Almost all of our participants reported that they had an FBM for both the news of the German invasion on April 9th, 1940 and the announcement of the liberation on May 4th, 1945. And indeed most of the participants were able to provide rather detailed descriptions of their personal contexts for the reception of these news events. Memory descriptions for the invasion suggested that, in many cases, the news of the invasion did not immediately cause an emotional reaction, consistent with the model presented in Figure 9.2.The appropriate reaction in many cases had to be learned through the observation of others, often authority figures such as school teachers or parents. Frequently, the emotional reaction was described to come delayed as an effect of such social learning processes. In such cases, the emotional reaction might be classified as contagion (see Figure 9.2). Consider a few examples to make the point (my translations): I was a student at the agriculture school . . . We saw the planes and later the ships from the windows. All the students were brought together in the lecture hall and a highly moved vice-chancellor announced that Denmark had

Flashbulb memories and social identity  195

become occupied. Then we all sang the national hymns with a sincerity that probably surprised us all.What made the strongest impression on me was that the vice-chancellor cried.This was the first time in my life that I saw an adult man crying. . . (Male, 81 years old). I had had my confirmation on the 7th of April and I was supposed to have my communion in the cathedral in Aarhus on the 9th of April [the day of the invasion]. I do not remember my parents’ reaction [to the news of the invasion] but my own was a worry that now I might not be able to go to Aarhus (we lived far out in the countryside). Juvenile egoism! Later the same day I realized how horrible the news in fact was, when I was in the cathedral, listening to the grave speech of the priest (Female, 77 years). (Berntsen, 2005) Compared to memories for the invasion, memories for the liberation were scored significantly higher on almost all measures of subjective clarity. They were also rated as more rehearsed and associated with stronger emotion at the time of the event (see Berntsen & Thomsen, 2005). Because various analyses ruled out that this advantage was due to the liberation taking place five years later in the participants’ lives, and because both of the events could be considered as highly important for this particular historical period (of which they represented the beginning and end, respectively), there was no obvious explanation of the advantage of the liberation memory. One possible, although speculative, explanation is that this advantage was due to the fact that the liberation according to historical analyses had been much more commemorated over the years than the invasion and the related fact that this event has played (and still plays) a much more positive role for the Danish national identity than the invasion.The liberation day agrees with central values of the Danish national identity, such as democracy and freedom, whereas the day of the invasion challenges such positive self-perception, because the Danish army gave in after only three hours of resistance and because the Danish government stayed in power and collaborated with the Wehrmacht for following three years (e.g. Kirchhoff, Lauridsen, & Trommer, 2002). In addition to being associated with more subjective clarity, many memory reports of the liberation appeared to be scripted in a way that seemed motivated by the positive historical value of this event in relation to the Danish national identity. Roughly half of the memory reports included one or more of the following details: The Danish flag, removing (and often burning) the dark shades that the Wehrmacht had demanded should cover the windows at night, lighting candles in the windows, finding “goodies” (e.g. tobacco, chocolate, coffee, wine) saved from the time before the war. The symbolic value of the Danish flag in relation to national identity is obvious. Replacing dark shades with burning candles is easily seen as a symbolic victory of light over darkness. Finding goodies saved from the time before the occupation can be seen as a symbolic restoration of a temporarily lost world. Such symbolic meaning may explain why exactly these details were salient in many of

196 Berntsen

our participants’ liberation memories, as illustrated by this report from an 83 year old female (my translation): I was at home, because we knew that something might happen soon. But even though we were prepared for it [the capitulation], we did not fully understand it right away. It was as if we were paralyzed. But then it came. We shouted and cried. Shortly after, my father came in with a wrinkled flag, which he nailed to a broomstick, and then with my father in front we – my mother and I – went out on our balcony and raised the Danish flag. I will never forget that moment. . . . Then we ran into the streets and celebrated with all the others. Around ten o’clock there was some shooting, and we went home. At home my mother had roasted, grinded and made real coffee. In all our windows, she had lit candles. Soon my home was filled with people, known as well as unknown – and then we were singing, while my brother was playing the piano, and we continued till 3 o’clock in the morning. May 4th and 5th [the liberation dates] are among the greatest memories from my life. (Berntsen, 2005) Such symbolic meaning of details may help to explain distortions in FBM reports. Even though most participants were able to provide accurate reports of the weather on the days of the invasion and liberations, systematic biases were also found. For the invasion, such biases presented the weather as worse than it had actually been (worse operationalized as less sun, lower temperature, more wind, more precipitation than the objective weather data), whereas the opposite was true for the day of the liberation (see Berntsen & Thomsen, 2005, for details). Rather than simply viewing such distortions as reflecting emotions at the level of the individual (light and warmth being standard metaphors for positive emotion, darkness and cold for negative emotions), the effects may reflect the role of the invasion and liberation in the national discourse. The years of the German occupation are often described as “the five dark years” and the day of the invasion was the beginning of the dark years, thus, itself a dark day (although it was bright and sunny according to the objective weather data).The liberation, on the other hand, has been celebrated as the victory of light over darkness.

Summary and concluding comments Ironically, a very nice summary of social identity theory is found in Roger Brown’s (1986) social psychology book. It is surprising that he and James Kulik ignored this theory when developing their model of FBMs. In Brown and Kulik’s (1977) original model of FBMs, the causal mechanisms are found at the level of the individual, in terms of individual experiences of surprise and consequentiality. The same is true for the subsequent revisions of this model. FBM research has thus shared that emphasis on cognitive and emotional processes at the level of the individual that is generally found in cognitive psychology (e.g. see Berntsen & Rubin, 2004; Ellemers et al., 2002, for discussions). At the same time, research into FBMs is characterized by a number of inconsistencies that may be resolved by reference to social factors

Flashbulb memories and social identity  197

that extend beyond cognitive and emotional variables at the level of the individual. Here I have suggested that the activation of a person’s social identity is a crucial factor for the development and maintenance of FBMs. I reviewed evidence that membership in social groups shows consistent effects in most studies that have examined group effects. Factors related to social identity and group maintenance may also be able to account for some biases and distortions in FBMs. In the present review, I have identified social groups post hoc on the basis of the existing FBM literature. Future studies should examine the effects of social identity on the formation of FBMs by identifying groups with different stakes in a major public event and then examine their FBMs for this event. Meta-analytic reviews of multiple studies will allow systematic control for mechanisms at the level of the individual, such as enhanced emotion. An FBM model based on the notion of social identity resolves some of the contradictions that are left unresolved by previous models. As illustrated in Figure 9.2, social identity may enhance memory consolidation and maintenance via three different paths, namely by (1) enhancing the initial emotional reaction to the new event; (2) enhancing initial emotional reaction as well as subsequent rehearsal; and (3) by enhancing rehearsal through public means without any remarkable initial emotional reaction by the subject at the time of the event. Invoking the notion of social identity as a cognitive and social construct therefore allows us to explain why FBMs sometimes occur in response to strong emotion without identifiable effects of rehearsal (Pillemer, 1984), why they sometimes appear to be an effect of both strong emotion and rehearsal (e.g. Finkenauer et al., 1998), and why they sometimes appear to be primarily a rehearsal effect (e.g. Hirst et al., 2015; Otani et al., 2005). The main point is that such traditional FBM factors correlate with the activation of social identity in ways that are partly dependent on the nature of the target event and its cultural and social context. Viewed in isolation, they are not the key mechanisms. FBMs develop as a result of a public event bringing to the foreground a person’s identification with a social group affected by the event. Having an FBM testifies “belongingness” and identification with the group.

Note 1 The work reported here was supported by a grant from the Danish Research Council for Culture and Communication and by the Danish National Research Foundation, DNRF89.

References Baumeister, R. F., & Hastings, S. (1997). Distortions of collective memory: How groups flatter and deceive themselves. In J. W. Pennebaker, D. Páez, & B. Rimé (Eds.), Collective memory for political events. Social psychological perspectives (pp. 277–293). Mahwah, NJ: Lawrence Erlbaum. Berntsen, D. (2002). Tunnel memories for autobiographical events: Central details are remembered more frequently from shocking than from happy experiences. Memory & Cognition, 30, 1010–1020.

198 Berntsen

Berntsen, D. (2005). Lidt skydning i gaderne tog vi ret roligt . . . En psykologisk tematisering af erindringsbilleder fra Danmarks besættelse 1940–45 [A little shooting in the streets didn’t bother us much . . . Autobiographical memories from the German occupation of Denmark 1940–45 from a psychological perspective]. Aarhus, Denmark: Klim. Berntsen, D. & Rubin, D. C. (2004). Cultural life scripts structure recall from autobiographical memory. Memory & Cognition, 32, 427–442. Berntsen, D., & Rubin, D. C. (2006). Flashbulb memories and posttraumatic stress reactions across the life-span: Age-related effects of the German occupation of Denmark during WWII. Psychology and Aging, 21, 127–139. Berntsen, D., & Thomsen, D. K. (2005). Personal memories for remote historical events. Accuracy and clarity for flashbulb memories related to WWII. Journal of Experimental Psychology: General, 134, 242–257. Bless, H., Clore, G. L., Schwarz, N., Golisano, V., Rabe, C., & Wölk, M. (1996). Mood and the use of scripts: Does happy mood really lead to mindlessness? Journal of Personality and Social Psychology, 71, 665–679. Bohannon, J. N. III. (1988). Flashbulb memories for the space shuttle disaster: A tale of two stories. Cognition, 29, 179–196. Bohn, A., & Berntsen, D. (2007). Pleasantness bias in flashbulb memories: Positive and negative flashbulb memories of the fall of the Berlin wall. Memory & Cognition, 35, 565–577. Brown, R. (1986). Social psychology.The second edition. New York: The Free Press. Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. Christianson, S. Å. (1992). Emotional stress and eyewitness memory: A critical review. Psychological Bulletin, 112, 284–309. Conway, M. A., Anderson, S. J., Larsen, S. F., Donnelly, C. M., McDaniel, M. A., McClelland, A. G. R., . . . Logie, R. H. (1994).The formation of flashbulb memories. Memory & Cognition, 22, 326–343. Curci, A., Lanciano,T., Maddalena, C., Mastandrea, S., & Sartori, G. (2015). Flashbulb memories of the Pope’s resignation: Explicit and implicit measures across differing religious groups. Memory, 23, 529–544. Curci, A., & Luminet, O. (2006). Follow-up of a cross national comparison on flashbulb and event memory for the September 11th attacks. Memory, 14, 329–344. Curci, A., & Luminet, O. (2009). Flashbulb memories for expected events: A test of the emotional integrative model. Applied Cognitive Psychology, 23, 98–114. Curci, A., Luminet, O., Finkenauer, C., & Gisle, L. (2001). Flashbulb memories in social groups: A comparative test-retest study of the memory of the French President Mitterand’s death in a French and a Belgian group. Memory, 9, 81–101. Day, M. V., & Ross, M. (2014). Predicting confidence in flashbulb memories. Memory, 22, 232–242. Deschamps, J. C., & Devos, T. (1998). Regarding the relationship between social identity and personal identity. In S. Worchel, J. F. Morales, D. Páez, & J. C. Deschamps (Eds.), Social identity: International perspectives (pp. 1–12). London: Sage. Edery-Halpern, G., & Nachson, I. (2004). Distinctiveness in flashbulb memory: Comparative analysis of five terrorist attacks. Memory, 12, 147–157. Ellemers, N., Spears, R., & Doosje, B. (2002). Self and social identity. Annual Review of Psychology, 53, 161–186. Er, N. (2003). A new flashbulb memory model applied to the Marmara Earthquake. Applied Cognitive Psychology, 17, 503–517. Finkenauer, C., Luminet, O., Gisle, L., El-Ahmadi, A., Linden, M, & Philippot, P. (1998). Flashbulb memories and the underlying mechanisms of their formation: Toward an emotional-integrative model. Memory & Cognition, 26, 516–531.

Flashbulb memories and social identity  199

Frijda, N. H. (1997). Commemorating. In J. W. Pennebaker, D. Paez,. & B. Rimé (Eds.), Collective memory for political events. Social psychological perspectives, pp. 103–127. Mahwah, NJ: Lawrence Erlbaum. Haslam, S. A., Jetten, J., Postmes, T., & Haslam, C. (2009). Social identity, health and wellbeing: An emerging agenda for applied psychology. Applied Psychology: An International Review, 58, 1–23. Hirst, W., & Phelps, E. A. (2016). Flashbulb memories. Current Directions in Psychological Science. 25, 36–41. Hirst, W., Phelps, E. A., Buckner, R. L., Budson, A. E., Cuc, A., Gabrieli, J. D., . . .Vaidya, C. J. (2009). Long-term memory for the terrorist attack of September 11: Flashbulb memories, event memories, and the factors that influence their retention. Journal of Experimental Psychology: General, 138, 161–176. Hirst, W., Phelps, E. A., Meksin, R.,Vaidya, C. J., Johnson, M. K., Mitchell, K. J., . . . Olsson, A. (2015). A ten-year follow-up of a study of memory for the attack of September 11, 2001: Flashbulb memories and memories for flashbulb events. Journal of Experimental Psychology: General, 144, 604–623. Kirchhoff, H., Lauridsen, J. T., & Trommer, A. (2002). Gads leksikon om dansk besættelsestid 1940–1945 [Gad’s encyclopaedia about the occupation of Denmark 1940–1945]. Copenhagen: Gad’s Forlag. Kvavilashvili, L., Mirani, J., Schlagman, S., & Kornbrot, D. E. (2003). Comparing flashbulb memories of September 11 and the death of Princess Diana: Effects of time delay and nationality. Applied Cognitive Psychology, 17, 1017–1031. Luminet, O., & Curci, A. (2009). The 9/11 attacks inside and outside of the US: testing four models of flashbulb memory formation and the specific effects of social identity. Memory, 17, 742–759. Luminet, O., Curci, A., Marsh, E. J., Wessel, I., Constantin, T., Gencoz, F., & Yogo, M. (2004). The cognitive, emotional, and social impacts of the September 11 attacks: Group differences in memory for the reception context and the determinants of flashbulb memory. The Journal of General Psychology, 131, 197–224. Mahmood, D., Manier, D., & Hirst, W. (2004). Memory for how one learned of multiple deaths from AIDS: repeated exposure and distinctiveness. Memory & Cognition, 32, 125–133. Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50, 370–396. Morse, C. K.,Woodward, E. K., & Zweigenhaft, R. L. (1993). Gender differences in flashbulb memories elicited by the Clarence Thomas hearings. The Journal of Social Psychology, 133, 453–458. Nachson, I., & Zelig, A. (2003). Flashbulb and factual memories: The case of Rabin’s assassination. Applied Cognitive Psychology, 17, 519–531. Neisser, U. (1982). Snapshots or benchmarks? In U. Neisser (Ed.), Memory observed: Remembering in natural contexts (pp. 43–49). San Francisco: Freeman. Neisser, U., & Harsch, N. (1992). Phantom flashbulbs: False recollections of hearing the news about Challenger. In E. Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb memories” (pp. 9–32). Cambridge: Cambridge University Press. Neisser, U., Winograd, E., Bergman, E. T., Schreiber, C. A., Palmer, S. E., & Weldon, M. S. (1996). Remembering the earthquake: Direct experience vs. hearing the news. Memory, 4, 337–357. Otani, H., Kusumi, T., Kato., K., Matsuda, K., Kern, R. P., Widner, R., & Ohta, N. (2005). Remembering a nuclear accident in Japan: Did it trigger flashbulb memories? Memory, 13, 6–20. Pezdek, K. (2003). Event memory and autobiographical memory for the events of September 11, 2001. Applied Cognitive Psychology, 17, 1033–1045.

200 Berntsen

Pillemer, D. B. (1984). Flashbulb memories of the assassination attempt on President Reagan. Cognition, 16, 63–80. Postmes, T., & Branscombe, N. (2010). (Eds.). Rediscovering social identity: Key readings. New York: Psychology Press. Rosch, E. (1978). Principles of categorization. In E. Rosch & B. Lloyd (Eds.), Cognition and categorization (pp. 27–46). Hillsdale, NJ: Lawrence Erlbaum. Rubin, D. C., & Kozin, M. (1984).Vivid memories. Cognition, 16, 81–95. Shapiro, L. (2006). Remembering September 11th: The role of retention interval and rehearsal on flashbulb and event memory. Memory, 14, 129–147. Sharot, T., Martorella, E. A., Delgado, M. R., & Phelps, E. A. (2007). How personal experience modulates the neural circuitry of memories of September 11. PNAS, 104, 389–394. Sierra, M., & Berrios, G. E. (1999). Flashbulb memories and other repetitive images: A psychiatric perspective. Comprehensive Psychiatry, 40, 115–125. Smith, M. C., Bibi, U., & Sheard, D. E. (2003). Evidence for the differential impact of time and emotion on personal and event memories from September 11, 2001. Applied Cognitive Psychology, 17, 1047–1055. Storbeck, J., & Clore, G. L. (2005). With sadness comes accuracy; with happiness, false memory: Mood and the false memory effect. Psychological Science, 16, 785–791. Tajfel, H. (Ed.). (1982). Social identity and intergroup relations. Cambridge: Cambridge University Press. Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks/Cole. Talarico, J. M., & Rubin, D. C. (2003). Confidence, not consistency, characterizes flashbulb memories. Psychological Science, 14, 455–461. Talarico, J. M., & Rubin D. C. (2007) Flashbulb memories are special after all; in phenomenology, not accuracy. Applied Cognitive Psychology, 21, 557–578. Tekcan, A. I., Ece, B., Gülgöz, S., & Nurham, E. (2003). Autobiographical and event memory for 9/11: Changes across one year. Applied Cognitive Psychology, 17, 1057–1066. Tekcan, A. I., & Peynircioglu, Z. F. (2002). Effects of age on flashbulb memories. Psychology and Aging, 17, 416–422. Thomsen, D., & Berntsen, D. (2003). Snapshots from therapy: Exploring operationalizations and ways of studying flashbulb memories for private events. Memory, 11, 559–570. Tinti, C., Schmidt, S., Sotgiu, I., Testa, S., & Curci, A. (2009). The role of importance/consequentiality appraisal in flashbulb memory formation: The case of the death of Pope John Paul II. Applied Cognitive Psychology, 23, 236–253. Turner, J. (1982). Toward a cognitive redefinition of the social group. In H. Tajfel (Ed.), Social identity and intergroup relations (pp. 15–40). Cambridge: Cambridge University Press. Winograd, E., & Killinger Jr., W. A. (1983). Relating age at encoding in early childhood to adult recall: Development of flashbulb memories. Journal of Experimental Psychology: General, 112, 413–422. Worchel, S., Morales, J. F., Páez, D., & Deschamps, J. C. (Eds.). (1998). Social identity: International perspectives. London: Sage. Wright, D. B., Gaskell, G. D., & O’Muircheartaigh, C. A. (1998). Flashbulb memory assumptions: Using national surveys to explore cognitive phenomena. British Journal of Psychology, 89, 103–121. Yarmey, A. D., & Bull, M. P. (1978). Where were you when President Kennedy was assassinated? Bulletin of the Psychonomic Society, 11, 133–135.

10 ALIGNING FLASHBULB AND COLLECTIVE MEMORIES1 William Hirst and Robert Meksin

The chapter explores Neisser’s claim that FBMs allow individuals to align their personal memories with history, which here is interpreted in terms of collective memory. An FBM, then, becomes a defining characteristic of membership in a community. Moreover, memory of the event itself is shaped by community practice, thereby aligning individual autobiographical memories with community practices. The implication for the intersection between individual and community identity is explored.

In his reflection on Flashbulb memories, Neisser (1982) wrote: [One] recalls an occasion where two narratives that we ordinarily keep separate — the course of history and the course of our lives — were momentarily put into alignment. . . Details are linked between our own history and History. . .. [Flashbulb memories] are the places we line up our lives with the source of history itself and say, “I was there”. (p. 48) In making this statement, Neisser offers a distinctive perspective on the nature of Flashbulb memories (FBMs). Most researchers examining Flashbulb memories are concerned with the conditions eliciting FBMs and their phenomenology. A few consider the functions of FBMs, mainly discussing the social bonds people form as they share their memories (Demiray & Freund, 2015; Rasmussen & Berntsen, 2009). The Neisser quote points to another function — that FBMs can serve to align the personal with the historical. In this paper, we go a step further than Neisser (1982) and argue that the alignment he identifies is not just about an individual lining up his or her life with history, but

202  Hirst and Meksin

about members of a community doing so.That is, Neisser’s alignment is a communitywide phenomenon. Although the formation of an FBM is no doubt an individual act, it is an action occurring within a community. After all, FBMs are usually not held by one member of the community in isolation. They tend to be held by most members of a community. Their presence does not just allow one to claim “I was there,” but also implicitly to assert “I, along with other members of the affected community, was there.” Moreover, the precipitating event is interpreted and mnemonically represented through the community’s lens. Most Americans interpret the attack of September 11, 2001 as an assault on American values, e.g. freedom of religion. Those in sympathy with the attackers view it as an attempt to regain a role for Islam in the political world and as a fight against American hegemony. That is, interpretation and memory for the event can often depend on the community to which one belongs. Consequently, FBMs are the places where autobiographical memories held across community members line up with community-shaped memories of an historically important event. That is why FBMs should be viewed not just as phenomena on an individual level, but also as deeply and fundamentally community-based. They serve a function not just for individuals, but also for the community as a whole. In what follows, we first provide specificity to Neisser’s concept of alignment, in particular, framing it in terms of collective memory rather than history per se. We then explore the community nature of FBMs, subsequently turn to community efforts to shape collective memory, and finally examine how the presence of an FBM, as a community-based phenomenon, increases the impact that a communityshaped collective memory has on collective identity. In arriving at this last step, we move from the individual to the collective, now connecting autobiographical Flashbulb memories with collective identity.Towards the end of the chapter, we are particularly interested in the way the social practices embedded in media coverage shapes the collective memory of a flashbulb event and, because of the presence of this Flashbulb memory, enhances the influence of collective memory on collective identity.

