Mapping "Race": Critical Approaches to Health Disparities Research 9780813561387

Citation preview

Mapping “Race”

Critical Issues in Health and Medicine Edited by Rima D. Apple, University of Wisconsin–­M adison, and Janet Golden, Rutgers University, Camden Growing criticism of the U.S. health care system is coming from consumers, politicians, the media, activists, and healthcare professionals. Critical Issues in Health and Medicine is a collection of books that explores these contemporary dilemmas from a variety of perspectives, among them political, legal, historical, sociological, and comparative, and with attention to crucial dimensions such as race, gender, ethnicity, sexuality, and culture.

For a list of titles in the series, see the last page of the book.

Mapping “Race” Critical Approaches to Health Disparities Research Edited by Laura E. Gómez and Nancy López

Rutgers University Press New Brunswick, New Jersey, and London

Library of Congress Cataloging-­i n-­P ublication Data

Mapping race : critical approaches to health disparities research / edited by Laura E. Gómez and Nancy López. p.  ;  cm. —­(Critical issues in health and medicine) Includes bibliographical references and index. ISBN 978–­0–­8135–­6137–­0 (hardcover : alk. paper) —­ISBN 978–­0–­8135–­6136–­3 (pbk. : alk. paper) —­ ISBN 978–­0–­8135–­6138–­7 (e-­book) I.  Gómez, Laura E., 1964–­II. López, Nancy, 1969–­III. Series: Critical issues in health and medicine. [DNLM: 1.  Continental Population Groups—­United States. 2.  Health Status Disparities—­United States. 3.  Health Services Accessibility—­United States.  WA 300 AA1] 362.1'0973—­dc23

2012042036

A British Cataloging-­in-­Publication record for this book is available from the British Library. This collection copyright © 2013 by Rutgers, The State University Individual chapters copyright © 2013 in the names of their authors All rights reserved No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, or by any information storage and retrieval system, without written permission from the publisher. Please contact Rutgers University Press, 106 Somerset Street, New Brunswick, NJ 08901. The only exception to this prohibition is “fair use” as defined by U.S. copyright law. Visit our website: http://rutgerspress.rutgers.edu Manufactured in the United States of America

Contents

List of Figures and Tables Foreword

vii ix

R. Burciaga Valdez

Preface

Chapter 1

xiii

Introduction: Taking the Social Construction of Race Seriously in Health Disparities Research 1 Laura E. Gómez

Part I

Charting the Problem

Chapter 2

The Politics of Framing Health Disparities: Markets and Justice 25

23

Jonathan Kahn Chapter 3

Looking at the World through “Race”-­Colored Glasses: The Fallacy of Ascertainment Bias in Biomedical Research and Practice 39 Joseph L. Graves Jr.

Chapter 4

Ethical Dilemmas in Statistical Practice: The Problem of Race in Biomedicine 53 Jay S. Kaufman

Chapter 5

A Holistic Alternative to Current Survey Research Approaches to Race 67 John A. Garcia

Part II

Navigating Diverse Empirical Settings

Chapter 6

Organizational Practice and Social Constraints: Problems of Racial Identity Data Collection in Cancer Care and Research 87

85

Simon J. Craddock Lee

v

vi Contents Chapter 7

Lessons from Political Science: Health Status and Improving How We Study Race 104 Gabriel R. Sanchez and Vickie D. Ybarra

Chapter 8

Advancing Asian American Mental Health Research by Enhancing Racial Identity Measures 117 Derek Kenji Iwamoto, Mai M. Kindaichi, and Matthew Miller

Part III

Surveying Solutions

Chapter 9

Representing the Multidimensionality of Race in Survey Research 133

131

Aliya Saperstein Chapter 10

How Racial-­Group Comparisons Create Misinformation in Depression Research: Using Racial Identity Theory to Conceptualize Health Disparities 146 Janet E. Helms and Ethan H. Mereish

Chapter 11

Jedi Public Health: Leveraging Contingencies of Social Identity to Grasp and Eliminate Racial Health Inequality 163 Arline T. Geronimus

Chapter 12

Contextualizing Lived Race-­Gender and the Racialized-­ Gendered Social Determinants of Health 179 Nancy López

Notes on Contributors 213 Index 217

Figures and Tables

Figures

3.1 Age-­Specific Mortality, Selected Years (1963–­2004), Black/White Americans 40 4.1 Mary Gets a Score of 100 on a Verbal Ability Test 62 6.1 Conceptual Schematic of Identity Data Collection in a Cancer Service Line 99 12.1 Multidimensional “Race” Data at the Individual, Micro-level 192 12.2 Multidimensional Ethnicity as Distinct from “Race” 198 12.3 Racialized and Gendered Social Determinants of Health: Multilevel “Race” Data at the Micro-­, Meso-­, and Macro-­levels 200 Tables