Approaching the topic of alignment Before we continue exploring the community-wide nature of the alignment between FBMs and history, we need to be clear about our terminology. For us, FBMs refer exclusively to memories of the circumstances in which a person learned about an emotionally charged, public event. The second author’s FBM of 9/11 is that he was working at Columbia University and, upon entering a lobby area, found a crowd raptly attentive to the news coverage on TV sets situated at the corners of the room. He then turned his own focus to the TV to watch the video footage of the second plane crashing into the World Trade Center. As to history, a more appropriate term might be lived collective memories.We mean by collective memory individual memories shared across a community that bear on the community’s identity (Hirst & Manier, 2002, 2008; see Olick, 1999, for an

Aligning Flashbulb and collective memories  203

alternative definition, also Hirst & Stone, 2016). We employ the modifier lived because the events involved in the formation of an FBM occur during the lifetime of the community members sharing the memory, even though the event itself is not directly experienced. Both authors’ recollections of the Fall of the Berlin Wall are lived national collective memories. What they know about the War of 1812 is also a collective memory, just not a lived one. Since Halbwachs’ (1992) groundbreaking work on collective memory, scholars have treated collective memories as distinct from history (see, for instance, Kansteiner, 2002; Nora, 1989; see also Olick,Vinitsky-Seroussi, & Levy, 2011). History is what professional historians write, whereas collective memories are what the community members hold. History must meet various professional standards. For instance, although history is often written to serve a community’s effort to come to terms with the past, in most instances, historians must account for the available facts, at least those fellow historians deem important. Moreover, these facts must conform to agreed-upon evidentiary standards. Collective memories, on the other hand, need not be accurate, may not account for the “important” facts, nor meet any evidentiary standards. Indeed, people can be stubbornly resistant to the available evidence when it comes to some aspect of their collective memories. Holocaust deniers may firmly believe that the Jewish genocide never took place, despite the overwhelming evidence that it did. Historians do not have the same leeway. History, then, connotes something outside the individual, something conceived by others, whereas collective memories are the mental representations individual members of a community have of the past, albeit shared across that community. For us, it seems more appropriate to align the mental representation of an FBM with the mental representation an individual has of a community’s collective memory, than to align it with something outside the individual, like history. That is why we argue that FBMs should be linked to collective memories. Fortunately, it is possible to study this alignment. Many FBM studies probe not only for memories of the circumstances in which participants learned of a public, emotionally charged event, but also for memories of the facts about the event, so-called event memories. For us, the facts assessed in an event memory test could potentially be part of the collective memories people form of the Flashbulb memory-eliciting event. Specifically, we treat a probed-for fact as part of a collective memory if a substantial portion of the sample in a study remembers it. To be sure, most FBM studies use convenience rather than representative samples, making it necessary to proceed cautiously when linking performance on an event memory task with claims about what is representative of a community’s collective memory (Corning & Schuman, 2015; Wright, Gaskell, & O’Muircheartaigh, 1998). Nevertheless, such a linkage seems like a reasonable way to obtain at least a preliminary look at the collective memories of Flashbulb memory-eliciting events. Consequently, in this chapter, we examine whether FBMs are indeed a community phenomenon and, using event memories as our probe, how community efforts might guide the formation of collective memories of flashbulb events.

204  Hirst and Meksin

Flashbulb memories as a community phenomenon Flashbulb memories are reported widely within an affected community As it appears to be used in the literature, a memory is treated as an FBM only if most members of a community report that they have a memory of the circumstances in which they learned of a public, emotionally charged event, such as the Brussels airport attack in March 2016. Memory for the reception event would not be classified as an FBM if only one or two people within a community reported such a memory. To be sure, the content of each FBM may differ from community member to community member, but most community members would report that they have an FBM.This observation underlies our claim that FBMs are not just a characteristic of an individual, but also a community. In order to examine the role FBMs might play when aligning with a collective memory of the event, one must first, then, classify a memory as an FBM by observing its presence in the larger community. Consider what most would treat as a prototypical FBM; the memories Americans have of the circumstances in which they learned of the attack of September 11, 2001. Hirst et al. (2009, 2015) tracked the 9/11 FBMs of a cohort of 202 Americans over a ten year period. They asked participants to respond to six questions, each of which probed a canonical feature Brown and Kulik (1977) assigned to FBMs.These probes essentially assessed the “who, what, when, where, and how” of the circumstances in which the participants learned of 9/11. If one uses the report within the first week as a baseline from which to compare subsequent responses, one can calculate a consistency measure for each response on the three additional surveys Hirst et al. administered, approximately one year, three years, and ten years after the attack. This consistency measure can approximate accuracy, that is, if one assumes that the initial first-week memory is largely accurate (Kvavilashvili et al., 2009, but see ­Winningham, Hyman, & Dinnel, 2000).There was a dramatic decline in consistency within the first year, with an average consistency measure of .61 at the 11th month. After ten years, the consistency measure of Hirst et al.’s participants was .60, suggesting that forgetting tapered off after the first year. Such a pattern of forgetting over the very long-term – that is, initial rapid forgetting followed by a leveling off – has been observed with a wide range of material (e.g. Bahrick, 1984). Clearly, FBMs can undergo forgetting, at least in the first year or so. We are interested here, however, in a different sense of forgetting; that people simply fail to report a memory, not whether it is accurate. There is little evidence for this “failure-to-report” sense of forgetting in Hirst et al. (2009, 2015). 99.5% of the participants answered at least five of their six probes in each of the three surveys reported in Hirst et al., including the survey administered ten years after the attack. Moreover, their participants assigned an average confidence rating of 4.35 out of 5 to their answers, suggesting that they were not merely being compliant and supplying an answer that they suspected might be wrong.

Aligning Flashbulb and collective memories  205

This pervasive, confidently held report suggests that 9/11 FBMs may be characteristic of the American public. Although Hirst et al.’s (2009, 2015) sample is not representative of the US population, it did involve participants from across the country. It seems fair to conclude from their results that FBMs of the 9/11 attack are widespread among Americans who were old enough to appreciate what was happening. Possessing an FBM of 9/11 captures at least one dimension of what constitutes being an American for this age cohort. Most Americans would no doubt find it surprising if a fellow American of a certain age did NOT have an FBM of 9/11.

Flashbulb memories are specific to the affected community One reason why we assert that an FBM can be characteristic of the affected community is that who reports an FBM appears to be community-specific. For instance, British participants in one study were more likely to report, after a delay of 51 months, FBMs for the death of Princess Diana than were Italians (Kvavilashvili et al., 2003). Similarly, French citizens were more likely to have an FBM for the death of President Mitterrand than were French-speaking Belgians (Curci et al., 2001). Moreover, African Americans were more likely to form FBMs of the assassination of Malcolm X than were European Americans (Brown & Kulik, 1977). These differences no doubt reflect, across different communities, alternative appraisals of the importance or consequentiality of the Flashbulb memory-eliciting event (Tinti, Schmidt, Sotgiu, Testa, & Curci, 2009). Luminet and Curci (2009; see also Berntsen, Chapter 9, this volume) also identified “activation of social identity” as another possible factor. We mean here by social identity “those aspects of an individual’s self-image that derive from the social categories to which he perceives himself as belonging” (Tajfel & Turner, 1979, p. 40). Whatever the exact reason for the community differences, there is no doubt that FBMs are often specific to a particular community, and not to others. Of course, caution is needed when describing what constitutes “an affected community.” Brown and Kulik’s (1977) finding that only African Americans, and not European Americans, reported possessing an FBM for the assassination of Malcolm X suggests that the affected community needs not map solely onto entities such as nations. An FBM may, of course, be confined to a community within a nation. Moreover, in some circumstances, the precipitating event may be of importance to many different nations. Studies by Curci, Luminet and their colleagues, for instance, found that FBMs for learning of the 9/11 attack were widespread across countries from North America, Europe, and Asia (e.g. Curci & Luminet, 2006; Luminet & Curci, 2009; Luminet et al., 2004). Finally, the group may be transnational, but not reflect a state structure at all, as illustrated by Tinti et al.’s (2009) study of religious involvement and FBMs of the death of Pope John Paul II. Although it is still unclear how one might specify in advance an affected community, there is little doubt that the bulk of the extant research underscores the community specificity of FBMs.

206  Hirst and Meksin

Community membership shapes characteristic features of Flashbulb memories Like ordinary autobiographical memories, as we noted in the Hirst et al. study (2009, 2015), FBMs can be inaccurate. However, also noted earlier, even when inaccurate, people report a high level of confidence (e.g. Talarico & Rubin, 2003). Indeed, high level of confidence has become a defining characteristic of FBMs (Hirst & Phelps, 2016). Is the characteristic high level of confidence associated with FBMs at least partially community-based? Although people assign a confidence rating to a memory for a variety of reasons, one commonly used feature is the ease with which a memory is retrieved. Echterhoff and Hirst (2006) employed ease of retrieval as a means of studying confidence judgment for FBMs. Experimentally, one can manipulate ease of retrieval by asking participants to remember either 5 or 15 aspects of a to-be-remembered event. Access should “feel” more difficult in the latter condition than the former. Consequently, as the extant experimental evidence indicates, people are less confident in their memory when trying to remember 15 aspects of a memory than when trying to remember five (Belli, Schuman, & Jackson, 1997).With these results in mind, Echterhoff and Hirst asked both Germans and New Yorkers to remember 5 or 15 activities on September 11, 2001, both six months and one year after the event. They were interested in the extent to which their participants used ease of retrieval when making their confidence ratings. Although all participants found the 15-activities task more difficult than the 5-activities task, at six months, the confidence ratings for both Germans and New Yorkers did not differ across the two conditions, demonstrating that, at this juncture, ease of retrieval was not a factor in determining confidence ratings. On the other hand, at one year, Germans began to base their confidence judgments on the ease of retrieval, inasmuch as their ratings were significantly lower in the 15-items than the 5-items condition. Americans, however, showed no signs of similarly using ease of retrieval to guide their confidence judgments. Their confidence ratings were the same across conditions. Echterhoff and Hirst (2006) accounted for the discrepancy between Germans and Americans by appealing to metamemory considerations. Specifically, social norms can dictate when one must remember (see Margalit, 2002, for a discussion of the ethics of memory). As a result, one might assign a high confidence rating to an ethically required memory, regardless of what other cues are suggesting. Such an ethical demand might be present for memories of learning about the attack of 9/11 for the American group. After all, as stated earlier, the data suggests that possessing an FBM of 9/11 is part of what characterizes being American. Americans may feel that they must possess – and confidently possess – an FBM of 9/11 after six months and after a year. To do otherwise would diminish your credentials as an American. Germans may not have held as firmly to such a belief, in that possessing an FBM of 9/11 need not be a defining characteristic of being German, especially as time passes.

Aligning Flashbulb and collective memories  207

The content of Flashbulb memories as a function of community membership Community membership can also play a role in shaping the content of FBMs. Berntsen (2009; also Chapter 9, this volume) studied this aspect of FBMs by examining Danes’ memory of circumstances in which they learned of the German invasion of Denmark during War World II and the subsequent German withdrawal (Berntsen & Thomsen, 2005). Given the difficulty in assessing the accuracy of the reported memories in every aspect, Berntsen and Thomsen cleverly focused on the weather, something precisely documented in the available archives. Danish participants in their study, all of whom had lived through the war, had overall surprisingly good memory for the weather, but the extant errors proved nicely consistent with what one would expect from this population. When the Danish participants made an error about the weather on the day of the invasion, they remembered it as worse than it actually was; when they made an error about the weather during the retreat, they remembered it as better than it was.

Summary FBMs, then, are not simply autobiographical memories of the circumstances in which one learned of a public, emotionally charged event. As we have been asserting, they are community-wide phenomena and, as such, their characteristics and content are, in part, community-based. The specifics of the memory may differ from one community member to another, but members of an affected community will inevitably report having an FBM. Consequently, the possession of this memory becomes a characteristic of this community. Indeed, people are confident in their memory, in part, because it is something they, as a member of this community, are expected to have. Moreover, the content of what they remember may reflect their community membership, in that they sometimes falsely remember the reception event in a manner that conforms with the perspectives of their community.

Community efforts to shape collective memories: Effects on event memories Social memory practices What about the other side of Neisser’s coin – what he referred to as “History,” and what we are calling lived collective memories? It is almost definitional that collective memories are community-based phenomena. More interesting is the question as to how a community shapes the collective memories its members hold. The study of collective memory routinely investigates this issue (see Olick et al., 2011). Particular emphasis is given to the role played by societal practices and artifacts, focusing largely on the influence of monuments, commemorations, texts, and the

208  Hirst and Meksin

media (e.g. Gillis, 1994; Hoskins, 2007; E. Zerubavel, 2012; Y. Zerubavel, 1995). Inasmuch as we are treating widely held event memories within a community as representative of a community’s collective memory, we follow the lead of the collective memory literature and explore the ways in which social memory artifacts and practices affect the event memories associated with flashbulb events. Before discussing social practices and event memories, we first want to discuss the accuracy of event memories. Although it may be surprising, given their strong emotionally charged nature, like FBMs themselves, memory for facts about the event declines over time. Studies differ as to the fate of such event memories after the first year, however. In one of the earliest studies examining event memory, which probed for memories of the Challenger explosion, Bohannon and Symons (1992) found that, whereas the ability to remember accurately essentially stopped declining at the eight-month interval for FBMs, accurate recollection of the facts about the event continued to decline, for instance, by 20% between the 15-month and 36-month intervals (for similar results, see, for instance, Smith, Bibi, & Sheard, 2003; Tekcan, Berivan, Gülgöz, & Er, 2003). On the other hand, in their large-scale, ten-year longitudinal study, Hirst et al. (2009, 2015) found a different pattern. In assessing event memory of the 9/11 attack, they asked questions such as “How many planes were involved?”; “What were the names of the airline carriers”; “Where was President Bush when the attack occurred?”; and “What were the crash sites?” Contrary to the other studies on event memory (Bohannon & Symons, 1992, Smith et al., 2003; Tekcan et al., 2003), in Hirst et al., the decline in accuracy of event memories for 9/11 leveled off after the first year. One week after the attack, participants remembered accurately, on average, 91% of the five questions they were asked about factual details of the terrorist attack, whereas they remembered, on average, 80%, 81% and 81% of them after delays of one, three, and ten years, respectively. It would appear, then, that the sustainability of collective memories of flashbulb events follows two distinct patterns. In some instances, such as the Challenger explosion, the collective memory becomes more streamlined over time, as pertinent facts are forgotten. In other instances, such as 9/11, the collective memory remains stable over time, at least in the sense that facts about the event are not forgotten. This difference provides a means of understanding the role social practices play in shaping collective memories. Several possible accounts of this difference do not make reference to social practices. They offer, at best, only partial explanations. For instance, the two patterns we just articulated cannot be explained simply by appealing to community relevance. Both the Challenger explosion and the 9/11 attack were important to the affected community. Moreover, a slight variation of this hypothesis, offered by Pezdek (2003), also has its limitations. She averred that people from different communities assign different levels of emotional attachment to the event. This claim, however, cannot account for the findings of Hirst et al., in that they found no substantial differences in event memory between participants living in New York City and those who resided in California cities (specifically, San Francisco, Los Angeles,

Aligning Flashbulb and collective memories  209

Palo Alto, and Santa Cruz), which, according to Pezdek, would constitute different communities with different emotional attachments. As to the role of social artifacts and practices, it appears that the retention of event memories depends on the execution of various social practices (although, interestingly, the retention of FBMs may not). Of course, one must have memories of at least some aspects of the event if one is going to have a memory for the circumstances in which one learned of the event. But we would argue that, to a large extent, an individual’s retention of the factual specifics of the flashbulb event (those that finally find their place in a community’s collective memory) appear to be the result of a social effort rather than simply a consequence of the way an individual encodes and stores a memory.

Media attention as a critical social practice When considering social practices, media coverage is of particular interest. Inasmuch as FBMs often involve public, emotionally charged events of national importance, the media usually provide extensive coverage of the precipitating event, at least initially. How the media cover these events and what they cover are quintessential examples of social practices (Bouvier, 2005; Dayan & Katz, 1992; Zelizer, 1992). Television networks can decide what to broadcast and how extensively to cover a topic. In doing so, they must follow certain rules, both those associated with the journalistic profession and those embodied in the legal framework in which media outlets function. But they also can have a political agenda. Whatever motivates the coverage, the resultant editorial decisions, intentionally or not, will affect what people know and remember about an event or a class of events (Hoskins, 2007; Iyengar & Kinder, 1987). In the last year or so, for instance, many critics have raised concerns about American television’s selective coverage of terrorist attacks, which devotes substantial air time to attacks on Western nations, such as the Charlie Hebdo and November 13 Paris attacks, while giving scant air time to attacks in non-Western countries, such as Turkey, Saudi Arabia, Iraq, Afghanistan, Pakistan, and Bangladesh, to name a few (e.g. Darling-Hammond, 2016). This concern is essentially about the practices of the television networks. It arises because there is general recognition that these practices can shape and maintain one collective memory – that the West is under attack – and not another – that much of the Muslim world is also under attack. Although most Flashbulb memory-eliciting events involving an event of national import begin with an explosion of media coverage, in many instances, coverage eventually declines to almost imperceptible levels. In others, it persists for years. Several studies of FBMs and their associated event memories have specifically probed for the extent to which participants attended to the media (e.g. Bohannon & Symons, 1992; Christianson, 1989, 1992; Finkenauer et al., 1998; Koppel, Brown, Stone, Coman, & Hirst, 2013; Luminet & Curci, 2009; Mahmood, Manier, & Hirst, 2004; McCloskey, Wible, & Cohen, 1988; Neisser et al., 1996; Otani et al., 2005; Wright, 1993). These studies generally find a positive correlation

210  Hirst and Meksin

between the self-reporting of the extent of media attention and the accuracy of event memory. Moreover, other studies have shown that the retention of event memories declines as media coverage wanes (e.g. Bohannon & Symons, 1992; Finkenauer et al., 1998; Nachson & Zelig, 2003; Smith et al., 2003; Tekcan et al., 2003; Tinti, Schmidt, Testa, & Levine, 2014; Wolters & Goudsmit, 2005). Most models reflect these correlations by positing that the media attention leads to a rehearsal of the event memory, which in turn leads to its retention (Conway et al., 1994; Curci, & Luminet, 2006; Er, 2003; Finkenauer et al., 1998; see also Roediger, Zaromb, & Butler, 2009). The role of the media can be clearly seen when comparing Bohannon and Symons’s (1992) three-year study of the Challenger explosion with Hirst et al.’s (2009, 2015) ten-year study of 9/11. Figure 10.1 plots the drop in accuracy of event memory over a three-year period for both events and an estimate of the media coverage of each event, as measured using the New York Times index service. We plotted z scores in order to make the accuracy measure and the measure of media coverage comparable. As Figure 10.1 indicates, this extent of media coverage nicely mirrors the rate of forgetting. For the Challenger explosion, both media coverage and accuracy continued to decline during the three-year period, whereas, for 9/11, both media coverage AND accuracy leveled off after the first delay period (approximately one year).

FIGURE 10.1 Relation

between media coverage and memory accuracy for facts about Flashbulb memory events: The attack of September 11th and the Challenger Explosion (From Hirst et al., 2009).

Aligning Flashbulb and collective memories  211

The media and critical facts in collective memories As Figure 10.1 suggests, the media can determine what will or will not be included in lived collective memories, even for the kind of emotionally charged events often associated with FBMs. Hirst et al. (2009, 2015) provided some specificity as to what about the event eventually finds its way into a community’s collective memory by tracking the extent to which participants accurately recalled what Hirst et al. identified in their analyses as critical and non-critical facts. If we posit that the critical facts are those considered essential to telling the story of the event, and the noncritical facts are those that hold less import in the narrative, then two examples of critical facts about 9/11 are (1) the number of planes involved in the attack, and (2) the crash sites. Both of these facts were included in most media renderings of the attack.That is, they received extensive and continuous media coverage. Examples of non-critical facts were (1) the location of President Bush when the attack occurred, and (2) the names of the involved airline carriers. These facts can be, and often are, omitted in accounts of the attack. For instance, the Wikipedia site on 9/11 does not include the location of President Bush. If non-critical facts receive less media coverage, and if the media is rehearsing, and thereby sustaining, memories, one would expect memory for the critical facts to remain robust across the ten-year period, whereas memory for the non-critical facts should decline, which is exactly what Hirst et al. found (see Figure 10.2).

FIGURE 10.2 Proportion

of accurately recalled critical and non-critical facts about the attack of September 11th within the first week and after one year, three years, and ten years (Based on Hirst et al., 2015).

212  Hirst and Meksin

One can observe a slight increase in performance for the non-critical items in the third and tenth years, but a closer inspection reveals again a role of the media. The third-year improvement for the location of President Bush can be attributed to the release of Michael Moore’s film Fahrenheit 9/11 a few months before the thirdyear survey. In his film, Moore included a striking video of President Bush learning of the attack. It showed President Bush, at an elementary school in Florida, looking confused and continuing to read to the assembled children The Pet Goat for several minutes after his aide Andrew Card whispered to him that there was a second plane crash and “America is under attack” (Yurdakul, 2009).This film no doubt improved the memory of our participants, in large part, because it was extremely popular, with the film grossing as much as the successful action-thriller The Bourne Identity (www.boxofficemojo.com). Moreover, the film was extensively discussed in the press (see Hirst et al., 2009, for a discussion of this point.) As to the improvement in the tenth year for the airline names, this could be attributed solely to the release, five years earlier, of the feature film United 93, which by its title alone prominently highlighted the fateful events of the airline carrier and flight number. Finally, to revisit Pezdek’s claim that emotional attachment could account for the differences in the accuracy of event memories, Hirst et al. (2009, 2015) found that those who reported personal loss and/or inconvenience evidenced better event memory than those who did not. However, a mediational analysis indicates that this finding could, at least in part, be attributed to the level of media attention these participants engaged in. That is, those who experienced personal loss and inconvenience attended to the media more, and this increased attention, rather than any emotional attachment elicited by personal loss and inconvenience, accounted, in large part, for their better event memory.

Collective memories, event memories, and ageing Recently, Meksin, Cyr, Mather, and Hirst (2016) examined the performance of the subset of participants in the Hirst et al. studies who were 65 years of age or older at the time of the 9/11 attack and who completed all four surveys. For these 21 participants, the consistency of their FBMs was comparable to the younger subsample across the entire ten-year period. There was indication that performance between these two subsamples was beginning to diverge after ten years, but the difference was not significant. For event memories, however, although no difference could be found between the two subsamples in the first week after the attack, significant differences could be found after a year for the non-critical facts. By year three, the accuracy of event memories, even for the critical facts, was lower for the older subsample. Interestingly, the older subsample did show a mnemonic benefit for their memories for the location of President Bush at the time of the attack with the release of Fahrenheit 9/11 (82% reported on the third-year survey that they saw the film). However, unlike their younger counterparts, this benefit was not sustained to the tenth-year survey: The difference in memory between the younger and older participants disappeared in year three, but re-appeared in year ten. As to the airlines’ names, none of the older subsample reported that they saw United 93.

Aligning Flashbulb and collective memories  213

Meksin et al. (2016) attributed this pattern of performance to the encoding difficulties the elderly can experience. Because of their encoding problems, they could not benefit from the re-exposure to the facts that the media promulgate. As a result, unlike what was the case for their younger counterparts, their memory for the critical facts declined over time. The decline for non-critical facts was even more marked. As to the mnemonic benefit their older subsample experienced with the release of Fahrenheit 9/11, Meksin et al. argued that, whereas the social practices of the media can impact older adults’ memory, the short-lived nature of this benefit likely reflects a less elaborate encoding on their part.Their results suggest that older and younger adults may have different collective memories for public, emotionally charged events, in part because encoding differences create differences in the contribution the media can make in preserving memories.

The consequences of alignment What are the consequences of a community-based alignment between FBMs and collective memories? Here we want to focus on two related possibilities.

Flashbulb memories, collective memories, and collective identity In their studies about collective memory, most scholars inevitably consider its role in shaping collective identity (Olick et al., 2011). Just as autobiographical memories can influence individual identity, so can collective memories influence collective identity. What, then, is the consequence of aligning an FBM with a collective memory for the impact a collective memory has on collective identity? It should be kept in mind that the formation of an FBM of a public event is exceptional. After all, the collective memory of a community contains many public, contemporaneous, historically important events without individual community members holding an associated FBM. Even when an event may be consequential for a large number of community members, the reception event may still be rapidly forgotten. The appointment of Samuel Alito to the US Supreme Court has had profound consequences for most Americans, but we suspect most Americans do not have an FBM of learning of the confirmation of Judge Alito. Our question, then, is about the exceptional formation of an FBM. Does its alignment with an associated collective memory influence the incorporation of collective memories into collective identities? Does the presence of an FBM “add value” to a collective memory so that it has a greater impact on a collective identity? Polletta and Jasper (2001) defined collective identity as “an individual’s collective, moral, and emotional connections with a broader community, practice, or institution.” (p. 284).That is, it is not something that the community as a whole possesses, but it is something the individuals hold and that connects the individual with the group. Our claim is that FBMs do indeed “add value” by giving personal resonance to an associated collective memory, thereby motivating people both to make a connection between themselves and the broader community and to strengthen

214  Hirst and Meksin

existing connections. For example, the two authors consider 9/11 as part of their identity as Americans because they have a distinct memory of “being there.” They might not have directly experienced the trauma, but in a telling way, they feel that they did (Galea et al. 2002). More critically, the two authors want to incorporate their memories of 9/11 into their identity as Americans because not only they, but also the entire community, was “there,” as evidenced by the widespread reports of FBMs within the American community. We were all there, and hence, for us, it is an experience that needs to be part of our collective identity. In this way, in the case of 9/11, the shared experience of having an FBM helps transform an individual trauma into a collective trauma. Moreover, it gives the members of the community a common ground on which to relate and, possibly, to heal. As to more positive FBMs, the individual experience presumably shapes collective identity in an affirming manner. As we have noted, some flashbulb events, such as 9/11, may be important for not just a single country, but for a larger international community. But whatever the community involved, FBMs may not be just places in which ’we line up our lives with the sources of history’, they may also be special locations for collective identity construction.

The impact of media influence The second consequence builds on the first. Many media theorists discuss the role media might play in shaping collective identity (Hoskins, 2007; Zelizer, 1992). We want to tie this claim to the formation of a Flashbulb memory. Up to this point, we have made two claims. First, building on Hirst et al. (2009, 2015), we asserted that the media shapes the collective memory people form of a Flashbulb memoryeliciting event. Second, in the previous section, we argued that the presence of an FBM increases the impact of this collective memory on collective identity. It would follow, then, that the presence of an FBM increases the influence of media-shaped collective memories on collective identity. That is, the media can play an enhanced role in the formation of a collective memory because of the presence of a Flashbulb memory. Hirst et al. (2009, 2015), of course, mainly examined how the media shapes the rate at which event memories are forgotten. But the media can also implant false memories. For instance, many people believe that the buildings of the World Trade Center collapsed because of planted bombs, an account with a substantial presence on the Internet (e.g., Architects and Engineers for 9/11 Truth, http:// www1.911truth.org/en/evidence.html), rather than, as official reports assert (Asquith, 2006), from the heat from burning fuel. Our claim is that these people are more susceptible to this conspiracy theory in part because they have an FBM of 9/11. As far as we know, no one has tested this claim. Indeed, continued research is needed to explore, in a variety of domains, the role FBMs play in structuring collective identity, especially in how they enhance the influence of media on collective memory. But whether through preservation of memories, forgetting, or the

Aligning Flashbulb and collective memories  215

implantation of false memories, the formation of a Flashbulb memory provides not only a location for collective identity construction, but in doing so, grants the media unintended, or intended, influence on the form this identity will take.

Concluding remarks In this chapter, we have argued that (1) people are much more likely to form FBMs if the eliciting event is consequential for their own community, not someone else’s community; (2) inconsistency in FBMs, and at least some of their characteristics, reflect in part the consequences of the social identity people hold; (3) event memories reflect in part the social practices of a community; (4) the presence of an FBM increases the impact of a collective memory on collective identity; and (5) the extensive media coverage of the flashbulb event, in turn, shapes this collective memory. Based on these observations, we averred that the presence of autobiographical FBMs allows one to state not only that “I was there,” but also that “I, along with other members of the affected community, was there.” Moreover, when aligning this autobiographical memory with a collective memory, I and my fellow community members do so with a collective memory built on the social practices of our community, in particular, the practices of the media. Finally, because of the interactions between FBMs, collective memories, and collective identity, the presence of an FBM allows these practices to have a greater influence on collective identity than might otherwise be expected. In the best sense, then, Neisser’s (1982) telling remarks about the relation between FBMs and History opened the door to an area where an individual’s psychology and social context meet. Our considerations of Neisser’s perspicacious observation are aimed at opening that door even wider, to demonstrate that there is so much more to FBMs than meets the “I.”

Note 1 The authors gratefully acknowledge the support from NIHM grant #0066972 and grants from the John S. McDonnell Foundation.

References Asquith, C. (2006, September 5). Who really blew up the twin towers? The Guardian. Retrieved from http://www.theguardian.com. Bahrick, H. P. (1984). Semantic memory content in permastore: Fifty years of memory for Spanish learned in school. Journal of Experimental Psychology: General, 113, 1–29. Belli, R. F., Schuman, H., & Jackson, B. (1997). Autobiographical misremembering: John Dean is not alone. Applied Cognitive Psychology, 11, 187–209. Berntsen, D. (2009). Flashbulb memories and social identity. In O. Luminet & A. Curci (Eds.), Flashbulb memories: New issues and new perspectives (pp. 187–205). New York: Psychological Press. Berntsen, D., & Thomsen, D. K. (2005). Personal memories for remote historical events: Accuracy and clarity of flashbulb memories related to World War II. Journal of Experimental Psychology: General, 134, 242–257.

216  Hirst and Meksin

Bohannon, J. N. III, & Symons, V. L. (1992). Flashbulb memories: Confidence, consistency, and quantity. In E. Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of ‘‘flashbulb’’memories (Vol. 4, pp. 65–91). Cambridge: Cambridge University Press. Bouvier, G. (2005). ‘Breaking’ news: the first hours of the BBC coverage of 9/11 as a media event. Journal for Crime, Conflict, and the Media, 1, 19–43. Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–79. Christianson, S. Å. (1989). Flashbulb memories: Special, but not so special. Memory and Cognition, 17, 435–443. Christianson, S. Å. (1992). Do flashbulb memories differ from other types of emotional memories? In E. Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of ‘‘flashbulb’’memories (pp. 191–211). Cambridge: Cambridge University Press. Conway, M. A., Anderson, S. J., Larsen, S. F., Donnelly, C. M., McDaniel, M. A., McClelland, A. G., . . . & Logie, R. H. (1994). The formation of flashbulb memories. Memory & Cognition, 22, 326–343. Corning, A., & Schuman, H. (2015). Generations and collective memory. Chicago: Chicago University Press. Curci, A., & Luminet, O. (2006). Follow-up of a cross-national comparison of flashbulb and event memory for the September 11th attacks. Memory, 14, 329–344. Curci, A., Luminet, O., Finkenauer, C., & Gisle, L. (2001). Flashbulb memories in social groups: A comparative test-retest study of the memory of French President Mitterrand’s death in a French and a Belgian group. Memory, 9, 81–101. Darling-Hammond, S. (2016, January 13). Lives fit to print: Exposing media bias in coverage of terrorism. The Nation. Retrieved from https://www.thenation.com/article/ what-happened-to-the-change-we-once-believed-in/. Dayan, D., & Katz, E. (1992). Media events: The live broadcasting history. Cambridge, MA: Harvard University Press. Demiray, B., & Freund, A. M. (2015). Michael Jackson, Bin Laden and I: Functions of positive and negative, public and private flashbulb memories. Memory, 23, 487–506. Echterhoff, G., & Hirst, W. (2006). Thinking about memories for everyday and shocking events: Do people use ease-of-retrieval cues in memory judgments? Memory & Cognition, 34, 763–775. Er, N. (2003). A new flashbulb memory model applied to the Marmara earthquake. Applied Cognitive Psychology, 17, 503–517. Finkenauer, C., Luminet, O., Gisle, L., El-Ahmadi, A., van der Linden, M., & Philippot, P. (1998). Flashbulb memories and the underlying mechanisms of their formation: Toward an emotional-integrative model. Memory and Cognition, 26, 516–531. Galea, S., Ahem, J., Resnick, H., Kilpatrick, D., Bucuvalas, M., Gold, J., & Vlahov, D. (2002). Psychological sequelae of the September 11 terrorist attacks in New York City. New England Journal of Medicine, 346, 982–987. Gillis, J. R. (Ed.). (1994). Commemoration:The politics of national identity. Princeton, NJ: Princeton University Press. Halbwachs, M. (1992). On collective memory. Chicago: University of Chicago Press. Hirst, W., & Manier, D. (2002). The diverse forms of collective memory. In G. Echterhoff & M. Saar (Eds.), Kontexte und Kulturen des Erinnerns [Contexts and cultures of remembering] (pp. 37–58). Konstanz, Germany: UVK. Hirst, W., & Manier, D. (2008). Towards a psychology of collective memory. Memory, 16, 183–200. Hirst, W., & Phelps, E. A. (2016). Flashbulb memories. Current Directions in Psychological Science, 25, 36–41.

Aligning Flashbulb and collective memories  217

Hirst, W., Phelps, E. A., Buckner R. L., Budson, A. E., Cuc, A., Gabrieli, J. D. E., . . .Vaidya, C. J. (2009).The long-term memory of the terrorist attack of September 11, 2001: Flashbulb memories, event memories, and the factors that affect their retention. Journal of Experimental Psychology: General, 138, 161–176. Hirst,W., Phelps, E. A., Meksin, R.,Vaidya, C. J., Johnson, M. K., Mitchell, K. J., . . . Mather, M. (2015). A ten-year follow-up of a study of memory for the attack of September 11, 2001: Flashbulb memories and memories for flashbulb events. Journal of Experimental Psychology: General, 144, 604–623. Hirst, W., & Stone, C. B. (2016). A unified approach to collective memory: Sociology, psychology, and the extended mind. In S. Kattago (Ed.), The Ashgate research companion to memory studies (pp. 103–116). New York: Routledge. Hoskins, A. (2007). Media and memory. New York: Routledge. Iyengar, S. & Kinder, D. R. (1987). News that matters. Chicago: Chicago University Press. Kansteiner, W. (2002). Finding meaning in memory: A methodological critique of collective memory studies. History and Theory, 41, 179–197. Koppel, J., Brown, A. D., Stone, C. B., Coman, A., & Hirst, W. (2013). Remembering President Barack Obama’s inauguration and the landing of US Airways Flight 1549: A comparison of the predictors of autobiographical and event memory. Memory, 21, 798–806. Kvavilashvili, L., Mirani, J., Schlagman, S., Foley, K., & Kornbrot, D. E. (2009). Consistency of flashbulb memories of September 11 over long delays: Implications for consolidation and wrong time slice hypotheses. Journal of Memory and Language, 61, 556–572. Kvavilashvili, L., Mirani, J., Schlagman, S., & Kornbrot, D. E. (2003). Comparing flashbulb memories of September 11 and the death of Princess Diana: Effect of time delays and nationality. Applied Cognitive Psychology, 17, 1017–1031. Luminet, O., & Curci, A. (2009). The 9/11 attacks inside and outside the US: Testing four models of flashbulb memory formation across groups and the specific effects of social identity. Memory, 17, 742–759. Luminet, O., Curci, A., Marsh, E., Wessel, I., Constantin, T., Gencoz, F., & Yogo, M. (2004). The cognitive, emotional, and social impacts of the September 11 attacks: Group differences in memory for the reception context and the determinants of flashbulb memory. The Journal of General Psychology, 17, 171–181. Mahmood. D., Manier, D., & Hirst, W. (2004). Memory for how one learned of multiple deaths from AIDS: Repeated exposure and distinctiveness. Memory & Cognition, 32, 125–134. Margalit, A. (2002). The ethics of memory. Cambridge, MA: Harvard University Press. McCloskey, M., Wible, C. G., & Cohen, N. J. (1988). Is there a special flashbulb-memory mechanism? Journal of Experimental Psychology: General, 117, 171–181. Meksin, R., Cyr, T., Mather, M., & Hirst, W. (2016). Age and the retention of flashbulb memories and their associated event memories. In preparation. Nachson, I., & Zelig, A. (2003). Flashbulb and factual memories: The case of Rabin’s assassination. Applied Cognitive Psychology, 17, 519–531. Neisser, U. (1982). Snapshots or benchmarks. In U. Neisser (Ed.), Memory observed: Remembering in natural contexts (pp. 43–48). San Francisco: Freeman. Neisser, U., Winograd, E., Bergman, E. T., Schreiber, C. A., Palmer, S. E., & Weldon, M. S. (1996). Remembering the earthquake: Direct experience vs. hearing the news. Memory, 4, 337–357. Nora, P. (1989). Between memory and history: Les lieux de mémoire. Representations, 26, 7–24. Olick, J. K. (1999). Collective memory: The two cultures. Sociological theory, 17, 333–348.

218  Hirst and Meksin

Olick, J. K., Vinitzky-Seroussi, V., & Levy, D. (2011). The collective memory reader. New York: Oxford University Press. Otani, H., Kusumi, T., Kata, K., Matsuda, K., Kern, R. P., Widner, R., & Ohta, N. (2005). Remembering a nuclear accident in Japan: Did it trigger flashbulb memories? Memory, 13, 6–20. Pezdek, K. (2003). Event memory and autobiographical memory for events of September 11, 2001. Applied Cognitive Psychology, 17, 1033–1045. Polletta, F., & Jasper, J. M. (2001). Collective identity and social movements. Annual Review of Sociology, 27, 283–305. Rasmussen, A. S., & Berntsen, D. (2009). Emotional valence and the functions of autobiographical memories: Positive and negative memories serve different functions. Memory & Cognition, 37, 477–492. Roediger, H. L., Zaromb, F. M., & Butler, A. C. (2009). The role of repeated retrieval in shaping collective memory. In P. Boyer & J. V. Wertsch (Eds.), Memory in mind and culture (pp. 29–58). New York: Cambridge University Press. Smith, M. C., Bibi, U., & Sheard, D. E. (2003). Evidence for the differential impact of time and emotion on personal and event memories for September 11, 2001. Applied Cognitive Psychology, 17, 1047–1055. Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–37). Monterey, CA: Brooks/Cole. Talarico, J. M. & Rubin. D. C. (2003). Confidence, not consistency, characterizes flashbulb memories. Psychological Science, 14, 455–461. Tekcan, A., Berivan, E., Gülgöz, S., & Er, N. (2003). Autobiographical and event memory for 9/11: Changes across one year. Applied Cognitive Psychology, 17, 1057–1066. Tinti, C., Schmidt, S., Sotgiu, I., Testa, S., & Curci, A. (2009). The role of importance/consequentiality appraisal in flashbulb memory formation: The case of the death of Pope John Paul II. Applied Cognitive Psychology, 23, 236–253. Tinti, C., Schmidt, S., Testa, S., & Levine, L. J. (2014). Distinct processes shape flashbulb and event memories. Memory & Cognition, 42, 539–551. Winningham, R. G., Hyman, I. E., & Dinnel, D. L. (2000). Flashbulb memories? The effects of when the initial memory report was obtained. Memory, 8, 209–216. Wolters, G., & Goudsmit, J. (2005). Flashbulb and event memory of September 11, 2001: Consistency, confidence and age effects. Psychological Reports, 96, 605–619. Wright, D. B. (1993). Recall of the Hillsborough disaster over time: Systematic biases of ‘flashbulb’ memories. Applied Cognitive Psychology, 7, 129–138. Wright, D. B., Gaskell, G. D., & O’Muircheartaigh, C. A. (1998). Flashbulb memory assumptions: Using national surveys to explore cognitive phenomena. British Journal of Psychology, 89, 103–121. Yurdakul, A. (2009). He told Bush that “America is under attack”. NBC News. September 10, 2009. Retrieved from http://www.nbcnews.com/id/32782623/ns/us_news-9_11_eight_ years_later/t/he-told-bush-america-under-attack/#.WNv-QWP8Tdk. Zelizer, B. (1992). Covering the body:The Kennedy assassination, the media, and the shaping of collective memory. Chicago: University of Chicago Press. Zerubavel, E. (2012). Time maps: Collective memory and the social shape of the past. Chicago: University of Chicago Press. Zerubavel,Y. (1995). Recovered roots: Collective memory and the making of Israeli national tradition. Chicago: University of Chicago Press.

11 FLASHBULB MEMORIES AND COLLECTIVE MEMORIES Psychosocial processes related to rituals, emotions, and memories Dario Páez, Anderson Mathias, Stefano Cavalli, Eduardo Guichard, José Joaquín Pizarro, Lander Méndez, and Sofián El-Astal The chapter discusses similarities and differences of Flashbulb memories (FBMs) and collective memories (CMs). Evidence suggesting that they are formed during formative years or a critical period and similarity in content is reviewed. Similarities in events features that elicit FBMs and CMs, the role of social rehearsal, collective behaviors and rituals, and relevance for social identity for both are reviewed. Differences are also discussed: FBMs imply autobiographical episodic memory analyzed at an individual level, while CMs are related to semantic memory, and are focused on distributed memory, societal antecedents, process and consequences.

This chapter intends to examine parallels existing between Flashbulb memories (FBMs) and collective memories (CMs), with a special accent on the impact on memory of social processes such as social rehearsal and social rituals. Our aim is first, to analyze the similarities and differences between FBMs and CMs in content. Second, to examine similarities in processes, like the relevance of the critical period of identity formation for the creation of FBMs and CMs. Third, we also review the importance of emotionality, relevance for society, and social identity for the formation of CMs and FBMs. Finally, we examine the role of social sharing, mass media rehearsal and commemorations and rituals in both FBMs and CMs. At an individual level, CMs are a semantic memory of historical events, including a memory of the event, but also, in the case of lived experiences, an autobiographical memory related to historical events. They are not only shared individual memories, but also social representations maintained by informal communication, commemorations, memorials, textbooks, and others, such as cultural tools that exist to preserve the past (Coman, Brown, Koppel, & Hirst, 2009; see also Hirst & Meksin, Chapter 10, this volume). CMs are a widely shared knowledge of past social events considered relevant for the collective identity, which may not have been

220 Páez et al.

personally experienced, yet they are collectively constructed through communicative social functions (Schuman & Scott, 1989). These social representations, or shared knowledge about the past, are elaborated, transmitted, and conserved in a society through both interpersonal and institutional communications. Social representations of the past are helpful to people for a variety of reasons. First, they maintain a positive image of the group to whi\ch they belong. Second, they preserve a sense of continuity of the group, able to endure through time. Third, they provide guidance to group values and norms by prescribing behaviors and contributing to what characterizes, or should characterize, the group. Fourth, they are a symbolic resource that can be mobilized politically (Páez & Liu, 2012).