3.1 Life of Homo sapiens Condensed into a Calendar Year 46 5.1 Heuristic Chart of Holistic Race Measures 70 6.1 Initial Draft Policy Categories 92 7.1 Determinants of Health Status among Full/Latino Sample (Ordered Logistic Regression) 111 8.1 PRIAS-­15 and PRIAS-­12 Items Derived from Confirmatory Factor Analysis on the Original PRIAS 122 8.2 Multiple Regression Using the PRIAS, PRIAS-­15, and PRIAS-­12 Measures as Predictors and Psychological Well-­Being as the Outcome 123 10.1 Some Examples of Risk Factors Used as Proxies for Race in Depression Research 150 10.2 Racial Identity Schemas for People of Color and White People 156

vii

Foreword

In 2008, the National Institutes of Health hosted over five thousand scholars at its first Science of Eliminating Health Disparities Summit. The subsequent annual summits were designed to encourage integration of science, practice, and policy to build a healthier society. As the opening plenary keynote speaker for the first one, I was asked to address “Health Disparities and the Intersection of Science and Policy.” I deliberately anchored my talk in an explanation of health disparities as a reflection of social stratification and inequitable resource allocations along “racial” and “ethnic” lines. But what do we mean by “racial” and “ethnic” lines? The scholars in this volume challenge us to be more precise about how we define and operationalize this type of stratification. Why? Because individuals are born into our society that neither treats people nor distributes opportunity equally. We observe discrimination, poverty, and other forms of oppression play out at the community level affecting overall community environment and opportunities. Especially in resource-­poor communities, the stresses of daily life bear down on the minds and bodies of residents, inviting illness through environmental exposures and other types of stress that are expressed biologically in what we recognize as disease. While Congress has required documenting these “racial” and “ethnic” disparities annually by the Agency for Health Research and Quality, it has done little to remedy the situation. Over the last few years, private philanthropies, such as the Robert Wood Johnson Foundation and the California Endowment, have recognized the need to prevent disparities by addressing the inequities in our society as well as addressing health care disparities—­creating a more just society and thus a more healthy nation. Interdisciplinary research plays an increasingly important role in addressing highly complex social issues such as health disparities. It allows scholars to reach beyond the boundaries of their own disciplines and to adapt or develop new methods of analysis. Institutional policies that support integrating the social and biological sciences could represent the twenty-­first-­century breakthrough similar to what we experienced in the twentieth century when biological and chemical scientists developed the fields of biochemistry, molecular biology, and molecular genetics that led to the sequencing of the human genome. Supporting interdisciplinary health and social science investigations

ix

x

R. Burciaga Valdez

and the development of transdisciplinary fields requires our nation’s funding institutions and universities to be more flexible and open to new ways of seeing things. The Institute for the Study of “Race” and Social Justice within the RWJF Center for Health Policy at the University of New Mexico is one example of such efforts. Sequencing the human genome provided definitive proof that there is no genetic or biological evidence for the concept of “race.” Yet in our society we often rely on assumptions about “race” that erroneously assume that it is interchangeable with genes to explain the world and our understanding of it. Scientists are not immune from using the myth of race as biology in dangerous ways. When I arrived in New Mexico in 2008, I had the opportunity to participate in the annual Cancer Center retreat. Researchers shared the marvelous molecular discoveries brought about by genomic research advances, which offer prospects for innovations in treatment for a wide variety of cancers; however, I was taken aback when colleagues resorted to explaining statistical propensities for groups of people in New Mexico using wildly inaccurate racial stereotypes. These genetic reductionist interpretations of health disparities denied American Indian spiritual beliefs, challenged the history of Hispanic settlement in the region, and posed findings in ways that could be misinterpreted in policy implementation. Despite all the problems inherent in the erroneous assumptions associated with the concept of “race,” it is imperative that we continue to collect data that captures the lived experiences and practices that contribute to the racialized inequality that is reproduced at all levels of society. While some may argue for a color-­blind society and an end to the collection of “race” or “ethnic” data, this book deliberately understands “race” as a social construction that can be mapped vis-­à-­vis legal efforts to eliminate racism in all its forms—­individual, institutional, and systemic. For example, in the areas of banking and housing we have largely eliminated public policies restricting groups of people from living in certain areas; however, informal practices in the mortgage and real estate industries continue to produce de facto segregation. Over the last several decades, we have negligently accepted racial categories in our work without the typical scrutiny we use to define variables for our analyses. Instead of interrogating and capturing the social construction of race in our research, we have at times blindly treated race as if it were a static characteristic that allows for the comparison of the health of different groups in our society. Or we have misused race as a proxy for many unmeasured factors in our studies, for example, class or culture, thereby masking what may be amenable to clinical intervention or social policy intervention.