FBMs and CMs: A comparison CMs often include negative events (in fact, more than half of historical recalled events have negative valence, see below). However, in the long term, a positive view of history prevails. For instance, military failures are transformed into moral victories, killed in-group members become martyrs and heroes, negative events are reappraised, minimized, or simply forgotten, and the general historical narrative helps to build a positive in-group view (Baumeister & Hastings, 1997). Conversely, FBMs are vivid and confident memories of the personal circumstances surrounding people’s discovery of shocking public events, such as assassinations of famous figures (Hirst et al., 2015). FBMs are distinguished from ordinary autobiographical episodic memories by their vividness and the confidence with which they are held. At odds with previous conceptions, evidence showed that they are not reliably different from ordinary autobiographical memories in longevity, accuracy, or consistency (see Talarico & Rubin, Chapter 4, this volume) See Table 11.1 for a comparison between FBMs and CMs. In general, FBM studies focused on one single major emotional event. In line with Brown and Kulik’s (1977) inaugural research, subsequent studies dealt predominantly with negative public events.These events elicited in some 40 to 100% of investigated persons, a clear and vivid remembering of what happened (e.g. the scene of the shooting of President Kennedy; the moment and circumstances of Franco’s death). In addition, these persons usually recalled with high perceptual clarity, the personal circumstances in which they learned about it (e.g. where they stood).These memories display, for decades after the event, higher vividness and confidence – even they are not more accurate (Talarico & Rubin, 2003) – than other everyday autobiographical memories (Berntsen, 2009). What is more, FBMs are of great importance because they are cases in which personal biography and identity meet, and are embedded in, “History” and collective identity (Neisser, 1982). To a large extent, people’s personal lives unfold locally and within a small social sphere. When an event such as the assassination of John F. Kennedy or the 9/11 attacks occurs, the personal timeline of one’s life and the historical timeline of a nation intersect. Hirst et al., 2015, p. 619

Flashbulb memories and collective memories  221 TABLE 11.1  Comparison between the key characteristics of CMs and FBMs

Definition

Type of memory Level of analysis Experience Focus Type of event

FBMs

CMs

Vivid and confident memories of personal circumstances surrounding the individual learning of shocking events. Autobiographical, episodic.

Shared memories of past events relevant for collective identity which are transmitted and conserved in a society.

Individual. Reported. Reception context of a public event. A major single event. Real and very specific.

Type of source

Other people, media news.

Number of sources Target event features

Usually a few or one. Unexpectedness, personal consequentiality, emotional intensity, and valence (negative, traumatic or extremely positive). Concrete, vivid. Generally very high. Important.

Memory quality Memory confidence Relevance to social group Identity level Media

Duration

Group, generational. Personal narratives (social sharing), mass media.

Generally long.

Semantic memory of events, but also autobiographical for lived events. Collective. Generally reported. Public event. Single or various events, processes or time periods. Real or symbolic. Various, including commemorations, memorials, monuments, etc. Various. Social consequentiality (major changes or threats), collective emotions and meanings (positive and negative). More abstract. High/consensual. Very important. Social, national. Mass media, cultural products (e.g. books, movies, art), rituals, institutions, group and personal narratives. Very long; years or generations.

Note: Based on Páez, Bellelli, and Rimé (2009, p 231).

While autobiographical memory serves a primary function of defining personal identity, collective memory sustains a nation, group, or community’s collective identity and makes possible the continuity and cohesion of a society (Wang & Aydin, 2009). FBMs are instances of autobiographical memory, but also intersect with CMs. The proper feature of FBMs is the fact that they mix personal and very idiosyncratic elements of private experience with the socially shared information pertaining to a collective event and, therefore, CMs usually imply FBMs for lived events, though FBMs are not equal to CMs. This is mainly because the latter often

222 Páez et al.

relates to extended events and even time periods, while FBMs refer to temporal bound events during the individual’s life. Also, CMs usually include cultural memories (e.g. myths and old events and periods) and cross-generational memories (e.g. parents’ and grandparents’ recollections of WWII). In the same vein, communicative collective memory, or oral memory, is supposed to integrate three generations, however evidence is limited and Stone, van der Haegen, Luminet, and Hirst (2014), examining the transmission of WWII related memories across three generations in Belgium, found limited transmission from grandparents and parents to their children and grandchildren. Evidence suggests that CMs are frequently intertwined with FBMs. Some studies on collective memory that ask people for important historical events in their life and use one canonical question of FBMs (i.e. where they were [location] when hearing about the event), found that a huge percentage answer positively. As an example, a study in Chile found that 99% reported to have remembered where they were when they learned about historical events like Pinochet’s putsch on September 11th, 1973 (Guichard, 2015).

FBMs and CMs: Similarities in content A review of the content of FBM studies (Páez, Rimé, & Bellelli, 2009) including recent ones related to socio-political collective and positive events (see Table 11.2), indicates that episodes of collective violence, such as the Gulf War and Operation Desert Storm or the 9/11 attacks, ranked first in frequency, being these kind of studies one the most repeated – for instance, Hirst et al. (2015) posits that for the 9/11 attacks alone, there are around 20 studies. Following the ranking, a nearly equal number of studies examined events of political violence, such as the assassination of political leaders (e.g. U.S. Presidents Abraham Lincoln and John F. Kennedy; Swedish Prime Minister Olof Palme), or assassination attempts (e.g. U.S. President Ronald Reagan). Collective catastrophes ranked third, with studies dealing with episodes such as the Challenger space shuttle disaster (1986), the San Francisco earthquake (1989), and the Chernobyl disaster (1986). Death of famous people ranked fourth, with studies focused on the natural deaths of important political figures, both unexpected (e.g. Belgium King Baudouin; the founder and the first president of Turkey, Mustafa Kemal Ataturk) or expected (e.g. Spanish Dictator, Francisco Franco and French President François Mitterrand). A smaller number of studies concerned political crises, such as the resignations of the Australian and Japanese Prime Ministers (Stone, Luminet, & Takahashi, 2015). Ranked last in frequency are positive political events, like the German withdrawal from Denmark at the end of WWII (Berntsen & Thomsen, 2005), the fall of the Berlin Wall (Bohn & Berntsen, 2007), and positive sport events like the Italian national football team’s victory in 2006 World Championship (Tinti, Schmidt, Testa, & Levine, 2014). These aforementioned events are public and they are certainly similar to the content of CMs.We contrasted these distributions of public events related to FBMs with the profile found in studies on CMs, using the CEVI (Changes and Events across the Life Course) international research program (see Guichard, 2015; Lalive

Flashbulb memories and collective memories  223 TABLE 11.2 Frequency distribution of types of events considered in FBM and CM studies

Types of events Collective violence (e.g. 9/11) and wars (e.g. WWII, Malvinas) Political violence – attacks, political leaders’ murder (e.g. JFK) Political turning points – negative events (e.g. State coups) Political turning points – positive events (e.g. democracy return) Positive sport events Social periods (e.g. military dictatorship) Collective catastrophes Death of famous people Political and social crisis (e.g. resignation of Thatcher) Economic crisis Creation institutions (e.g. EU) Elections, governments Social mobilizations (e.g. May 68) Technological advances Discovering and exploration Other

FBM

CEVI

n studies

% studies

n events

% events

20

23.0

24

20.0

17

20.0

4

3.3

2

1.7

3

4.0

9

7.5

3 16 14 7

4.0 18.5 17.0 7.0

4 7 10 3 3

3.3 5.8 8.3 2.5 2.5

6.5 100.0

9 7 19 8 5 4 2 120

7.5 5.8 15.8 6.7 4.2 3.3 1.8 100.0

5 85

Note: First column for number of studies, second for percentage of 85 FBMs studies (actualization Páez, Bellelli & Rimé, 2009). Third column for number of events, and fourth, for percentage of events in 120 CEVI studies.

d’Epinay, Cavalli, & Aeby, 2008). Participants in CEVI were asked the following question: “Consider the main changes and events which occurred in your country and in the world during your life. What are the ones which most struck you? (Please, mention up to four events)”. Data were collected during 2004 and 2012 in large convenience samples from 18 to 85 years old in ten countries (Argentina, Belgium, Brazil, Canada, Chile, France, Italy, Mexico, Switzerland and Uruguay). Based on these nations, and the ten most mentioned events, similarities between FBMs and CMs events are evident: the most mentioned or recalled events are wars, political and collective violence-related events, and catastrophes.The 9/11 attacks appear in all countries. In collective violence, mentioning of recent wars (e.g. Gulf War or Malvinas) or WWII is important – which are marginal in FBM studies. Elections and new governments ranked second among the most cited events in CEVI: first Obama’s election is internationally mentioned, but likewise government-related events associated with important changes, like the first Peron’s government in Argentina, Fox’s election in Mexico, or the recent governments of Lula da Silva and Dilma Roussef in Brazil – FBM studies on political changes are marginal but exist (e.g. Koppel et al., 2013; FBM study on Obama’s election). Catastrophe-related

224 Páez et al.

events include local earthquakes and disasters, although there are also international ones like the Indian Ocean earthquake and tsunami in 2004. Positive political turning points have an important presence: the return to democracy in Latin America and the fall of the Berlin Wall in Western Europe. Further, socio-economical periods and economical crises are remembered in CMs (e.g. depression, the Crisis of 2008, inflation) and political changes, including negative turning points (e.g. instauration of military dictatorships in Latin America, the Years of Lead in Italy, the Quiet Revolution in Québec). Social mobilizations (e.g. May 68, Zapatista mobilization, movements for free elections in Brazil), creation of institutions (e.g. introduction of the Euro currency), technological advances and the exploration of the Moon are also considerable. Sports events include an important presence of football and the organization of the Olympic Games. The fact that CMs could also refer to extended periods and not only to bounded and lived events explains more of these differences. In other cases, like government changes (e.g. studies in government changes considered positive by a part of the population like Mitterrand’s, Allende’s or Lula’s first governments), FBMs are rare, probably because they are related to extended institutional changes and not limited to a specific event. One might argue that relevant specific events (e.g. when the candidate wins) are less salient, or even more complex and symbolic (e.g. Obama’s first speech; see Koppel, Brown, Stone, Coman, & Hirst, 2013), and thus, not enough to form an FBM. Studies on collective memory have found that across diverse cultures there are some trends in common among CMs and FBMs. Evidence suggests that dominance of war and politics, relevance of traumatic events, euro-centrism or Westerncentrism and socio-centrism are common to FBMs and CMs (Wang & Aydin, 2009). Wars, political and military leaders and events are cross-culturally referred to as historically important. Revolutions and wars are mentioned in world history (Liu et al., 2005; Pennebaker, Páez, & Deschamps, 2006) as the most important events, while science and technology are secondary in significance. In 24 nations from America, Europe, and Asia, collective violence accounted for 48% of events nominated as important, whereas 45% of the named leaders were known for their roles in violent acts (Liu et al., 2005). Even though wars accounted for only 2% of the 20th century premature death toll (Layard, 2005), people tend to overvalue the role of political violence in world history because of the extreme and negative impact events like wars have. For example, when asked for the most important historical events of the last century in seven nations (data was collected at the end of the 20th century), wars represented 40% of the 70 events, space exploration 10%, computer technology, industrial and medical advances 9%, USSR collapse 7%, social movements like Women’s movement and May 68 6%, and the atomic bomb, sociopolitical periods like the Cold War and Fascism, economic depression, HIV/AIDS epidemic and JFK’s assassination, 3% each. 38% of the events have a positive valence and 62% a negative one. When analyzing different events individually reported in FBM and CM studies, it is clear that both can refer to negative (e.g. assassination, instances of collective violence) and positive (e.g. space exploration, the fall of the

Flashbulb memories and collective memories  225

Berlin Wall, and the USRR collapse) extreme events. Conversely, CMs can include not only these events, but also long-term periods and social changes. Continuing with Western-centrism, some of them recalled as important for world history are predominantly related to Europe and North America (e.g. World Wars, Pennebaker et al., 2006), even by subjects of African and Asian countries (Liu et al., 2005; Pennebaker et al., 2006). Historiography (Goody, 2007) and news are focused on the USA and Western events, and similar findings are described in FBMs (Wang & Aydin, 2009). Together, such findings reflect the representational power of the West. Finally, research has also revealed a partial tendency towards socio-centrism. Most nations consider their national historical events as more important than events unrelated to their own history. In Western countries, for instance, participants in the U.S. list the American Civil War as one of the most important events of the last millennium. Similarly, participants from non-Western countries exhibit ethnocentrism in mentioning as important events in world history, the ones that are related to the creation of their own state (e.g. decolonization), and devaluing events linked to the history of neighboring states that did not affect their own historical trajectory (Liu et al., 2005). Still, even if people globally exhibit some ethnocentrism in their view of universal history, the socio-centric bias does not hold for all nations. American “new nations” exhibit stronger socio-centrism than European countries (Concha & Henriquez, 2011; Guichard & Henriquez, 2011; Oddone & Lynch, 2008; Paredes & Oberti, 2015). Similar findings are described in FBMs (see Wang & Aydin, Chapter 12, this volume).

FBMs and CMs: Similarities in the relevance of a critical period Events that generate FBMs are similar to the typical content of CMs and also factors which participate in the formation of FBMs are similar to those that create CMs (Páez & Liu, 2012). If attention is then centred to the moment when the event was lived, one can realise that people normally remember a high number of public events that occurred when they were between 10–30 years old, the socalled “reminiscence bump”. This is called a formative or critical period, because people are forming their adult identity, entering into large social networks, and experiencing for the first time, or in a fresh manner, some events (e.g. sexual initiation, political experiences, leaving home and so on). This implies that CMs are, as Mannheim (quoted in Schuman & Scott, 1989) suggests, cohort-dependent: people remember better historical events experienced during adolescence or early adulthood, a formative period in one’s social identity (see Wang & Aydin, Chapter 12, this volume). Confirming this phenomenon, in 1989, older Americans mentioned the Great Depression and WWII more as an important historical event, whereas younger participants mentioned more frequently JFK’s assassination and the Vietnam War – events that had occurred during participants’ early adulthood (Schuman & Rodgers, 2004).

226 Páez et al.

Based on studies on remembering important national and world historical events during participants lives (CEVI research program), and during the last 50–60 years (Schuman & Scott, 1989; Schuman, Akiyama, & Knäuper, 1998), a meta-analytical synthesis was performed. For instance, mention of the Vietnam War as a historical event by participants that were 10–30 years old when the Vietnam War occurred (33.1%) were compared with mention of the Vietnam War by participants who were less than ten years old and over 30 years old during the war.These two groups were collapsed in a unique “others” categories and 14.3% of these participants mentioned the Vietnam war as an historical event (see Table 11.3). PHI coefficient, a dichotomous equivalent of correlation, was used to check if there was a significant association between the mentioning of the event and the cohort. Weighted effect size for the critical period was r = .18, p < .001 (for 28 studies and N = 15.486). Similarly to CMs, people have been reported to form FBMs at younger ages, during the formative years of identity (Denver, Lane, & Cherry, 2010). When asked to remember FBMs from their lives, older adults usually remember events from the formative years of between 10–30 (Rubin, Rahhal, & Poon, 1998). On the other hand, young people report more FBMs of recent events than older adults. For example, FBMs scores for people who lived the death of the Turkish Prime Minister in 1993 at a young age was higher (M = 7.49) than the elderly group (M = 6.40) who lived the same event when they were 60 to 80 years old.Young participants reported better memory for the event, and shared or talked to people more than older ones (Tekcan & Peynircioğlu, 2002). Similar results were found for the 9/11 attacks (Denver, Lane, & Cherry, 2010). However, other studies found that FBMs are reported at a similar level during different periods of life (see Wang & Aydin, Chapter 12, this volume). FBMs and CMs not only share the relevance of the critical period, but they share other processes as will be developed in the next two sections.

FBMs and CMs: Similarities in predictors: Novelty, surprise, emotional intensity, and valence Models of FBMs included different predictors for the formation of memories: (1) novelty and the reaction of surprise when learning about the event; (2) the intensity and valence of emotional reactions; (3) the consequentiality of the original event and appraisal of importance; (4) rehearsal, internal or rumination; external interpersonal and mediated by mass media; and (5) pre-existing knowledge, attitudes, and event memory (Luminet, 2009). In this section, we examine evidence for the similarities between FBMs and CMs for the first and second types of predictors, particularly emotional valence, and in the next section we look at evidence for the other predictors. Novelty and surprise were supposed to be important predictors for the formation of both CMs and FBMs (but see Talarico & Rubin, Chapter 4, this volume, whose review concluded that novelty and surprise had a limited role in FBMs). CMs and FBMs evolve from public events that collectively affect a large number of people – even if FBMs are reported for personal events, like admission into

1985

1985 1993 1997

2004

2004

2004

2004

2004 2009 2004 2009 2004 2009 2004 2009 2008

United States

United States

Argentina

Argentina

Argentina

Argentina

Switzerland

Chile

Sweden

Switzerland

Switzerland

Switzerland

2009

1985

United States

Spain

Year

Country

623

946

1301

1301

1301

1301

579

579

579

579

750

348

1243

1280

N

old) by country and nature of event

Democracy return

Collapse of Communism

May 1968

Man on the Moon

Fall of Berlin Wall

9/11

2001 crisis

9/11

Alfonsín’s government

Democracy return

Democracy return

Civil rights

World War II

Vietnam War

Event

45.8

11.6

14.9

9.1

19.7 17.2

13.0

11.1

20.2

23.5

13.9

12.5

25.1

24.2

9.4

26.2

14.5

% others

25.2

37.8

58.3

56.9

44.0

27.7

44.8

33.3

17.7

41.8

33.1

% Crit. age

(Continued)

72.654** 0.22 22.978** 0.14 4.998* 0.12 3.725 (n.s) 0.07 16.151* 0.19 14.204* 0.18 52.891** 0.31 68.212** 0.34 73.451** 0.34 52.821** 0.29 8.913* 0.15 8.688* 0.15 2.387 (n.s) 0.05 75.025** 0.35

Chi2 Phi

TABLE 11.3 Percentage of remembered events as a function of being within or outside the critical period (experiencing the event between 10 and 30 years

2009

2009

2009

2009

2009

2009 2010 2009 2010 2010

2011

2011

2011

2011

2012

2012

Chile

Chile

Italy

Italy

Italy

France

Brazil

Brazil

Brazil

Brazil

Uruguay

Uruguay

1690

1690

637

637

637

637

580

621

621

670

670

670

623

623

N

9/11

2002 Economic crisis

Collor’s impeachment

9/11

Dilma’s government

Lula’s governments

Sichuan earthquake

Fall of Berlin Wall

9/11

Years of Lead

Fall of Berlin Wall

9/11

Bachelet’s government

9/11

Event

Note: Chi square and dichotomic orrrelation PHI (n.s.) non significative *p 60

Age at encoding

FIGURE 12.2 Lifespan

retrieval curves of a) public event memories and b) FBM details in five cultures.

256  Wang and Aydin

from 40 to 49 years for all groups when recalling public events, possibly reflecting a recency effect. Surprisingly, only the Chinese group had a peak recall that occurred from 30 to 39 years.The Chinese group also showed a distinct pattern in their recall of FBM details (Figure 12.2b), whereby they recalled significantly more details for events from the periods of 40 to 49 years and 50 to 59 years than those from other periods. In contrast, other cultural groups recalled similar numbers of FBM details across the lifespan. These findings suggest that the lifespan retrievals of FBM and public event memory may follow different courses, possibly underlined by different mechanisms. Also, it is not entirely clear whether and how the distinct pattern of lifespan retrievals in the Chinese participants may be associated with the timing of their generational identity formation. Perhaps the late bump in the Chinese entails the establishment of a “mature” self with sincere commitment to social needs and moral rectitude, which, according to Confucian teachings, does not appear until age 30, when one can hold on firmly to what he or she has learned (san-shi-er-li, 三十而立). These findings further suggest a complex relation among culture, self, and memory recollection. Further research on the lifespan retrievals of FBM and public event memory within and across cultures is crucial to corroborate the findings.

FBM as a cultural practice: Conclusions In this chapter, we outlined a culturally embedded model of FBM to delineate various cultural factors in the macro and micro contexts that may substantially shape the process and consequence of FBM. Specifically, the ways of information transmission and news coverage in a culture, together with the values and beliefs embedded in the larger cultural ambience, may largely determine how individuals receive public event news, as well as what types of events individuals value, attend to, and subsequently remember. Cultural beliefs and practices may further act on basic-level mechanisms of remembering, influencing the degree of importance or significance individuals perceive of an event, the type and intensity of emotions individuals experience and report upon learning about the news, and individuals’ attitude toward, and frequency of, memory sharing following the event. These varied processes, in turn, moderate the manifestation of commonly assumed universal mechanisms in affecting FBM formation and public event memory. Furthermore, cultural belief systems, particularly cultural conceptions of selfhood, shape the selfdefining functions that FBM, as well as public event memory, assume, which further affects the content, structure, and lifespan distribution of memories across cultures. We must acknowledge, however, that in spite of the extant theories and empirical findings, research on FBM with cultural sensitivity and methodological sophistication is still very much in need. Wherever appropriate, we have suggested directions for future research throughout the chapter. In addition to what we have discussed, several issues deserve further attention. One is related to the accuracy of FBM and its cultural determinants. It will be particularly rewarding to examine whether memory distortion about public events and their reception contexts may

Culture in Flashbulb memory  257

take different forms across cultures, such that individuals in a given culture may tend to make systematic omission or commission errors for certain types of information. Moreover, in addition to examining FBM for news events pre-selected by researchers, studies need to sample a wide variety of public events that are deemed important, consciously or unconsciously, by individuals from different cultures. And in addition to examining the accuracy, quality, and long-term retention of FBM, a greater emphasis should be placed on the content of memory narratives individuals provide, both about the news event itself and about the contextual aspects of receiving the news (Rice, Hamamouche, & Bohannon, Chapter 5, this volume). As research on autobiographical memory has demonstrated (Wang, 2013; Wang & Ross, 2007), memory content opens up a window through which we can observe cultural influences on the ways that individuals attend to, represent, organize, retrieve, and share event information. Furthermore, simultaneously studying FBM and public event memory across cultures will help us better understand the shared and distinct processes underlying the two types of memories. Patently, FBM not only takes place in a larger cultural context where the public event occurs and the news is then transmitted to and received by individuals, it also takes shape in proximal social settings during face-to-face interactions among individuals. It is thus essential to examine the dynamic interactions among individuals upon receiving the news and afterwards, and to examine how such interactions, while mediating cultural influences, shape the nature of FBM. Studies on collective memory, memory for conversations, and the effect of social sharing on individual memory have yielded important findings informative to FBM. For instance, individual recollections of shared experiences were strongly influenced by memory conversations with the presence of a narrator who initiated or dominated the conversations (Cuc, Ozuru, Manier, & Hirst, 2006; Gabbert, Memon, & Wright, 2006), and by the specific information the narrator selectively said or left unsaid (Cuc, Koppel, & Hirst, 2007). It follows that in cultures that promote social hierarchy among individuals, individual recollections of FBM may be more likely to conform to memories of those who hold greater power and resources in the society at large and within the context of memory sharing, compared with cultures that value equality and individuals’ ownership to their experiences (Wang, 2013). This remains an interesting empirical question in future research. Furthermore, studies that examine within-cultural variations, such as the influence of ethnicity, generational status of immigrants, and multiculturalism, will shed critical light on the cultural embeddedness of FBM. Ultimately, remembering of public news events and their reception contexts is an important cultural practice. Not only do the process and consequence of the remembering vary as a function of cultural influences, such memories, once formed, serve to constitute a shared cultural identity among individual members of a society or a social group, as well as a shared generational identity among individuals of a historical time that may transcend culture. Such memories further interweave our personal history with the history of our time and place, thus sealing our individual identities with a clear mark of culture.

258  Wang and Aydin

Note 1 Statistics acquired on September 12, 2016, at http://www.statista.com/statistics/282087/ number-of-monthly-active-twitter-users/ and http://expandedramblings.com/index.php/ weibo-user-statistics/.

References Alba, J. W., & Hasher, L. (1983). Is memory schematic? Psychological Review, 93, 203–231. Alwin, D. F., & Krosnick, J. A. (1991). Aging, cohorts, and the stability of sociopolitical orientations over the lifespan. American Journal of Sociology, 97, 169–195. Barrett, J. L. (2000). Exploring the natural foundations of religion. Trends in Cognitive Science, 4, 29–34. Basabe, N., Páez, D., Valencia, J., Gonzalez, J. L., Rimé, B., & Diener, E. (2002). Cultural dimensions, socioeconomic development, climate, and emotional hedonic level. Cognition & Emotion, 16, 103–125. Benson, K. A., Jarvi, S. D., Arai,Y., Thielbar, P. R. S., Frye, K. J., & McDonald, B. L. G. (1992). Socio-historical context and autobiographical memories: Variations in the reminiscence phenomenon. In M. A. Conway, D. C. Rubin, H. Spinnler, & W. A.Wagenaar (Eds.), Theoretical perspectives on autobiographical memory (pp. 313–322). Boston, MA: Kluwer Academic. Bohannon, J. N. III (1988). Flashbulb memories for the space shuttle disaster: A tale of two theories. Cognition, 29, 179–196. Bohannon, J. N. III, & Symons, L. V. (1992). Flashbulb Memories: Confidence, consistency, and quantity. In E. Wingograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb” memories (pp. 65–91). Cambridge: Cambridge University Press. Bond, M. H. (1991). Beyond the Chinese face. Hong Kong: Oxford University Press. Bourdon, J. (2003). Some sense of time: Remembering television. History & Memory, 15, 5–35. Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. Carr, N. (2010). The shallows: How the internet is changing the way we think, read and remember. London: Atlantic Books Ltd. Chao, R. K. (1995). Chinese and European American cultural models of the self reflected in mothers’ childrearing beliefs. Ethos, 23, 328–354. Christianson, S. A. (1989). Flashbulb memories: Special, but not so special. Memory and Cognition, 17, 435–443. Christianson, S. A., & Safer, M. A. (1996). Emotional events and emotions in autobiographical memories. In D. C. Rubin (Ed.), Remembering our past: Studies in autobiographical memory (pp. 218–243). New York: Cambridge University Press. Cohen, A. B. (2009). Many forms of culture. American Psychologist, 64, 194–204. doi:10.1037/ a0015308 Cole, M. (1996). Cultural psychology. Cambridge, MA: Harvard University Press. Conway, M. A. (1995). Flashbulb memories. Hove, UK: Lawrence Erlbaum. Conway, M. A., Anderson, S. J., Larsen, S. F., Donnelly, C. M., McDaniel, M. A., McClelland, A. G. R., . . . Logie, R. H. (1994).The formation of flashbulb memories. Memory & Cognition, 22, 326–343. Conway, M. A., & Pleydell-Pearce, C.W. (2000).The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261–288. Conway, M. A., & Rubin, D. C. (1993). The structure of autobiographical memory. In A. F. Collins, S. E. Gathercole, M. A. Conway, & P. E. Collins (Eds.), Theories of memory (pp. 103–137). Hove, UK: Lawrence Erlbaum.

Culture in Flashbulb memory  259

Conway, M. A., Wang, Q., Hanyu, K., & Haque, S. (2005). A cross-cultural investigation of autobiographical memory: On the universality and cultural variation of the “Reminiscence Bump”. Journal of Cross-Cultural Psychology, 36, 739–749. Conway, M.,Wang, Q., Hou,Y., Kulkofsky, S., Mueller-Johnson, K., Aydin, C., & Williams, H. (2006, July). Public event memories: A multi-national investigation. In Q. Wang & M. A. Conway (Co-Chairs), Memory & Culture. Symposium conducted at the 4th International Conference on Memory, Sydney, Australia. Cuc, A., Koppel, J., & Hirst, W. (2007). Silence is not golden: A case for socially-shared retrieval-induced forgetting. Psychological Science, 18, 727–733. Cuc, A., Ozuru,Y., Manier, D., & Hirst, W. (2006). On the formation of collective memories: The role of a dominant narrator. Memory and Cognition, 34, 752–762. Curci, A., & Luminet, O. (2006). Follow-up of a cross-national comparison on flashbulb and event memory for the September 11th attacks. Memory, 14, 329–344. Curci, A., Luminet, O., Finkenauer, C., & Gisle, L. (2001). Flashbulb memories in social groups: A comparative test-retest study of the memory of French President Mitterrand’s death in a French and a Belgian group. Memory, 9, 81–101. Erikson, E. H. (1963). Childhood and society. New York: Norton. Fernandez, I., Carrera, P., Sanchez, F., Páez, D., & Candia (2000). Differences between cultures in emotional verbal and nonverbal reactions. Psicothema, 12, 83–92. Fivush, R. (1994). Constructing narrative, emotions, and self in parent-child conversations about the past. In U. Neisser & R. Fivush (Eds.), The remembering self: Construction and accuracy in the self-narrative (pp. 136–157). New York: Cambridge University Press. Gabbert, F., Memon, A., & Wright, D. B. (2006). Memory conformity: Disentangling the steps toward influence during a discussion. Psychonomic Bulletin and Review, 13, 480–485. Gaskell, G., & Wright, D. (1997). Group differences for memory of a political event. In J.W. Pennebaker, D. Páez, & B. Rimé (Eds.), Collective memory of political events: Social psychological perspectives (pp. 175–189). Hillsdale, NJ: Lawrence Erlbaum. Halbwachs, M. (1992). On collective memory. Chicago, IL: University of Chicago Press. Heine, S. J., Lehman, D. R., Markus, H. R., & Kitayama, S. (1999). Is there a universal need for positive self-regard? Psychological Review, 106, 766–794. Hirst, W., & Phelps, E. A. (2016). Flashbulb memories. Current Directions in Psychological Science, 25, 36–41. doi:10.1177/0963721415622487 Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions and organizations across nations. Thousand Oaks, CA: Sage Publications. Holmes, A., & Conway, M. A. (1999). Generation identity and the reminiscence bump: Memory for public and private events. Journal of Adult Development, 6, 21–34. Jiang, T. L., Hou, Y., & Wang, Q. (2016). Does micro-blogging make us “shallow”? Sharing information online interferes with information comprehension. Computers in Human Behavior, 59, 210–214. Julian, M., Bohannon, J. N III., & Aue, W. (2009). Measures of flashbulb memory: Are elaborate memories consistently accurate? In O. Luminet & A. Curci (Eds.), Flashbulb memories: New issues and new perspectives (pp. 99–122). Hove and New York: Psychology Press. Koppel, J., Brown, A., Stone, C. B., Coman, A., & Hirst, W. (2013). Remembering President Barack Obama’s inauguration and the landing of US Airways Flight 1549: A comparison of the predictors of autobiographical and event memory. Memory, 21, 798–806. doi: 10.1080/09658211.2012.756040 Kulkofsky, S., Wang, Q., Conway, M.A., Hou,Y., Aydin, C., Mueller-Johnson, K., & Williams, H. (2011). Cultural variation in the correlates of flashbulb memories: An investigation in five countries. Memory, 19, 233–240.

260  Wang and Aydin

Kulkofsky, S., Wang, Q., & Koh, J. B. K. (2009). Functions of memory sharing and motherchild reminiscing behaviors: Individual and cultural variations. Journal of Cognition and Development, 10, 92–114. Kvavilashvili, L., Mirani J., Schlagman, S., & Kornbrot, D. E. (2003). Comparing flashbulb memories of September 11 and the death of Princess Diana: Effects of time delays and nationality. Applied Cognitive Psychology, 17, 1017–1031. Luminet, O., Bouts, P., Delie, F., Manstead, A. S. R., & Rimé, B. (2000). Social sharing of emotion following exposure to a negatively valenced situation. Cognition and Emotion, 14, 661–688. Luminet, O., Curci, A., Marsh, E. J., Wessel, I., Constatin, T., Gencoz, F., & Yogo, M. (2004). The cognitive, emotional, and social impacts of the September 11 attacks: Group differences in memory for the reception context and determinants of flashbulb memory. The Journal of General Psychology, 131, 197–224. Markus, H., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224–253. Marsh, E. J. (2007). Retelling is not the same as recalling: Implications for memory. Current Directions in Psychological Science, 16, 16–20. Martini, M. (1996). “What’s New?” at the dinner table: Family dynamics during mealtimes in two cultural groups in Hawaii. Early Development and Parenting, 5, 23–34. McGaugh, J. L. (2003). Memory and emotion:The making of lasting memories. New York: Columbia University Press. Mullen, M. K., & Yi, S. (1995). The cultural context of talk about the past: Implications for the development of autobiographical memory. Cognitive Development, 10(3), 407–419. Neisser, U. (1982). Snapshots or benchmarks? In U. Neisser (Ed.) Memory observed: Remembering in natural contexts (pp. 43–48). San Francisco: W. H. Freeman & Company. Nelson, K. (1993). Explaining the emergence of autobiographical memory in early childhood. In A. F. Collins, S. E. Gathercole, M. A. Conway, & P. E. Morris (Eds.), Theories of memory (pp. 355–385). Hillsdale, NJ: Lawrence Erlbaum. Otani, H. Kusumi, T., Kato, K., Matsuda, K., Kern, R. P., Widner, R., Ohta, N. (2005). Remembering a nuclear accident in Japan: Did it trigger flashbulb memories? Memory, 13, 6–20. Pasupathi, M. (2001). The social construction of the personal past and its implications for adult development. Psychological Bulletin, 127, 651–672. Pennebaker, J. W., & Banasik, B. (1997). On the creation and maintenance of collective memories: History as social psychology. In J. W. Pennebaker, D. Páez, & B. Rimé (Eds.), Collective memory of political events: Social psychological perspectives (pp. 3–19). Mahwah, NJ: Lawrence Erlbaum. Pezdek, K. (2003). Event memory and autobiographical memory for the events of September 11, 2001. Applied Cognitive Psychology, 17, 1033–1045. Pillemer, D. B. (1984). Flashbulb memories of the assassination attempt on President Reagean. Cognition, 16, 63–80. Pillemer, D. B. (1998). Momentous events, vivid memories. Cambridge, MA: Harvard University Press. Reisberg, D., & Hertel, P. (2004). (Eds.). Memory and emotion. New York: Oxford University Press. Rimé, B., Corsini, S., & Herbette, G. (2002). Emotion, verbal expression, and the social sharing of emotion. In S. R. Fussell (Ed.), The verbal communication of emotions: Interdisciplinary perspectives (pp. 185–208). Mahwah, NJ: Lawrence Erlbaum. Rimé, B., Finkenauer, C., Mesquita, B., Pennebaker, J. W., Singh-Manoux, A., & Yogo, M. (2001, July). The social sharing of emotion: An overview of cross-cultural data. In

Culture in Flashbulb memory  261

Q. Wang & M. A. Conway (Co-Chairs), Memory & Culture. Symposium conducted at the 3rd International Conference on Memory,Valencia, Spain. Rubin, D. C., Rahhal, T. A., & Poon, L. W. (1998). Things learned in early adulthood are remembered best. Memory and Cognition, 26, 3–19. Rubin, D. C., & Kozin, M. (1984).Vivid memories. Cognition, 16, 81–95. Rubin, D. C., Wetzler, S. E., & Nebes, R. D. (1986). Autobiographical memory across the adult lifespan. In D. C. Rubin (Ed.), Autobiographical memory (pp. 202–221). Cambridge: Cambridge University Press. Sahin, B., & Mebert, J. (2013). The role of culture and self-construal in autobiographical memories of US and Turkish college students. Memory, 21, 1004–1017. Schaefer, E. G., Halldorson, M. K., & Dizon-Reynante, C. (2011). TV or not TV? Does the immediacy of viewing images of a momentous news event affect the quality and stability of flashbulb memories? Memory, 19, 251–266. doi:10.1080/09658211.2011.558512 Schmolck, H., Buffalo, E. A., & Squire, L. R. (2000). Memory distortions develop over time: Recollections of the OJ Simpson trial verdict after 15 and 32 months. Psychological Science, 12, 39–45. Schuman, H., Akiyama, H., & Knauper, B. (1998). Collective memories of German and Japanese about the past half century. Memory, 6, 427–454. Schuman, H., Belli, R. F., & Bischoping, K. (1997). The generational basis of historical knowledge. In J. W. Pennebaker, D. Páez,  & B. Rimé (Eds.), Collective memory of political events: Social psychological perspectives (pp. 44–77). Hillsdale, NJ: Lawrence Erlbaum. Schuman, H., & Scott, J. (1989). Generations and collective memories. American Sociological Review, 54, 359–381. Segev E., & Blondheim, M. (2013). America’s global standing according to popular news sites from around the world. Political Communication, 30, 139–161. Shapiro, L. R. (2006). Remembering September 11th: The role of retention interval and rehearsal on flashbulb and event memory. Memory, 14, 129–147. Shoemaker, P. J., & Cohen, A. A. (2006). News around the world. New York: Routledge Taylor & Francis Group. Stone, C. B., Luminet, O., & Takahashi, M. (2015). Remembering public, political events: A cross-cultural and -sectional examination of Australian and Japanese public memories. Applied Cognitive Psychology, 29, 280–290. doi:10.1002/acp.3106 Tajfel, H. (1982). Social identity and intergroup relations. Cambridge: Cambridge University Press. Tinti, C., Schmidt, S., Testa, S., & Levine, L. J. (2014). Distinct processes shape flashbulb and event memories. Memory & Cognition, 42, 539–551. doi:10.3758/s13421-013-0383-9 Tsang, P. S., & Velazquez, V. L. (1996). Diagnosticity and multidimensional subjective workload ratings. Ergonomics, 39, 358–381. doi: 10.1080/00140139608964470 Tweed, R. G.,White, K., & Lehman, D. R. (2004). Culture, stress, and coping: Internally- and externally-targeted control strategies of European Canadians, East Asian Canadians, and Japanese. Journal of Cross-Cultural Psychology, 35, 652–668. Wang, Q. (2001a). “Did you have fun?”: American and Chinese mother-child conversations about shared emotional experiences. Cognitive Development, 16, 693–715. Wang, Q. (2001b). Cultural effects on adults’ earliest childhood recollection and selfdescription: Implications for the relation between memory and the self. Journal of Personality and Social Psychology, 81, 220–233. Wang, Q. (2003). Emotion situation knowledge in American and Chinese preschool children and adults. Cognition & Emotion, 17, 725–746. Wang, Q. (2004a). The cultural context of parent-child reminiscing: A functional analysis. In M. W. Pratt & B. Fiese (Eds.), Family stories and the life course: Across time and generations (pp. 279–301). Mahwah, NJ: Lawrence Erlbaum.

262  Wang and Aydin

Wang, Q. (2004b). The emergence of cultural self-constructs: Autobiographical memory and self-description in European American and Chinese children. Developmental Psychology, 40, 3–15. doi:10.1037/0012–1649.40.1.3 Wang, Q. (2006). Relations of maternal style and child self-concept to autobiographical memories in Chinese, Chinese immigrant, and European American 3-year-olds. Child Development, 77, 1794–1809. doi:10.1111/j.1467–8624.2006.00974.x Wang, Q. (2008). On the cultural constitution of collective memory. Memory, 16, 305–317. doi:10.1080/09658210701801467 Wang, Q. (2013). The autobiographical self in time and culture. New York: Oxford University Press. doi:10.1093/acprof:oso/9780199737833.001.0001 Wang, Q. (2016). Why should we all be cultural psychologists? Lessons from the study of social cognition. Perspectives on Psychological Science, 11, 583–596. doi: 10.1177/1745691616645552 Wang, Q., & Conway, M. A. (2004). The stories we keep: Autobiographical memory in American and Chinese middle-aged adults. Journal of Personality, 72, 911–938. Wang Q., Conway, M., Kulkofsky, S., Hou, Y., Mueller-Johnson K., Aydin, C., & Williams, H. (2009). The “egocentric” Americans? Long-term memory for public events in five countries. Cognitive Sciences, 4, 111–120. Wang, Q., & Fivush, R. (2005). Mother-child conversations of emotionally salient events: Exploring the functions of emotional reminiscing in European American and Chinese Families. Social Development, 14, 473–495. Wang, Q., Lee, D., & Hou, Y. (2016). Externalizing the autobiographical self: Sharing personal memories online facilitated memory retention. Memory. doi: 10.1080/ 09658211.2016.1221115. Wang, Q., & Ross, M. (2005). What we remember and what we tell: The effects of culture and self-priming on memory representations and narratives. Memory, 13, 594–606. doi:10.1080/09658210444000223 Wang, Q., & Ross, M. (2007). Culture and memory. In H. Kitayama & D. Cohen (Eds.), Handbook of cultural psychology (pp. 645–667). New York: Guilford Publications. Wickens, C. D. (2002). Multiple resources and performance prediction. Theoretical Issues in Ergonomics Science, 3, 159e177. doi: 10.1080/14639220210123806. Wright, D. B. (1993). Recall of the Hillsborough disaster over time: Systematic biases of flashbulb memories. Applied Cognitive Psychology, 7, 129–138. Wright, D. B., & Gaskell, G. D. (1995). Flashbulb memories: Conceptual and methodological issues. Memory, 3, 67–80. Wu, H. D. (2000). Systemic determinants of international news coverage: A comparison of 38 countries. Journal of Communication, 50, 110–130.

13 THE STUDY OF FLASHBULB MEMORIES The good, the bad, and the way forward Lucy V. Justice, Shazia Akhtar, and Martin A. Conway

One of the problems of the past 40 years of Flashbulb memory research is the extreme variability of published papers in terms scientific quality – ranging from the unacceptable to the outstanding. In this review chapter we take the opportunity to appraise the field and highlight the good and the bad. Major advances have been made in modelling and in collecting high quality data with impressive statistical power and we conclude that theorising about Flashbulb memories should be based on these high quality research programs and that the findings of other lower quality projects can be safely ignored.

Flashbulb memories (FBMs) are unusual memories; of the thousands, probably tens or even hundreds of thousands of items of news a person is exposed to during the course of their lifetime, very few result in memories of the reception event (who, what, where, when, and maybe some idiosyncratic detail) in which the item of news was first encountered. In our laboratory, over a series of years, very few participants have been able to recall more than a handful of news events for which they can also remember any details of the reception event. So within the domain of memory for news events, although many items of news can be recalled, vanishingly few of those news memories have the flashbulb quality of memory for the reception event. As such they make the ones that do, the FBMs, unusually rare and worthy of study simply for that reason. Interestingly, it may well be the case that over a period of hours, let’s say in the course of the same day, most people almost certainly can remember the reception event for a recently encountered items of news, showing that a memory was created in long-term memory. As the retention interval increases over weeks, months, years, etc., in nearly all cases, although memory for the news items remains, memory for the reception event is lost. This too is a peculiarity of long-term memory for news. One possibility is that memory for news items becomes assimilated to knowledge

264 Justice et al.

about the news of a person’s life (Brown, Shevell, & Rips, 1986) and in the process of this schematization, memory for the reception event is lost. Something like this seems to occur in the acquisition of many types of knowledge (Conway, Gardiner, Perfect, Anderson, & Cohen, 1997), that is, that episodic memories of learning new information lose their episodic qualities as the new knowledge is processed repeatedly and in different contexts. Possibly the associated episodic memory is simply forgotten, deleted from long-term memory in some way. Or, intriguingly, perhaps we simply lose access to them although they remain in long-term memory. Why there should be such selective forgetting/loss of access of long-term episodic memories of reception events is not known but as the chapters in this book show there may be good reasons for the selective retention of some episodic memories for reception events in which certain items of news were first encountered. However, there is a caveat; it unfortunately turns out that the study of FBMs is fraught with problems and this is hardly helped by entrenched theoretical positions. To their credit, several of the preceding chapters were explicit about these and sought genuine solutions to further the field.