Foreword xi

It is important that a paradigm shift take place in health disparities research, and Mapping “Race” helps catalyze it. Researchers should depart from the premise that racial and ethnic identities are fluid in both governmental and social settings. In order to combat discrimination, in 1976 Congress created the term “Hispanic” to group together a population that differs enormously by history, nationality, social class, legal status, and generation in the United States; it has little or no meaning outside of the U.S. political context. As scholars, scientists, and policy makers, we must challenge ourselves to be more precise with our categories of analysis and interpretation. The subjective meaning of such labels and whether they are situationally asserted are open empirical questions that need to be investigated. R. Burciaga Valdez July 2012

Preface

The story of how this book came to exist is somewhat unusual, and thus we include it here as a way of charting the intellectual genesis of Mapping “Race.” We intentionally invoke the word “mapping” in the book’s title as a metaphor to capture the complexity of “race” and to emphasize the need for researchers to grapple with that complexity. In early 2010, we received funding from National Institutes of Health (NIH) National Center on Minority Health and Health Disparities (NCMHHD) to hold a workshop called “Mapping ‘Race’ and Inequality: Best Practices for Theorizing and Operationalizing ‘Race’ in Health Policy Research.”1 Under the auspices of the Institute for the Study of “Race” and Social Justice at the University of New Mexico, which we co-­founded in 2009 and then co-­directed (2009–­2011), and which López now directs, we convened a group of nineteen scholars for a meeting in Albuquerque held April 29–­30, 2011. The participants in that intensive dialogue included scholars from the health, social, and biological sciences. They were an almost even mix of scholars whose primary methodological orientation was quantitative analysis versus those who used mixed methods or purely qualitative methods. Beyond those parameters, we self-­consciously sought to include scholars who represented diverse race, gender, and ethnic backgrounds, as well as scholars at varied career stages. It is from that initial group of nineteen scholars that the contributors to Mapping “Race” emerged.2 Mapping “Race” has benefitted from the collective excitement and synergy we experienced leading up to the workshop, during our two-­day meeting, and in our subsequent conversations. One of the striking things was the degree to which scholars from different disciplines and different methodological traditions were vexed about the same phenomenon: the frequent refusal by scholars doing research on race-­based disparities to explicitly conceptualize (much less define) “race” and the accompanying problem of how to most effectively operationalize socially constructed “race.” We were stunned by the similar stories told by participating scholars in the biological sciences (including biology and genetics), the medical sciences (including epidemiology, medicine, and public health), and the social sciences (including anthropology, political science, psychology, and sociology) about how existing research in their fields typically engages race in superficial ways that largely ignores theoretical insights relating

xiii

xiv

The Editors

to conceptualizing race as socially constructed and, therefore, as historically contingent, dynamic, and multifaceted. And yet, while we agreed on the problem, it was not obvious what the solutions were. Over the course of many conversations, disagreements, and even heated intellectual exchanges, we pushed each other to identify potential solutions and to speak across the traditional boundaries that separated us, such as method, discipline and scientific domain. Mapping “Race” embodies our progress and signals our desire to take this important conversation to a broader audience. We hope that it encompasses the voices of the original workshop participants and the exciting synergy of that collective conversation, as well as the contributions of the authors whose work has evolved into these chapters. We are grateful to many people for helping us realize this publication. We are deeply indebted to the faculty, fellows, administrators, and staff at the Robert Wood Johnson Foundation (RWJF) Center for Health Policy at UNM. At the Center, we found a generous intellectual community that welcomed us as relative outsiders to health research when we began this journey and that encouraged our efforts to cross disciplinary, science, and methodological boundaries. We are especially grateful to Executive Director Robert Valdez for seeing early promise in our ideas and for supporting us at each step of the way, including in our working group and speakers’ series and symposium on topics related to this research.3 Many Center staffpersons played a role in helping us at various stages, and we are especially grateful to Lia Abeita-­Sanchez, Lila Chavez, Sheri Lessanese, Thu Luu, Anita Parmar, Gina Sandoval, Vanessa Tafoya, Maria Vahtel, and Denise Wallen. Antoinette Maestas was especially helpful, working closely with us at several different stages. We thank Cirila Estela Vasquez Guzman, an RWJF Center for Health Policy doctoral fellow in sociology and our research assistant at the Institute (2010–­2012), for her assistance on multiple facets of this project. Estela and three other RWJF doctoral fellows contributed immensely as observers at the NIH workshop: Sonia Bettez (sociology), Yahaira Pena-­Esparza (psychology), and Vickie Ybarra (political science). We would also like to thank chairs and staff at the School of Law and in the Department of Sociology at the University of New Mexico for their assistance at various stages of this project, including Beverly Burris, Dorothy Esquivel, David Fricke, Margaret Harrington, Cyndi Johnson, Dona Lewis, Melissa Lobato, and Richard Santos. We thank our deans at UNM and University of California, Los Angeles (UCLA) for supporting our research, including Brenda Claiborne, Rachel Moran, Mark Peceny, and Kevin Washburn. For research assistance during various stages of this project, we thank librarians and student assistants at the UNM School of Law and the UCLA School of Law. For help