Measuring Flashbulb memories As several authors point out, the field of FBM research is populated by published research of highly variable quality (see Rice, Hamamouche, & Bohannon, Chapter 5, this volume).What this means, in our view, is some research can be safely discounted. For example studies that have any (or usually most) of the following characteristics: small sample sizes, bad statistics, introspective ratings that are massively open to influence by beliefs and demand characteristics of the settings in which they are collected (all, of course, unacknowledged), no justification for why the news event selected might be expected to lead to widespread FBM formation in the population selected for study, highly selective coverage of the literature, and entrenched partisan theorizing. The fact that many of these misleading and low quality papers have seen publication in some of the top journals in Psychology is, sadly, a testament, to utterly inept (and in some cases, we suspect, biased) journal editing. We believe that this sorry state of affairs arises from a major academic schism between Neisser and Brown and, more generally between Neisser and the rest of the memory research world. Neisser (see Neisser, 1976) strongly believed and persuasively argued that all memory was schematic. (He also held that memory research in the verbal learning tradition, i.e. at the time, the past 80 years or so of memory research, had discovered nothing of importance). The Brown and Kulik (1977) paper and the notion of FBMs that it introduced flatly contradicted Neisser’s view. As did, it has to be said, Tulving’s notion of episodic memory (Tulving, 1972) that eventually provoked Neisser’s counter-proposal of “repisodic” memory (Neisser, 1981); a concept that subsequently has had little impact, unlike the concept of episodic memory. It is unfortunate that Neisser’s adherence to a schematic view of memory so polarized the field, leading researchers to play what Curci and Conway, 2012, termed the Flashbulb memory game, i.e. on the basis of questionable research,

The study of Flashbulb memories  265

failing to find evidence to support the FBM hypothesis (see Conway, 1995). Neisser’s influence upon memory research was, in fact, significant and profound and best embodied in his brilliantly edited work Memory observed (Neisser, 1982a).1 A book that changed memory research forever and in a wholly positive way, opening up as it did whole new vistas of memory research. One of those vistas was FBM research. Issues of measurement, methods, and modeling were covered in Chapters 1 (Curci), 2 (Wright & Carlucci) and 3 (Luminet). Curci’s excellent review of dimensional vs. categorical modeling of variables thought to define and characterize FBMs notes that categorical models highlight a discontinuity in memory between Flashbulb and other types of autobiographical memories. Curci’s preference is for categorical models, although they do raise a somewhat fundamental issue. In these models FBMs are considered “whole” units in long-term memory, (a position the schematic view of memory would find unacceptable), rather than memories constructed by retrieval processes from multiple sources of knowledge in memory. Wright and Carlucci (Chapter 2), on the other hand, are concerned that appropriate measures are taken depending on the nature of the questions asked: causal or associative. Whether such a distinction can be sustained in practice and not just in theory will, we suspect, turn out to be far from straightforward, although the strong recommendation to carefully consider what measures might in fact measure is without doubt wise counsel and a recommendation that should be taken up widely in the study of FBMs. We will return to this point shortly. At the heart of this chapter is a highly contentious issue and that is the issue of accuracy. As we have tried to show in previous work, the issue of accuracy in autobiographical memory research is a highly complex one (Conway, 2005). FBM researchers perhaps need to develop a concept that we term psychological accuracy. A stable and consistent mental representation that gives rise to a feeling of remembering, provides information relevant to the self (this can be non-conscious), and which intersects with the past to at least some degree, is in our view psychologically accurate. The relationship between psychological accuracy and veridicality is, however, complex and will have associative, causal, and other types of qualities. Luminet (Chapter 3) surveys and contrasts a number of structural equation models (SEMs) for a range of events that were thought have given rise to FBM formation. Of particular interest in this review is how the pathways between different latent structures vary in their degree of relatedness, sometimes showing strong correlations, e.g. rehearsal, and sometimes not (rehearsal again being a prime example). As Luminet points out, these difference across models clearly relate to the type of event selected for study with events of high social relevance giving rise to models in which emotion and self-relevance play important roles in the encoding and subsequent retention of FBMs. This is important because type of event is not a fixed variable but rather it is a random variable and because of that cannot be expected to show high consistency across different types of events. Following Luminet’s lead, we suggest that a distinction should be made between media events and FBM events. The later will have high social- and self-relevance, consequentiality for individuals, groups, and, perhaps, society, and lead to the widespread formation of FBMs in a target population.

266 Justice et al.

Media events, in contrast, although attracting much media coverage will not lead to the widespread formation of FBMs. For example, in our study of FBMs for the unexpected resignation of the British Prime Minister Margaret Thatcher (Conway et al., 1994) we also included, as a control event, a questionnaire on memory for the resignation of Thatcher’s Foreign Secretary Geoffrey Howe (effectively Thatcher’s second in command). Howe’s resignation which occurred about six months prior to Thatcher’s, was sensational and attracted intense media coverage, and his famous resignation speech in the House of Commons was widely viewed as sounding the death knell for Thatcher’s Prime Ministership. However, of the hundreds of people we sampled, few had any memory for the reception event of the news of Howe’s resignation. Thus, one distinguishing feature of media events may be intense media coverage that does not trigger widespread FBM formation in a selected population. We suggest that the news of the Challenger Disaster was a media event for populations who had not grown up in the ‘space age’, e.g. the students who were mostly used as respondents in the Challenger studies. Other media events are the acquittal of O.J. Simpson, which surely was not a flashbulb event for wealthy middle class white students, but may well have been for African Americans. Similarly the death of the pop singer Michael Jackson may have been a media event for most people but was, perhaps, a flashbulb event for devoted fans of the deceased singer. News events take on a media or flashbulb nature for different groups and FBM research needs to take this carefully into account. Indeed, the results of studies that were not of FBM causing events in the populations sampled should be discounted and that is simply because they are not studies of FBMs but, rather, are studies of memory for media events (Day & Ross, 2014; McCloskey, Wible, & Cohen, 1988; Neisser & Harsch, 1992; Schmolck, Buffalo, & Squire, 2000; Weaver, 1993; may be examples of studies of memory for new events but not flashbulb events). These three opening chapters highlight important issues in the study of FBMs: the nature of mental representation of FBMs, their accuracy (or rather, what accuracy might mean in this context), and how different types of SEM models might indicate FBMs and non-FBMs. Chapters 4, 5, 6, and 7 deal with important issues too but each in a different way and accordingly we briefly comment on each of them separately.

Confidence, consequentiality, and the experimental and clinical study of Flashbulb memories In Chapter 4, Talarico and Rubin pursue Rubin’s long and firmly held view that FBMs and, indeed, memories of traumatic events generally, are the product of “ordinary” memory processes. Exactly what these “ordinary memory processes” are is unclear and although in memory research generally we do now know quite a lot about memory retrieval, the nature of knowledge representation in long-term memory, and the underlying neurological substrate, we know comparatively little about the formation of memories – which is what the study of FBMs is largely concerned with. Thus, most of what we do know about memory formation/encoding

The study of Flashbulb memories  267

in the study of autobiographical memory comes from the study of FBMs.To a large extent Talarico and Rubin wrestle with an old ghost – Brown and Kulik’s original neurological model of FBM formation, the famous “Now Print!” hypothesis. It is fair to say that no one now subscribes to this model although it does not follow that there is not something different about the formation of Flashbulb and, indeed, trauma memories at the neurological level. Quite simply we do not know, the views of these authors notwithstanding. Intriguingly, however, Phelps’ group (see Sharot et al., 2006) in a neuroimaging study found areas of the amygdala to be more active in people recalling FBMs of the 9/11 attack on the World Trade Centre in New York. As the amygdala is associated with negative emotion, and many FBMs are of negative events, it might be that processes mediated by amygdala networks are specialized in some way for the formation and long duration of memories of such negative events and, importantly, that these same processes do not mediate formation, retention, and retrieval of events that do not feature negative emotions. One of Talarico and Rubin’s main criteria for identifying FBMs is that they should differ in some marked way from “ordinary” autobiographical memories. But what are “ordinary” autobiographical memories? We suggest that there is no such thing. All memories differ in their content and the ways in which they are formed and later constructed vary both within and between people. Thus, it seems to us that the proper control in an FBM study is not just any other memory that is not of a news event, but rather a memory that is of another news event. Or perhaps, a memory that is of an emotionally negative event. A further control is sampling across at least two populations, one of which is expected to show widespread FBM formation and the other of which is not as, indeed, Brown and Kulik did in their original study. In Chapter 5, Rice, Hamamouche, and Bohannon III, in a salutatory piece, provide a brief review of research that has generally been critical of the FBM hypothesis (e.g. McCloskey et al., 1988; Neisser & Harsch, 1992; Schmolck, Buffalo, & Squire, 2000; Weaver, 1993, among others). Characteristic features of these types of studies is that they largely use small groups of young adults all from the same group (North American university students). No attempt is made to justify or identify why a chosen event is likely to cause FBMs in this group and no comparisons can be made to other groups who would not show widespread FBM formation. Thus, few, and usually no, controls are possible. But of even more concern are the very small groups of respondents who take part. Low sampling sizes produce highly variable and unstable data sets that often do not replicate. Results from studies that have unacceptably low power and that do not justify the event they have chosen to study should in our view be discounted. They are simply contributing noise to the field and do not help us advance our understanding of the flashbulb phenomenon. These critical but justifiable points by Rice et al., are well taken and should be acted on. This chapter then moves onto a well-informed account of the modern reconstructive view of memory and develops a sophisticated approach to consequentiality and shows how this may play a significant role in the formation of FBMs.

268 Justice et al.

In Chapter 6, Lanciano, Curci, and Sartori give a very useful review of survey studies of FBMs compared to approaches that are more laboratory based. They then develop an autobiographical Implicit Association Test that aims to implicitly investigate the content of memories by contrasting verification time to true and false statements, some featuring details of a memory. It will be most interesting to see how this approach develops, particularly as implicit measures, when used with care, may be less prone to uncontrolled influences. The measures typically taken by FBM researchers such as ratings of vividness, personal importance, self-relevance (which is why Chapter 5 is important), and so on, can hardly be described as “process pure”. Indeed one urgent research agenda should be to establish what exactly it is these introspective measures are in fact measuring. Take for example ratings of vividness. Cultural demand characteristics may have a powerful effect on such ratings. If one is a U.S. citizen over the age of about 20, is it possible not to have a vivid memory for learning of 9/11? Similarly for many middle-aged UK citizens, is it possible not to vividly remember the death of Princess Diana? Many measures of factors thought to be relevant to FBMs are in fact open to such demand characteristics.Yet others may involve acts of introspection that are beyond the abilities of many people (i.e. judging memory perspective) and are therefore responded to on how a respondent understands what is being asked, rather than on the basis of a perceived mental state or property of such a state. Implicit measures might just get round these problems although as Lanciano et al. point out, these too have their own problems. In Chapter 7, Tat, Gold, and Budson provide a valuable review of what little is known about FBMs in clinical populations with brain damage and suffering from psychological illness. The findings as might be expected are mixed. Amnesic patients, especially those with damage to the medial temporal lobe memory system, show no persistent memories of flashbulb events in contrast to patients with damage elsewhere, such as the frontal lobes. Nevertheless, in all cases there are differences from non-brain damaged controls. Interestingly patients with posttraumatic stress disorder had FBMs broadly similar to those of controls with other psychological illnesses.

The social-cultural functions of Flashbulb memories Finally, the last five Chapters (8, 9, 10, 11, and 12) all deal with social aspects of FBMs. These chapters essentially take as their starting point Neisser’s suggestion that FBMs represent the intersection of the personal with the historical. They are moment’s where we line up our lives with the history of our (lived) time and in effect say “I was there” (Neisser, 1982b). Stone and Jay (Chapter 8) make the useful point that flashbulb events are typical of negative occurrences but that there are also positive flashbulb events; they cite the example of the election of President Barack Obama, the first African American president of the United States of America. A positive event for some, but by no means all, U.S. citizens. The fact that there are FBMs for emotionally positive news events is hardly surprising, however, as these authors point out, the fact they have only rarely been studied is a lacuna in

The study of Flashbulb memories  269

the field of FBM research and one that needs rectifying. Positive flashbulb events and the FBMs to which they give rise may have an important effect on collective self-confidences, perhaps even on a collective subjective sense of well-being. More studies should, then, aim to address these important issues. Berntsen in Chapter 9 persuasively argues that FBMs arise because they impact on social identity. Indeed, this was one of the main points of the original Brown and Kulik (1977) paper: African Americans had more FBMs for learning about the murder of Martin Luther King than did Caucasian Americans. This too was the main point of our own work showing widespread FBM formation for Thatcher’s resignation in UK citizens but not in citizens from other countries (Conway et al., 1994). Social identity is, however, complex and although it would be fascinating to map the memories that help support and define social identity across a wide range of groups, there are some flashbulb events that seem to lead to FBM formation across different groups. For example, virtually all of Brown and Kulik’s respondents, black and white Americans, had FBMs for the assassination of the American President John F. Kennedy. We suspect news of 9/11 similarly lead to FBM formation across many groups with different social identities. In a similar vein, Hirst and Meksin (Chapter 10) argue that FBMs are shared by “communities” (although what is or is not a “community” is not entirely clear). This is undoubtedly true but how exactly it works will extend beyond FBMs to other shared memories and knowledge. Collective FBMs may be important in a community but the degree to which they are shared by their significance or meaning for individuals will vary very considerably. Research exploring such individual differences in shared memories could prove of considerable importance to understanding the memorial basis of community/social identity. Páez and colleagues (Chapter 11) consider the nature of FBMs and what they term “collective” memories. From their analysis it seems that FBMs are simply a distinctive type of collective memory and contribute along with other shared memories and shared knowledge to collective identity important qualities such as collective self-esteem. These authors also suggest that collective identity may form during a critical period, age 10 to 30, the period of the reminiscence bump (for similar suggestions see Conway, 1997). Wang and Aydin (Chapter 12) insert a much-needed cultural dimension into theorizing about the social or collective functions of FBMs. In their model, culture and the type of self-focus supported by different cultures takes a central role. To their credit, these authors are aware that many other factors such as beliefs, culturally specific practices, shared knowledge, and so forth contribute to cultural generational identity but argue that FBMs may play an important part in this nexus of knowledge and memory that helps constitute cultural identity.

The way forward In briefly reviewing the chapters in this book, a number of more general points occurred to us and we summarize those here: •

FBM researchers need to identify those studies that are scientifically sound and develop theories based on the findings of those studies, and those studies only.

270 Justice et al.

• •

•

•

Large samples sizes yielding good power should always be used in surveys of FBMs. Possibly one of the most important ways forward will be in the development of high quality neuroimaging studies (this might include more experimental procedures such as those outlined by Curci in Chapter 2). A caveat here, however, is that brain imagining could not be used to study encoding, although it might be used to study consolidation. Better models and more detailed formal models of the social and cultural functions of FBMs need to be developed (derived perhaps from the models described by Luminet in Chapter 3). Research into what exactly is “shared” in so called collective memories needs to be conducted. We all know 9/11 happened but our (the authors’s) memories differ from each other’s and also from yours. So, what is collective? What is shared?

We are completing this brief overview in January 2017, 40 years after Brown and Kulik coined the term “Flashbulb memory” in their seminal paper. Studying FBMs is a difficult endeavor, however there have been some excellent large-scale surveys since that original paper and these have facilitated the development of models that are more explanatory. What seems to be a shortcoming is the integration of FBM theory and research with other memory research, particularly autobiographical memory research. Focusing on the functions of FBMs, especially their role in underpinning identity may help develop and integrate FBM theory more widely.

Note 1 Martin Conway was fortunate enough to meet Dick Neisser at Cornell some years ago and, being the person he is, recounted his FBM of buying Memory observed.The great man was not amused, but nonetheless graciously took Conway and his colleague Qi Wang out for a very pleasant dinner.

References Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. Brown, N. R., Shevell, S. K., & Rips L. J. (1986). Public memories and their personal context. In D. C. Rubin (Ed.), Autobiographical memory (pp.137–158). Cambridge, UK: Cambridge University Press. Conway, M. A. (1995). Flashbulb memories. Brighton, Sussex: LEA. Conway, M. A. (1997). The inventory of experience: Memory and identity. In D. Jodelet, J. Pennebaker, D. Páez, & B. Rimé (Eds.), Political events and collective memories: Social psychological perspectives (pp. 21–46). London: Routledge. Conway, M. A. (2005). Memory and the self. Journal of Memory and Language, 53, 594–628. Conway, M. A., Anderson, S. J., Larsen, S. F., Donnelly, C. M., McDaniel, M. A., McClelland, A. G. R., . . . Logie, R. H. (1994).The formation of flashbulb memories. Memory & Cognition, 22, 326–343.

The study of Flashbulb memories  271

Conway, M. A., Gardiner, J. M., Perfect, T. J., & Anderson, S. J., & Cohen, G. M. (1997). Changes in memory awareness during learning: The acquisition of knowledge by psychology undergraduates. Journal of Experimental Psychology: General, 126, 1–21. Curci, A., & Conway, M. A. (2012). Playing the flashbulb memory game: A comment on Cubelli and Della Sala. Cortex, 49, 352–355. Day, M. V., & Ross, M. (2014). Predicting confidence in flashbulb memories. Memory, 22, 232–242. McCloskey, M., Wible, C. G., & Cohen, N. J. (1988). Is there a special flashbulb-memory mechanism? Journal of Experimental Psychology: General, 117, 171–181. Neisser, U. (1976). Cognition and reality. San Francisco: W. H. Freeman. Neisser, U. (1981). John Dean’s memory: A case study. Cognition, 9, 102–115. Neisser, U. (1982a). (Ed.). Memory observed: Remembering in natural contexts. San. Francisco: W. H. Freeman. Neisser, U. (1982b). Snapshots or benchmarks? In U. Neisser (Ed.), Memory observed: Remembering in natural contexts (pp. 43–48). San Francisco: W. H. Freeman. Neisser, U., & Harsch, N. (1992). Phantom flashbulbs: False recollections of hearing the news about Challenger. In E. Winograd & U. Neisser (Eds.), Affect and accuracy in recall: Studies of “flashbulb” memories (pp. 9–31). New York: Cambridge University Press. Schmolck, H., Buffalo, E. A., & Squire, L. R. (2000). Memory distortions develop over time: Recollections of the O.J. Simpson trial verdict after 15 and 32 months. Psychological Science, 11, 39–45. Sharot,T., Martorella, E. A., Delgado, M. R., & Phelps, E. A. (2006). How personal experience modulates the neural circuitry of memories of September 11. PNAS, 104, 389–394. doi: 10.1073/pnas.0609230103 Tulving, E. (1972). Episodic and semantic memory. In E. Tulving & W. Donaldson (Eds.), Organization of memory (pp. 381–403). New York: Academic Press. Weaver, C. A. III. (1993). Do you need a “flash” to form a flashbulb memory? Journal of Experimental Psychology: General, 122, 39–46.

INDEX

Bold page numbers indicate figures, italic numbers indicate tables. 9/ll terrorist attacks see World Trade Center attacks accessibility 22, 23, 51, 171, 206, 242, 249, 251, 252, 254, 264 accuracy of FBMs 35, 41; autobiographical memory 127, 265; and confidence 78–9, 127; and consistency 67; control procedures 126; event memory 208; experimental approach 121; eyewitness testimony 127, 132; and measurement of FBM 75–6; models of FBM formation 63–4, 66, 67, 68; objectivity 76; ordinary/special encoding hypothesis 75–6; Pearl Harbor attack 76, 99; peripheral/central details in emotional memories 125–6; and positive/ negative emotion 81, 249; psychological 265; stability of memory over time 127 AD see Alzheimer’s disease affect 51; see also emotional ... (entries starting with) Affect and accuracy in recall: Studies of FBM (Winograd and Neisser) 9 affective impact 39, 52, 59, 61, 67, 79, 81–2, 87, 110–11, 114–15 age: and collective memories 212–13; social identity 225–6, 227–8 Agypt, B. 108 Ahmetzanov, M.V. 43, 123, 124 AIDS, family death due to 80, 97–8 alcoholism 146 Alito, Samuel 213 Alzheimer’s disease: emotional memories 139–40, 152–3; and FBM 138–46, 142,

143, 144; future studies 150–1; US election 2008 143–5, 145 amnesia 106–7, 146–7 amygdala 99, 139, 140, 267 aneurysm 147 anthropological studies 40 anxiety 20, 171, 246; see also worry appraisal: FBM models 54, 59; and feeling states 54; importance of the event 54, 120, 186, 188, 205, 226, 230, 231, 233; of novelty 51–2, 59, 66 arousal see emotional arousal associative hypotheses 38 Atatürk, Mustafa Kemal, death of 226, 231 attention-capturing properties 73 attention focusing, negative events as 172, 175, 229 attitudes: and culture 249, 254; FBM models 53, 55, 56, 66, 68; Implicit Association Test (IAT) 127–8; see also cross-cultural studies; cultural issues; psychosocial processes Aue, W. 76, 97 Australia, resignation of PM 66, 249, 251 autism spectrum disorders 20 autobiographical Implicit Association Test (aIAT) 76, 128, 129–32, 130, 268 autobiographical memory: accuracy 127, 265; Alzheimer’s disease dementia 138–9; constructivist model 22–3, 25; cross-cultural studies 252, 253, 254, 257; differences to FBMs 74; discontinuity 23–5; emotional features 186; and FBM 20–5; functions of 172; knowledge base structures 22, 23–4, 26, 244; limited

Index  273

theory of 87; ordinary 74; remember/ know memory distinction 34–5; surprise 84 autonoetic consciousness 34 aversive events 108–10 Bable, R. 109 Banaji, M.R. 44 Bandura, A. 102 Barga, S. 106 Bartholomew, D.J. 36, 37 Baudouin, King 222, 230, 231, 234 Belgium, WWI memories 3 Belgium, WWII memories 222 Bellelli, G. 231 Belli, R.F. 254 belongingness 187, 197 Benedict XVI, Pope, resignation of 131, 192 Berlin wall, fall of 193–4, 224, 227, 228, 233 Berntsen, D. 75, 76, 82–3, 88, 165, 173, 184, 190, 192, 193–6, 207 between-person rehearsal 110–11, 114, 243 biases: distinctiveness 171–2; fading affect 110, 111, 114, 171; memory 40, 125, 193, 196–7 (see also distortion); negative, in FBM research 170–2, 229; pleasantness 81, 171–2; selection 171, 172; sociocentrism 225; spatial memory 125 Bilbao, M.A. 173 BILOG-MG software package 17 biological mechanisms 184–5; see also adaptational value, FBM biopsychosocial functions of memories 172–6 Bischoping, K. 254 Bluck, S. 106 Boals, A. 81, 167 Bohannon, H. 103, 109 Bohannon III, J.N. 21, 76, 78, 99, 100, 104, 106, 107–8, 108, 109, 120, 208, 210 Bohn, A. 165, 184, 193 brain tumours/lesions 146, 147, 154–5 Breslin, C.W. 166 Brewer, W.F. 81 Brewin, C.R. 87 Brown, A.D. 80, 167 Brown, A.S. 85 Brown, N. 100 Brown, R. see Brown & Kulik model Brown & Kulik model 2, 34, 35, 37, 42, 44n1, 51, 54, 57, 73, 80–1, 96, 105, 126, 183–5, 185, 189, 190, 196, 246, 264, 267; citations/use of phrase, FBM 41, 42; critiques 12–13, 33; FBM formation 49; main findings 12–13; and measurement

of FBM 12–13; multiple events, use of 41; operational criteria 12; strengths/ limitations 12–13, 33 Budson, A.E. 139, 140–3, 142, 143, 143–5, 150–1, 152–3 Bull, M.P. 192 Buschmann, F. 148 Bush, George, President 211, 212 Butler University 104 California earthquake 58, 190 Campbell, A.M. 103 Candel, I. 146–7, 154 canonical: categories 64, 77, 86, 120, 126, 184; features 112–13, 204; life periods 254; variables 1 Carlson, E.B. 20 Carmona, J. 175 case study approach 40–1 categorical models: issue with 265; vs dimensional models 2–3, 14–19, 16, 17, 24 categorical variables 13, 24 CatPCA (categorical principal component analysis) 13 causal: inferences 39–40; relationships 39–40 central information 76–7, 125–6 ceteris paribus conditional 39 CFA (confirmatory factor analysis) 24, 49 Challenger disaster: collective memory 208; and consistency of memory 97, 99; culture/psychosocial processes 266; and event memory 97, 208; involvement in event 21; media coverage of event 210, 210; patterns of forgetting 97; rehearsal and initial arousal 108; self vs the other distinction 103; sustainability of collective memories 208; weighted attribute scores (WAS) 13 Change in Life Events and Identity (CEVI) research program 222–3 Charron, A. 109 Chatelain-Jardon, R. 175 Cherry, K.E. 75 children: divorce of parents 108; emotional memory 248–9, 250–1; past injuries, memories of 102 Christianson, S. 77, 103 Chwalisz, K. 171 clarity of memory 35, 36 Clark, C. 3 clinical populations 268; Alzheimer’s disease dementia 138–46, 142, 143, 144, 152–3; amygdala pathology 138, 139, 140; brain tumours/lesions 146; dementia 138–46; depression 18; emotional memories 139–40, 152–3;

274 Index

Korsakoff syndrome/focal brain lesions 146–7, 151, 154; limitations of research 150–1; mild cognitive impairment (MCI) 138–46, 152–3; post-traumatic stress disorder (PTSD) 148–50, 155, 157; temporal lobe epilepsy 147–8, 151, 156, 157; term definitions 137–8; variables measured in studies 138 cluster analysis 18 cognitive overload 244 cognitive processes 189, 196, 250 cognitive psychology 3, 5, 40, 41, 196 Cohen, A.A. 242 Cohen, N.J. 97 cohort design 157 Colegrove, F.W. 1 Collective: behaviors 231–2; catastrophes 222, 223; functions of FBMs 175–6; violence 222, 223, 223, 224; see also rituals collective identity: and collective memories 213–14; media shaping of 214–15 collective memories: affected community, defining 205; and age 212–13; Challenger disaster 208; characteristics of 220–2, 221; and characteristics of FBMs 206, 220–2, 221; community phenomenon, FBMs as 204–7; congruence with cultural values 232–3; consequentiality 230–2; consistency 206; content of, FBM comparison 222–5, 223; content of FBMs as function of community membership 207; critical period 225–6, 227–8; defining communities 205; emotional intensity 229–30; and event memory 203; future studies 269; and history 203; importance of the event 230–3; lived 202–3; media, and critical facts in 211–12; media coverage of events 209–10, 210; media exposure 230–1; negative events 81; novelty and surprise 226, 229; positive/ negative events 229–30; predictors, comparison with FBMs 226, 229–36, 234, 235; relevance of events for social identity 232; reporting of FBMs within community 204–5; self-esteem 232–3; social artifacts and practices 207–8, 209; as social representations 219–20; social sharing 230–1; specificity of memories to a community 205; sustainability of 208; valence 229–30 collective rituals see rituals collectivism 43, 66, 83, 247, 248–9, 252–3; see also individualism; social identity

Columbia space shuttle disaster 97; see also Challenger disaster Coman, A. 80, 167 commemoration 207, 231–2 communication see verbal communication community: affected, defining 205; and characteristics of FBMs 206; consistency of memories 206; content of FBMs as function of membership 207; defining communities 205; and FBMs 43, 201–2, 204–7; future studies 269; reporting of FBMs within 204–5; social artifacts and practices 207–8, 209; and social identity 205; specificity of memories to 205; see also collective memories; cultural context; cultural issues Comparative Fit Index (CFI) 69n3 conditions for formation of FBMs 79–82, 87 confidence in memory 48, 50, 78–9; and accuracy 78–9, 127; attachment to subject of FBM 86; and consistency 64–6, 65; and emotional intensity 85; and FBM as distinct category 79; mutable, memory as 100; positive/negative events 163, 169; and rehearsal 85 consequentiality 53, 58–9; Barack Obama, election of 110–15; Baudouin, King 231; collective memories 230–2; and consistency 80; Conway et al. model 100; cultural issues 240, 243–4, 245, 246, 251, 256; first-hand experiences 102–3; model comparisons 79–80; ordinary/special encoding hypothesis 79–80; see also taxonomy of consequentiality consistency 35, 41, 48–9; and accuracy 67; collective memories 206; and confidence in FBM 64–6, 65; and consequentiality 80; and delay 77; and emotional intensity 85; and forgetting 99, 204; measurement of FBM 76–7; in memory 99; ordinary/ special encoding hypothesis 76–7; positive/negative events 169, 169; and rehearsal 66; space shuttle disaster 97; World Trade Center attacks 2, 204 consolidated memories 143, 231 construct validity 11; see also measurement of FBM constructivist view: cyclic retrieval 22; models of FBM formation 22–3, 25 continuous measures 14, 19, 24, 36–7 conversations see verbal communication Conway, M.A. 104, 190, 192, 246, 247, 254

Index  275

Conway et al. model 50–2, 52, 57, 87, 100, 185–6 Cook, S.P. 147 correlation and causation 39–40 critical period 225–6, 227–8 Cross,V.S. 21 cross-cultural studies: autobiographical memory 257; consequentiality 80; emotionality 247–9; memory and selfidentity 252–3; personal importance 245–7; rehearsal/social sharing 249–51; relevance of events for social identity 232; World Trade Center attacks 192, 206; see also cultural issues; psychosocial processes Crowder, R.G. 44 cultural ambience 246–7 cultural beliefs 256; see also cross-cultural studies; cultural issues; psychosocial processes cultural context: memory and self in 252–3; predictors of FBM formation 66; significance of events 67 cultural issues: collective memories 232–3; collectivism 247; consequentiality 240, 243–4, 245, 246, 251, 256; culturally embedded FBM model 240–1, 241, 245; emotional determinants 247–9; emotional expressivity 248–9, 250–1; emotionality 248–9; future research 243, 256–7; importance effect 245–7; individualism 247; information transmission/mass media 242–5; lifespan memory retrieval 253–6, 255; macrolevel variables 242–3; memory and self-identity 252–3; personal importance 245–7; phenomenological qualities of FBMs 246; rehearsal 249–51; selfdefining functions of FBM 251–3; social identity 269; social sharing 249–51; universal mechanisms, cultural variations in 245–51; see also community; crosscultural studies; psychosocial processes cultural relevance vs irrelevance 50; see also cross-cultural studies Curci, A. 13, 21, 24–5, 37, 43, 50, 55–6, 59–60, 77, 87, 105, 122–4, 123, 125–6, 131, 186, 190, 192, 205, 232, 246 cyclic retrieval 22 Cyr, T. 212–13 da Costa, S. 173 Davidson, P.S.R. 83, 147, 154–5 Day and Ross model (2014) 64–6, 65, 86

de Waal, F.B.M. 104 death: of famous people 222, 223 (see also specific individuals); of a parent 102 (see also first-hand experiences) degenerative research 44 delay to first memory assessment 75, 78, 100 delayed recall 77 dementia 152–3; see also Alzheimer’s disease Demiray, B. 174 Denmark, invasion/liberation, WWII 76, 84, 173–4, 192, 193–6, 207, 230, 233 Denver, J.Y. 75 depression 18, 20 detail: positive/negative events 169, 169–70; symbolic meaning of 196 developmental perspective see children Diamond, D.M. 103 Diana, Princess 105; collective memory 205; conditions for FBM formation 3; cross-cultural studies 192; epilepsy, patients with 148 diencephalon amnesia 147, 154 Diener, E. 171 dimensional models vs categorical models 2–3, 14–19, 16, 17, 24 directive function of FBM 173–4 discontinuity in autobiographical memory 23–5 discovery: impending divorce 108; infidelity 109; narratives 109 discrete measures 17, 18, 19, 20, 35, 37 dissociation, pathological/nonpathological 20 distinctiveness: of memory 80–1; negative events 171–2 distortion, memory: clinical populations 141, 142, 144, 144–5, 150, 152, 157; and culture 245, 251, 256–7; measures 121–2; post-encoding factors 245; and social identity 197; symbolic meaning of details 196 divorce 108 documentary method 39 DRM procedure 99, 100 durability 183, 184, 187, 229–30 earthquakes see California earthquake; Iran earthquake; Kobe earthquake; Marmara earthquake; San Franciso earthquake; Turkey: earthquake eating disorders 20 economic crises 223, 224 Edery-Halpern, G. 81

276 Index

egalitarian values 257 Eisenberger, N.I. 103–4 elaboration, memory 77 elections 222, 223, 223 emotion, positive/negative see emotional valence emotional activation 233, 235; arousal: clinical populations 144–5; and cultural issues 248–9; and emotional memory 103; FBM models 52; measures of FBM 20–1; and social identity 189, 231 emotional affect: clarity of FBMs and 38–40; and measurement of FBM 20–1; model comparisons 81–2; ordinary/ special encoding hypothesis 81; structural equation modelling (SEM) 53–4; twopath model 53–4; feeling states 248–9 emotional attachment 208–9, 212 emotional determinants 247–9; factors, influence on FBMs 43 emotional events 43, 73, 120, 124–5, 146, 147, 186, 220, 229 emotional expressivity 248–9, 250–1 emotional-integrative model see Finkenauer et al. model emotional intensity 52, 53, 54, 58, 84–5; collective memories 229–30; future studies 88; surprise 84 emotional memory: Alzheimer’s disease patients 139, 140; children 248–9, 250–1; clinical populations 139–40; positive/ negative emotion 139 emotional narrative 139 emotional numbing 148; see also posttraumatic stress disorder (PTSD) emotional processes 54, 189, 196 emotional reaction: Baudouin, King 231; cross-cultural studies 249; Denmark, invasion/liberation, WWII 194; and FBM 2, 87; and individualism/ collectivity 253; and memory 38; and personal importance 51, 57, 186, 245; prior interest and knowledge 186; and psychosocial processes 188; and rehearsal 51, 188, 236; social identity 188, 197; see also surprise emotional socialization 248–9, 250–1 emotional valence 43; see also positive/ negative emotion emotions, social sharing of 231 encoding factors 51, 82, 121, 124–5, 132, 236, 245; see also consequentiality; importance; novelty and surprise; prior knowledge; special (flashbulb/memory) mechanism

epilepsy 147–8, 151, 156, 157 episodic memory: long-term memory 264; Tulving’s notion of 264; vs semantic memories 20–2 (see also autobiographical memory) Er, N. 97, 190 Er (2003) model 56–9, 58, 186 error sampling 38, 39 errors, memory 44, 73, 86, 121, 125, 207, 244, 257 euro-centrism 224 event memory: accuracy 67, 208; assessing, vs FBM 131; and collective memories 203; as distinct from FBMs 57; as form of semantic memory 21; maintenance of FBMs 54–5; measures 20–1; media coverage 209–10; and rehearsal 60–1, 62–4; World Trade Center attacks 20–1 event-specific knowledge 23 events: aversive 108–10; cultural significance of 67; physical vs social 101, 103–7; see also exceptionality of events; negative events; positive events everyday memory movement 44 evolutionary: function of negative events 174–5; evolutionary/survival value, FBM 98–9, 115; theory 102 exceptionality of events: accuracy 132; collective memory 213; consequentiality 213; consistency 66; novelty 56, 68; rehearsal 66; vs mundane events 50, 66, 68 expected events see novelty and surprise experimental approaches 43; accuracy of FBMs 121; autobiographical Implicit Association Test (aIAT) 128, 129–32, 130; contextual details investigated 121–6, 123; emotional memory 43, 123–5; explicit FBM measures 126–7; implicit measures 127–9; measurement of FBMs 122–4; methodology 43–4; peripheral/central details in emotional memories 125–6; reproduction of FBM characteristics 124–5; Stroop paradigm 124; vs naturalistic approaches 120–6, 122, 123 explicit FBM measures 126–7 expressive writing techniques see emotional expressivity Extraordinary memories for exceptional events (Schmidt) 89 eyewitness testimony 41–2, 127, 132–4 factor analysis models 14, 19 fading affect bias 110, 111, 114, 171

Index  277

Fahrenheit 911 (Moore) 212 false memories 214–15 falsification 38, 41 family and cultural issues 248–9, 250–1 faulty consolidations 140 FBM see flashbulb memories feelings: clinical populations 34; and cultural issues 248–9; and emotional memory 52; models 52, 54, 59, 60; see also affect; emotional.. (entries starting with) Fernández, J. 79 Finkenauer, C. 13, 53, 53–6, 192, 231, 232, 246 Finkenauer et al. model 13, 53, 53–6, 57, 68, 121, 186 first-hand experiences 48; and consequentiality 102–3; death of a parent 102; Er (2003) model 56–9, 58 Fisher, C. 109 fit indices 50, 60; Comparative Fit Index (CFI) 69n3; comparison 57 Fivush, R. 4 flashbacks see post-traumatic stress disorder (PTSD) flashbulb events: individual processing 82–7; see also collective memories; event memory; flashbulb memories flashbulb memories: biopsychosocial functions of 172–6; characteristics 75–9, 220–2, 221; and collective memories 220–2, 221; conditions for formation 79–82; conflicting research findings 183–6, 185; definitions 48–9, 89, 183; developing debate over 96–8; media popularity 1; methodological issues (see methodology); model comparisons (see models of FBM formation); negative (see negative events); positive (see positive events); processing 82–6, 87; prototypical case (Kennedy assassination) 183; reconceptualization 186–8, 188; as separate category 87; simplified model 32, 32; special vs mundane character 3; types of FBM hypotheses 37–43, 42; see also autobiographical memory; Brown & Kulik model; models of FBM formation flashbulb metaphor, use of term 41, 42 focal brain lesions 146–7 Fong, Bobby, resignation of 104 forgetting 48–9, 57, 99, 204; see also Alzheimer’s disease formation of FBMs: conditions for 79–82; see also models of FBM formation formative years of identity 225–6, 227–8

four Ws 35–6, 36 Franco, Francisco, death of 164, 220, 222, 230 Franz Ferdinand, Archduke, assassination of 3 Frederik IX, King 182–3 Freund, A.M. 174 frontal lobe lesions 147, 154–5 functions: biopsychosocial 172–6; collective 175–6; directive 173–4; self 172, 173; social 172–4 Galea, S. 171 Gallagher, D. 171 Gaskell, G.D. 36 gender differences: infidelity, discovery of 109; social identity 191 general knowledge 254 generational effect(s) 222, 241, 245, 253, 254–6, 257, 269 GENLAT software package 17 genotypic terms 19 Gil-Rivas,V. 171 Gisle, L. 13, 192, 232, 246 Glisky, E.L. 83, 147 goal structures 22; see also constructivist view goals: and cultural issues 240, 248, 252; generational differences 254–5; personal 22, 83 Goldstein, F.C. 139 Gooding, G. 107 Gratz, S. 21 group commitment 187, 194, 252 group differences 42, 189, 190, 191–3 group membership: exclusion from 103–4; future studies 89; identification with group 161–2, 247; in-group, effects on FBM 104–5; out-group, effects on FBM 105–6; perceptions of 187; proximity to events 106; and significance 82–3; see also community; social identity groups, comparison of 50, 51 Hamann, S.B. 139 Harsch, N. 13, 63, 97 hearing the news vs memories of first-hand experience 4 Heckman, D. 108 Hillman, J. 102 Hillsborough football stadium disaster 24, 40, 85 Hinton, M. 108 Hippocampus 99, 140 Hirst, W. 20–21, 80, 85, 167, 190, 191, 193, 204–5, 208, 210, 210, 211, 212–13, 214, 220, 222

278 Index

historical significance 3–4, 224 history, and collective memories 203 Holman, E.A. 171 Hornstein, S.L. 85 Hou,Y. 243–4 Howe, Geoffrey, resignation of 266 hypotheses: associative 38; causal inferences 39–40 ICC see Item Characteristic Curve (ICC) identity: autobiographical 172; collective (see collective identity); generational 254–5; individual/personal (see social identity); self-defining functions of FBMs 251–3 IIF see Item Information Function (IIF) Ikeda, M. 139–40 impaired coding 143; see also clinical populations Implicit Association Test (IAT) 127–9 implicit FBM measures 127–9; autobiographical Implicit Association Test (aIAT) 76, 128, 129–32, 130, 268; Implicit Association Test (IAT) 127–9 importance 51, 245–7; see also personal importance independent self 252–3 indicators (proxies of FBM accuracy): explicit/self-report 126; implicit 127–9 individualism 66, 247, 248, 252–3 inferences: associative 38; causal 39–40; local 38 infidelity, discovery of 109 information transmission 242–5; see also media exposure injury, personal 106–7 intensity, emotional 52, 53, 54, 84–5, 88 interdependent self 252–3 interest, prior 51, 186; see also prior knowledge Internet 243–4 interpretive approach 19, 81–2, 202, 229 intrusive thoughts 148, 149, 150; see also post-traumatic stress disorder (PTSD) inverted-U-function 103 Iran earthquake 183 IRT see item response theory (IRT) IRTPRO software package 17 Israel, terrorist attacks 81 item characteristic curve (ICC) 15–16, 16, 17 item characteristic function 15 item information function (IIF) 17 item response theory (IRT) 14–17, 16, 17, 36; see also latent class analysis Jackson, Michael, death of 64–6, 86, 266 Japan: nuclear accident 249; resignation of PM 66, 249, 251

Jasper, J.M. 213 Jelicic, M. 146 Jennings, E. 107 Jennings, L. 104 Jiang, T.L. 243–4 John Paul, Pope, death of 24–5, 56, 130–1, 192, 205 Jordan, H. 104 Julian, M. 76, 77, 100 Kennedy, President J.F. 233; assassination 2, 81, 192, 230; as prototypical FBM case 183, 186 Kensinger, E.A. 143–5, 165 Kiburz, K. 100 Killinger, W.A. 13 King, Martin Luther 105 Kizilöz, B.K. 77–8, 120 Knott, M. 36 knowledge: base structures 22, 23–4, 26, 244; prior 51, 59, 186 (see also novelty and surprise); remember/know memory distinction 34–5; semantic 146, 154, 231; structures 51; transmission of 242–5 Kobe earthquake 140 Kock, N. 175 Koh, J.B.K. 250 Koppel, J. 67, 80, 167 Korean War 229 Kornbrot, D.E. 246 Korsakoff syndrome/focal brain lesions 151, 154 Kozin, M. 78, 85 Kraha, A. 81, 167 Kulik, J. see Brown & Kulik model Kulkofsky, S. 249, 250 Kvavilashvili, L. 66, 86, 190, 192, 246 La Violencia 41 Lanciano, T. 21, 24–5, 37, 43, 56, 82–3, 87, 105, 123, 124–6, 130–1, 192 Lane, S.M. 75 Laney, C. 106 Lang, A. 104 Larsen, S.F. 21 latent class analysis (LCA) 18, 36–7 latent constructs 12, 14–20, 16, 17, 23–4, 49, 122 Latent GOLD software package 18 latent trait analysis (LTA) 14–17, 16, 17, 36–7 latent variable approach 23–4, 36, 53, 86, 122–4 Lazar, N. 97 LCA see latent class analysis (LCA) Lee, D. 244