Preface xv

getting this manuscript to the final stages of production, we thank Tal Grietzer and especially Rusty Klibaner at the UCLA School of Law. It has been a pleasure to work with the editors and staff at Rutgers University Press, and we thank acquisitions editor Peter Mickulas, David Takeuchi and other reviewers of our manuscript, and the entire staffs in the production and marketing departments. We are honored that Mapping “Race” is included in the Rutgers University Press series “Critical Issues in Health and Medicine,” edited by Rima D. Apple and Janet Golden. Last, but certainly not least, we thank our families for their love and patience (and for putting up with too many early mornings, late nights, and weekends when they expected us to be “free”): Alejandro Gómez, Antonio Gómez, Eloyda Gómez, Luna Romero, Sierra Romero, Emma Romero and Augustine Romero. Laura Gómez and Nancy López July 2012 Notes 1. Funding for the workshop (April 29–­30, 2011) that made this book possible (in part) was by 1R13MD006054–­01 from the National Center on Minority Health and Health Disparities (NCMHD), the Agency for Healthcare Research and Quality (AHRQ), and the Eunice Kennedy Schriver National Institute of Child Health and Human Development (NICHD). The views expressed in written conference materials or publications and by speakers and moderators do not necessarily reflect the official policies of the Department of Health and Human Services; nor does mention by trade names, commercial practices, or organizations imply endorsement by the U.S. government. 2. The participating scholars and their institutional affiliations were the following: John Garcia, University of Michigan; Arline Geronimus, University of Michigan; Laura Gómez, (then) University of New Mexico; Joseph Gone, University of Michigan; Joseph Graves Jr., North Carolina A&T State University and University of North Carolina, Greensboro; Clarence Gravlee, University of Florida; Janet Helms, Boston College; Kimberly Huyser, University of New Mexico; Derek Iwamoto, University of Maryland; Camara Jones, Centers for Disease Control; Jonathan Kahn, Hamline University School of Law; Jay Kaufman, McGill University; Sandra Lee, Stanford University; Simon J. Craddock Lee, University of Texas Southwestern Medical Center; Nancy López, University of New Mexico; Michael Montoya, University of California, Irvine; Maribel Rodriguez-­Torres, Fundación de Investigación and Universidad de Ponce; Gabriel Sanchez, University of New Mexico; Aliya Saperstein, (then) University of Oregon. 3. For more information, see Trans-­disciplinary Guidelines for Researching “Race,” Institute for the Study of Race and Social Justice, RWJF Center for Health Policy, University of New Mexico, http://healthpolicy.unm.edu/about/initiatives/isrsj, last accessed July 25, 2012.

Mapping “Race”

Chapter 1

Laura E. Gómez

Introduction Taking the Social Construction of Race Seriously in Health Disparities Research

Typing the word “disparities” into the search engine of the American Diabetes Association web site generates nearly 1,700 hits that relate to racial and/or ethnic gaps in diabetes rates and care (http://www.diabetes.org, accessed July 5, 2012). Similarly, an agency of the U.S. Health and Human Services Department proclaims that the risk of diabetes is “much greater for minority populations than the white population” (“Diabetes Disparities among Racial and Ethnic Minorities Fact Sheet” 2012). These warnings reflect multiple biomedical studies that have identified differential rates of diabetes among Whites (6.2 percent), American Indians (9 percent), Mexican Americans (10.6 percent), and African Americans (10.8 percent; Mokdad et al. 2000), as well as those that have reported that some non-White groups have diabetes-related complications at rates as much as 50 percent higher than Whites (Carter et al. 1996). The temptation is to attribute such disparities to genetic differences because people often assume that “racial” groups correspond to biological differences. Moreover, in a capitalist society in which much medical research is driven by pharmaceutical companies’ pursuit of individualized solutions to health problems (see Kahn, this volume), we often look for a genetic basis for health outcomes. Yet much data suggests that the notion of biological race is a poor proxy for other social dynamics. For example, epidemiologist Thomas LaVeist and colleagues have challenged the conventional wisdom that differential rates of diabetes reflect essential, biological differences (2009). They studied diabetes in a racially mixed Baltimore neighborhood that included large numbers of both African American and White residents who were of the same socioeconomic class and who had comparable access to healthcare. In contrast to the

1

2

Laura E. Gómez

government studies previously referenced, they found that African Americans and Whites in this neighborhood had quite similar rates of diabetes (LaVeist et al. 2009). “I don’t mean to suggest that genetics play no role in race differences in health,” LaVeist said, explaining the study’s conclusions, “but before we can conclude that health disparities are mainly a matter of genetics we need to first identify a gene, polymorphism or gene mutation that exists in one race group and not others. And when that gene is found we need to then demonstrate that that gene is also associated with diabetes. On the other hand, there is [already] overwhelming evidence that behavior, medical care and the environment are huge drivers of race differences in health” (“Racial Disparities . . .” 2009). In other words, looking for race-­based health disparities may at best jump-­start a productive scientific inquiry when it leads researchers, policy makers, and health care providers to ask further questions about why race seems to be important in the context of a specific disease or health problem. But, at its worst, looking for race-­based health disparities blinds us to seeing the full range of possible causes of health inequalities. The broader point that we collectively make in this book goes a step further: we must be skeptical of claims about race-­based health disparities precisely because “race” is the product of historically rooted ideas and political contestation (Gómez 2012). Anthropologist Michael Montoya puts it this way: “the ascertainment of ethnicity or race is a profoundly social enterprise anchored in contemporary history,” and racial categories, both historically and today, “correspond best to the imaginations of the scientists and not the presumably defining and stable features being measured” (Montoya 2007). Using the example of diabetes in his book Making the Mexican Diabetic, Montoya explores how the process of racializing diabetes—­that is, the process of scientists and health professionals learning to take for granted that diabetes has a distinct impact and perhaps even etiology in people of different races—­has occurred in laboratories, in government funding circles, in peer-­reviewed scientific publications, and in the practice of medicine (2011). What explains biomedical researchers’ categorization of humans into groups, and then the linkage of those groups to specific health problems such as diabetes, is the social process of making race—­of constituting race as socially, politically, and scientifically important. Thus, our research agenda must include actively studying this racialization process; says Montoya: “when we carefully examine the selection of a group to study, the labeling of that group, the representation of that group in scientific papers, we see a science of population labeling based squarely on sociocultural factors particular to each group, each region, and each historical period” (Montoya 2007).

Introduction 3

As the diabetes example shows, there is a thriving literature documenting what appear to be enduring race-­based health disparities in the United States. In their comprehensive review of the literature on health disparities, sociologist David Williams and his colleagues confirm that the long-­standing gap between health outcomes among Whites and other racial groups has persisted into the twenty-­first century (2010). In particular, a wide variety of data sources show a continuing gulf between Whites and Blacks; for example, as a general measure of health, consider life expectancy. Whites’ life expectancy is 78.3 years, compared to 73.1 years for Blacks, so that it would take another quarter-­century to close the current White/Black life expectancy gap (Williams et al. 2010, 70). Similarly, the data show widening contemporary disparities between Native Americans and Whites (Williams et al. 2010, 74). Although, for a variety of reasons, the data is more limited than for African Americans and Native Americans, other data show that Latinos and Asian Americans have inferior health outcomes relative to Whites in many categories (Williams et al. 2010, 71).1 To a large extent, the current boom in the study of race-­based health disparities is the result of the process of institutionalizing, at the federal level, research and data collection along gender and racial lines. Since 1990, Congress has mandated a number of policy changes that have impacted how biomedical researchers and social scientists study health disparities (C. Lee 2009). For example, in 1990 the National Institutes of Health (NIH) created the Office of Research on Minority Health (C. Lee 2009, 1185).2 In 1993, when Congress funded the NIH, it directed the agency to require all grant awardees to include women of all races and minority men in clinical research (C. Lee 2009, 1185). During the same era, the federal government also created specialized departments to explore racial disparities within the Department of Health and Human Services and the Centers for Disease Control (Abu El-­Haj 2007, 292–­93). Sociologist Stephen Epstein (2007) has documented the fascinating political context and institutional dynamics that produced these changes, but no one doubts that these norms have become entrenched in today’s biomedical research establishment. If the federal government mandates various types of data collection to document such gaps and that data show racial gaps in health outcomes, what is wrong with health disparities research on race? Nothing, so long as that research is scientifically rigorous and accurate. The purpose of Mapping “Race” is to improve how health disparities research is conducted by challenging some central premises. The book’s core argument is that biomedical researchers and social scientists have not sufficiently grappled with how the conceptualization of race as socially constructed implicates how we operationalize and analyze