Index  279

Leider, M. 103 lEM software package 18 Levine, L.J. 50, 168 Lieberman, M.D. 103–4 Life: epochs 23, 100–1, 254–6, 255; narratives 253 lifespan memory retrieval 253–6, 255 Lincoln, Abraham, assassination 1, 81, 233 linear structural equation models 109 Liu, A. 166 ‘live’ quality 126 local: independence 14; inference 38 long-term: consistency 193–6; durability 229, 244, 263–4, 265 longevity 12, 75 LTA see latent trait analysis Luminet, O. 13, 50, 56, 59–60, 63, 66, 77, 80, 83, 186, 190, 192, 205, 222, 232, 246, 249 Luo, J. 166 Lyons, K. 109 McCarthy, R.A. 148 McCloskey, M. 97 McIntosh, D.N. 171 MacKay, D.G. 43, 123, 124 macro-level culture 242, 243, 256 Maddalena, C. 21, 105, 192 Madrid terrorist attacks 231 Mahmood, D. 80 Malcolm X 205 Manier, D. 80 Mannheim 225 Marmara earthquake 56–7 Martinez-Zelaya, G. 173 Martini, M. 250 mass media see media; media exposure Mastandrea, S. 21, 105, 192 Mather, M. 212–13 McClure, L. 100 MCI see mild cognitive impairment (MCI) meaning making process 244 measurement of FBM: accuracy 75–6; additive model 63; affect 20–1; Brown & Kulik model, limitations of 12–13; categorical models vs dimensional models 14–19, 24; categories as real entities/discontinuity 19–20; CatPCA (categorical principal component analysis) 13; cluster analysis 18; conclusions 25–6; confidence 12; consistency 12, 76–7; continuous measures 14, 19, 24, 36–7, 122, 124; delay to initial test 75, 78, 100; detailed 13, 23, 25; dimensional models vs categorical models 14–19, 16, 17, 24; discrete measures 17–18, 19, 20, 35, 37;

experimental approaches 122–4; explicit measures 126–7; factor analysis models 14, 19; implicit FBM measures 127–9; item response theory (IRT) 14–17, 16, 17; latent class analysis (LCA) 18–19; latent trait analysis (LTA) 14–17, 16, 17; latent variable approach 23–4; longevity 12, 75; media exposures 21, 25; methodological/ theoretical issues 14–19, 16, 17; model (see Brown & Kulik model; models of FBM formation); Pearl Harbor attack comparison 76, 99; scoring 113, 126, 146; semantic vs episodic memories 20–2 (see also autobiographical memory); source of discovery 21; types of memories 34–5; vividness 12; weighted attribute scores (WAS) 13 media: coverage of events 209–10, 210; creation of emotions 248; critical facts in collective memories 211–12; and event memories 64; information transmission 242–5; popularity of FBMs in 1; reporting 78; shaping of collective identity 214–15 media events and FBM events, distinction between 265–6 media exposure: Alzheimer’s disease patients 151; Baudouin, King 230; collective memories 211–12, 213, 230–1; cultural issues 242–5; event accuracy 193; factual recall 114; FBM 21; flashbulb events 106; increased omissions 86; to information 25, 248; and measurement of FBM 25; models of FBM 63; public mourning 248; reconstruction of memory 59; and rehearsal 62–3, 66, 67, 86, 145, 193, 236, 250; and social factors 209–10, 210, 213 medial temporal lobe amnesia 147 Meehl, P.E. 20 Meksin, R. 212–13 memorability 32–3 memorialization 219, 221, 269 memories: episodic 264; remember/know memory distinction 34–5; traumatic 57, 79, 162, 171, 172, 176, 267 (see also posttraumatic stress disorder (PTSD)) Memories for Obama’s Presidential Election: delay, effects of 110, 113, 114; political views, effects of 114, 115; race, effects of 112, 113, 114, 115; source by information type interaction 111, 114; The Source by Information type interaction 114 memory: biases (see biases); consolidation model 98, 143, 231; developmental perspective (see children); distortion

280 Index

(see distortion, memory); elaboration (see elaboration, memory); for flashbulb events (see event memory; flashbulb events; flashbulb memories); as mutable 100; narratives 78; probes (see probed recall technique); reconstructive errors 125; selective 98–9; and self-identity 252–3; semantic (see autobiographical memory; semantic memory); shared (see social sharing; verbal communication); source 114, 147; spatial biases 125; see also emotional events; emotional memory; vividness Memory observed (Neisser) 265 Merckelbach, H. 146 meta-analyses 197, 226 methodology: associative hypotheses 38; case study approach 40–1; causal inferences 39–40; clarity and emotional affect 38–40; clarity ratings 35, 36; degeneration, avoidance of 44; experimental approach 43–4; four Ws 35–6; hypotheses/theory 37–43, 42; latent class/trait analysis 36–7; local inference 38; population of events 33–4; questioning 31; remember/know memory distinction 34–5; sampling 38–9, 97–8; types of FBM hypotheses 38–40; see also measurement of FBM Metternich, B. 148, 156 micro-blogging sites 243–4 mild cognitive impairment (MCI) 138–46, 150–1, 152–3 Mirani, J. 246 Mitterrand, François, President: collective memories comparison 205; consistency of memory 77; cross-cultural studies 192; death 20, 55–6, 173, 192, 232, 246; electoral victory 224 mobilization-minimization hypothesis 229 modelling see structural equation modelling (SEM) models of FBM formation: accuracy of FBM 63–4, 66, 67, 68; comparisons 55, 59–62, 67–8; consistency of FBM 48–9; constructivist view 22–3, 25; Conway et al. model 23, 50–2, 52, 100; Day and Ross model (2014) 64–6, 65, 86; dimensional 2–3, 14–19, 16, 17, 24; Er (2003) model 56–9, 58; Finkenauer et al. model 53, 53–6, 68; fit indices comparisons 57; future studies 68–9; prediction, direct/indiect paths 54; structural equation modelling (SEM) 49; Tinti et al. model (2014) 62, 62–4; twopath model 53, 53–6; types of events 66;

vividness of FBM 57; see also Brown & Kulik model; categorical models Modesti, A. 108 momentous public events see specific examples Monarch, E.S. 139 moon landing 162, 224, 227, 234 Moore, K.M. 81, 166 Mori, E. 140 Morse, C.K. 190, 191 motivation to remember 185, 195, 213–4, 253 mourning, public 248 Moustaki, I. 35, 36 MPLUS software package 18 Mullen, M.K. 250 Mulligan, N.W. 85 MULTILOG software package 17 Nachson, I. 21, 81 narcissism 20 narratives: discovery 109; length of 78; life 253; rules for 88; skills, children’s 102, 108 national identity see cross-cultural studies; cultural issues; psychosocial processes naturalistic vs experimental approaches 120–6, 122, 123 Neal, A.G. 230, 233 negative emotion: research focus on 162, 170–2; see also positive/negative emotion negative events 81, 84; biopsychosocial functions of memories 172–6; characteristics of FBMs compared to positive 176; collective memories 229–30; confidence in memory 163; consistency 169; detail 169–70; distinctiveness 171–2; evolutionary function of 174–5; future studies 176–7, 268–9; positive events comparison 162–3, 169–70; survival function 174–5; traumatic events 171; vividness of memory 163, 169 Neisser, U. 13, 58, 63, 76, 88, 97, 99, 100, 162, 185, 190, 191, 201, 264–5; Pearl Harbor memory 76, 99 network models 39 news coverage see media; media exposure Niedźwieńska, A. 83, 121–2, 123 9/11 Memory Consortium questionnaires 141, 149 9/11 terrorist attacks see World Trade Center attacks novelty and surprise 186; collective memories 226, 229; expected events 55–6; formation of FBMs 56, 59, 60;

Index  281

model comparisons 51; structural equation modelling (SEM) 51; two-path model 53–4; World Trade Center attacks 60 “Now Print” mechanism 184–5 nuclear accident, Japan 249 Obama, Barack, President: election of 110–15, 112, 113, 143–5, 145, 161–2; inauguration of 67 observational learning 102 O’Muircheartaigh, C.A. 36 ordinary (autobiographical) memories see autobiographical memory; exceptionality of events ordinary/special encoding hypothesis 74 Otani, H. 83, 190 Páez, D. 173 Pajkos, S. 104 Palme, Olof, (Swedish Prime Minister), assassination of 222 PANMARK software package 18 Paradis, C. 83 Park, C.R. 103 PARSCALE software package 17 patterns of forgetting see forgetting Pearl Harbor attack 76, 99 peripheral details/information 76–7, 125–6 personal events 109, 176, 178n3, 183, 226, 244, 253 personal identity 187; see also social identity personal importance 51, 52, 53, 58–9; Baudouin, King, death of 231, 234; cross-cultural studies 83; cultural variations 245–7; and emotionality 52; and social identity 82–3; World Trade Center attacks 83; see also significance personal memory, term definition 137 personal relevance: vs irrelevance 50, see also personal importance personal significance see personal importance; significance personal trauma see post-traumatic stress disorder (PTSD); trauma personality assessment 19, 20 Petersen, R.C. 138 Peynircioγlu, Z.F. 231–2 Pezdek, K. 20–1, 80, 190, 191, 208, 212 phenetic classes 19 phenomenological/phenomenal characteristics 79; cultural variations 246; future studies 89; revival of research into FBMs 2; see also confidence in memory; vividness of FBM Pillemer, D.B. 77

pleasant events 107–8 pleasantness bias 81, 171–2 Pokorney, J.J. 104 Poletta, F. 213 political: behaviours 254; beliefs 254; changes 223–4; crises 222, 223; events 222; rituals 231; violence 222, 223, 224 Ponsoda,V. 164 population of events 33–4 positive events 50, 63, 64, 81, 84, 107–8, 268–9; biopsychosocial functions of memories 172–6; characteristics of FBMs compared to negative 176; collective memories 229–30; confidence in memory 163, 169; consistency 169, 169; detail 169, 169–70; future studies 176–7, 268–9; negative events comparison 162–3, 169–70; sports events 222; studies examining 164–8; vividness of memory 163, 169, 169 positive/negative emotion: accuracy of FBM 249; Berlin wall, fall of 162, 165, 193; collective memories, requirement for 203, 222; Denmark, invasion/ liberation, WWII 84, 196; emotional memory 139; predisposition to negative 229; ritual 170; see also emotional valance; surprise post-encoding factors see rehearsal; social sharing post-event information 122, 241 post-traumatic stress disorder (PTSD) 57, 148–50, 155, 157 Poulin, M. 171 prediction models 49, 60–1; Conway et al. model 50–2, 52, 57; Day and Ross model (2014) 64–6, 65, 86; Er (2003) model 56–9, 58; Tinti et al. model (2014) 62, 62–4 prior interest 51 prior knowledge 51, 59, 186; see also novelty and surprise probed recall technique 113, 203, 204, 208, 209 proximity, effects of 105–6, 191 psychological accuracy 265 psychosocial processes: collective memories 232–6, 234, 235; rituals 230–2; types of event causing FBM 223; see also crosscultural studies; cultural issues PTSD see post-traumatic stress disorder (PTSD) public mourning 248 public trauma see trauma Putnam, F.W. 20

282 Index

Qin, J. 149–50, 155 quantity of recall and consistency/accuracy 77, 126 questionnaires: 9/11 Memory Consortium 141, 149; improving 177; standardized 68 R package e1071 software package 18 R package Itm software package 17 Rabin, Itzhak, assassination of 21 Rapcsak, S.Z. 147 rape 157 Rasmussen, A.S. 173 Reagan, President 222 reappraisal 220; see also appraisals reception: contexts 12–13, 20, 48, 186; events 204, 207, 213, 263–4, 266 recollective narrative technique not found reconstructive memory theory 121 recounting 107, 108, 111, 112, 113, 113–14, 115, 250; see also rehearsal Reed, J. 107 rehearsal: between-person 110–11, 114, 243; bypassing subjective emotional reactions 188; collective/factual 62–4; and confidence 85; consistency of FBMs 66, 85–6; cross-cultural studies 249–51; cultural issues 249–51; emotionality 51, 53, 54–5; event memory 60–1; forgetting 48–9, 57, 99, 204; future studies 88; individual/personal 62–4; and initial arousal 108; maintenance of FBM 50; media exposure 86; model comparisons 62–4; motivation for 185; ordinary/ special encoding hypothesis 85–6; and positive/negative emotion 81; social 230; social media 244; social sharing 230 (see also verbal communication); structural equation modelling (SEM) 50, 51, 52, 53, 54–5; two-path model 53, 54–5; and vividness 85; World Trade Center attacks 60–1 remember/know memory distinction 34–5; see also memory reminiscence bump 75, 225–6, 227–8, 254–6, 255 research: academic schism 264–5; controls in 267; early 1–2; future studies 68–9, 88, 89, 150–1, 157, 176–7, 197, 243, 256–7, 268, 269; ‘golden age’ 2; quality, low 264 revolutions 224 Rice, J. 104 Ride, Sally 103 rituals: collective behaviours 231–2; commemoration 231–2; and FBM/ communal memories 230–2; positive/

negative emotion 170; social 207–8, 209, 219 romantic relations: divorce 108; failed romances 108; FBM-like recall of events in 107–8; infidelity, discovery of 109; rejections 104–5 Ross, P. 102 Ross, S.A. 102 Rubin, D.C. 75–6, 78, 85, 97, 107, 184 Ruiz-Vargas, J.M. 164 rumination 60, 62, 67, 226, 248, 250 Ryff, C.D. 173 Safer, M.A. 166 sampling: methodology 38, 97–8; size of samples 267 San Franciso earthquake 222 Sartori, G. 21, 105 SAS macros 18 Satori, G. 192 Sauer, K. 103, 104, 107–8, 108 Schacter, D.L. 165 Schlagman, S. 246 Schlenger, W.E. 171 Schmidt, S. 50, 168, 192 Schmidt, S.R. 89 Schulze-Bonhage, A. 148 Schuman, H. 254 Scott, D. 164 Seiger, A.N. 143–5 selective memory 98–9 self: -categorization theory 187; -concepts 187; -construal 241, 241, 252–3; -defining functions of FBM 172, 173, 251–3; -esteem 173, 232–3, 252, 269; -identity see social identity; -referential effect 35, 83; -report ratings 3, 126; vs the other 101, 102–3 SEM approach (structural equation modelling) see structural equation modelling (SEM) semantic knowledge 154, 231; event memories 21; vs episodic memories 20–2; see also autobiographical memory Semin, G. 43, 123, 124–5 September 11 attacks see World Trade Centre attacks sexual abuse 191 shared emotions see social sharing shared knowledge see social sharing shared memories see social sharing shared values see social sharing Shoemaker, P.J. 242 significance: historical 4; personal 82–3 (see also personal importance); threshold of 51

Index  283

Silver, R.C. 171 Simpson, O.J., trial of 266 single-case study approach 40–1 sins of memory 98–100 social: artifacts and practices 207–8, 209; cohesion 88, 221, 230; exclusion 103–4; factors see collective memory, cultural issues, social identity; functions of FBMs 172–4; integration 230; media 243–4; membership, effects of 103–4; rehearsal 230; relevance 188, 265; representations 219–20; rituals 207–8, 209; see also rituals social identity: Berlin Wall, fall of 193–4; as causal factor 187–8, 188; and community 205; cultural issues 269; Denmark, invasion/liberation, WWII 193–6; formative years 225–6, 227–8; future studies 89, 197; gender differences 191; group differences in formation of FBMs 173–4; long-term of maintenance of FBMs 193–6, 197; personal importance 82–3, 183; proximity to events 191; reconceptualization 186–8, 188; relevant events 232; research conflicts 183–6, 185; research consistences 189, 190, 191–3; social relevance 188; surprise and FBMs 61; theory of 187; vignettes 193–6; World Trade Center attacks 60–1 social sharing 173; Baudouin, King 230; collective memories 230–1; cultural issues 249–51; emotions 231; and memory 88; social media 244; verbal communication 257 socialization, emotional 248–9, 250–1 socio-centrism 224, 225 socio-economic periods 224 Soleti, E. 56 Solomon, P.R. 143–5 Sotgiu, I. 192 source memory 114, 147 space shuttle disaster 97; see also Challenger disaster spatial biases 125 special (flashbulb/memory) mechanism: challenges to 73–4 (see also exceptionality of events); class vs. trait question 37; evidence summary 75; FBM models 32; initial/revised hypothesis 73–4, 120; and memorability 32 sports events 222, 224 stability of memory 126, 127, 184, 188 Stata plug-in 18 statistical modelling 2, 49; see also structural equation modelling (SEM) Stehman, A. 107

Stone, C.B. 66, 80, 83, 167, 222, 230, 249, 251, 253 stress 171; see also post-traumatic stress disorder (PTSD) stroke 147 Stroop paradigm 124 structural equation modelling (SEM) 2, 86, 265; Conway et al. model 50–2, 52; Day and Ross model (2014) 64–6, 65; Er (2003) model 56–9, 58; Finkenauer et al. model 13, 53, 53–6; limitation of 49; Tinti et al. model (2014) 62, 62–4; use of 49 substance addiction 20 surprise 51, 52, 53, 54, 59, 84, 186; social identity 61; survival function 174–5; see also novelty and surprise survival value, FBM 32, 32, 33, 174–5 symbolic meaning 188, 195–6 symbolic values 89, 195 Symons,V.L. 99, 208, 210 Takahashi, M. 66, 249 Talarico, J.M. 75–6, 81, 97, 166, 167, 184 taxonicity 19 taxonomic models/taxometrics 19–20, 24–5 taxonomy of consequentiality 101; Barack Obama, election of 110–15; event type: physical vs social 101, 103–7; private vs public 101–2; self vs the other 101, 102–3; valance: pleasant vs social 101, 107–1–; victim: self vs the other 101, 102–3 Taylor, S. 229 technological advances 223, 224; social media 243–4 Tekcan, A.I. 76–7, 77–8, 120, 164, 231–2 temporal lobe epilepsy 147–8, 151, 156, 157 test information function (TIF) 17 test-retest consistency 41; Challenger disaster 97; criterion 50, 127; and social identity 50; Thatcher’s resignation 50 Testa, S. 50, 168, 192 Thatcher, Margaret, resignation of 23–4, 50, 104, 105, 173, 192, 232, 246, 266 The seven sins of memory (Schacter) 98–100 thiamine deficiency 146 Thomas, Clarence, hearings 190, 191 Thompson, L. 104 Thompson, R.G. 140 Thomsen, D.K. 75, 76, 165, 192, 193–6 threshold model of memory retrieval 36–7 TIF see test information function (TIF)

284 Index

Tinti, C. 50, 83, 168, 190, 192, 205 Tinti et al. model (2014) 62, 62–4 to-be-remembered (TBR) information 206 transition period 254 trauma: clinical populations 137, 139, 146, 147, 151, 152, 157; negative events 171; personal 57, 79; public 214; see also posttraumatic stress disorder (PTSD); World Trade Center attacks tunnel memory 186 Turkey: earthquake 56–7; Prime Minister, death of 226, 232 two-path model 53, 53–6 US Airways Flight 1549 emergency landing 67 Vaclavik, D. 107–8, 108 valence: Alzheimer’s disease dementia 138–9; collective memories 229–30; durability of memories 229–30; of emotion 249; of FBMs 173–4; pleasant vs social 101, 107–10 validity, statistical 11, 59, 128, 131 values, and cultural issues 187, 193, 195, 202, 232–6, 247, 256 van der Haegen, A. 222 verbal communication of shared experiences 257 Verfaellie, M. 147 Vietnam War 227 vividness of FBM 23, 35, 48; closeness to event 80; and emotional intensity 85; and FBM as distinct category 79; measurement of FBM 77–8; models of FBM formation 57; ordinary/special encoding hypothesis 77–8; personal importance 83; and positive/negative emotion 81; positive/ negative events 163, 169, 169; ratings of 268; and rehearsal 85 von Restorff effect 81 Wagner, K. 148 Waller, N.G. 20 Wang, Q. 242, 243–4, 247, 250, 251, 255, 255–6 Waring, J.D. 143–5, 153 wars 224 Wasserman, R. 97 Weaver III, C.A. 97, 127 weighted attribute scores (WAS) 13

well-being 173, 230, 269 Wester, A. 146 Westerncentrism 224, 225 Wible, C.G. 97 Williamson, K.D. 103–4 Winograd, E. 13 Woodson, J.C. 103 working self 22 World Cup football 2006, Italian win 62 World Trade Center attacks 173; age, and memories of 226; Alzheimer’s disease patients 140–3, 142, 143, 144; amygdala activity 267; clinical populations 149, 152–5; conditions for FBM formation 3; consistency 2, 99, 204; critical period 227, 228; cross-cultural studies 192, 206; event memories 20–1; as extraordinary event 66; forgetting patterns 204; George Bush, location of at time of attacks 211, 212; ‘golden age’ of FBM research 2; Korsakoff syndrome patients 146–7; measures of FBM 59–62; media, and critical facts concerning 211–12; media coverage of event 210, 210; media exposure 193; model comparisons 59–62, 68; novelty and surprise 60; personal importance 83; personal life impact 43; and post-traumatic stress disorder (PTSD) 171; as prototypical event 61, 204; proximity to 106; rehearsal 60–1; and social identity 60–1, 191, 214; sustainability of collective memories 208; US citizenship status 83; weighted attribute scores (WAS) 13 World War II see Denmark, invasion/ liberation, WWII worry 149–50; see also anxiety Wright, D.B. 13, 23–4, 36, 40, 41–2, 85, 190 writing paradigm see emotional expressivity wrong time slices 76 X, Malcolm 205 Xcalibre software package 17 Yarmey, D. 106, 190, 192 Yerkes Dodson law 103 Yi, S. 250 Ying, X. 166 Zelig, A. 21 Zur, Judith 41