4

Laura E. Gómez

race in health disparities research. Although they represent a variety of disciplines across the branches of the medical, biological, and social sciences and take varied methodological approaches, the authors in this book agree that serious negative consequences will result if researchers continue to follow the current course. Significantly, it is not our claim that research on race or race-­based health disparities is illegitimate or inappropriate. We align ourselves with those scholars who have called for more and better research on race and racism (American Sociological Association 2003), rather than with those who have argued that the race is too complicated and fraught to be effectively studied scientifically (American Anthropological Association 1998). Thinking Critically about Racial Disparities

In order to make the point that researchers must chart a different course, let us start with an example that seems to be a classic case of a race-­based health disparity. While White women in Chicago are slightly more likely than Black women in Chicago to get breast cancer, Black women are twice as likely to die from it. As legal scholar Dorothy Roberts has noted, what is most shocking about this fact is that the Black/White gap emerged after 1980, when public awareness of breast cancer was at an apex and when there were great advances in detection and treatment (2011, 123). What happened between 1980 and the turn of the twenty-­first century to widen the gap between White and Black women’s breast cancer outcomes? If we were to think in narrow terms about “race,” we might be tempted to explore genetic (or epigenetic) differences as a symbol of what many people believe to be the biological differences among racial groups. A different approach is to think about the social context that produces racial disparities, which in turn involves thinking about race and racial dynamics as socially constructed. In Chicago, that means understanding the difference in access to health care that White and non-­White women typically receive. When Roberts interviewed Dr. Steven Whitman, whose research first documented the breast cancer survival disparity, she did so in his office at Mount Sinai Hospital, which she described as “small and shabby” and without air conditioning. Roberts noted that the hospital is located in “an all-­black community called North Lawndale on Chicago’s West Side, a block from the border of South Lawndale, which is predominantly Mexican . . . The patient population here at Mount Sinai is about half black and half Mexican,” Whitman told Roberts (2011, 124). He went on to say that Mount Sinai has, at most, one day of cash on hand to operate the hospital, compared to Northwestern University Hospital (located in majority-­White, affluent Evanston, a northern

Introduction 5

suburb of Chicago), which he estimated has four hundred days of cash on hand (and daily spends $5 to $10 million; Roberts 2011, 125); in the health care sector, hospitals are rated by bond agencies according to how many days of cash they have on hand. Whitman interprets the data as showing that while Black women’s breast cancer treatment stayed stagnant in the last two decades of the twentieth century, White women’s breast cancer prognoses dramatically improved due to early detection via mammograms and advances in treatment including radiation therapy and new medication regimes (Roberts 2011, 125). Roberts found that Black neighborhoods in Chicago had few facilities with mammograms (so that women had to travel long distances to get them, decreasing the likelihood that they would do so); that public hospitals used older, inferior mammogram equipment (often lacking both digital technology and trained mammographers); that until the 2010 Patient Protection and Affordability Act (President Barack Obama’s healthcare initiative), even middle-­class women with insurance were deterred from getting mammograms because of high insurance copayments (the 2010 law mandated the elimination of payments for women over age fifty); and that racially segregated neighborhoods meant that White women more frequently had access to the best breast cancer treatment facilities, which Whitman said were usually located in “the fancy institutions—­they are all in white neighborhoods” (2011, 125–­27). How are we to understand the role of race in the Chicago breast cancer example? The typical approach is to view race as an individual characteristic. For example, researchers often speak of self-­identified race, wherein a research subject selects her or his racial category from a list of limited options (which almost always correspond to the Office of Management and Budget [OMB] racial categories, themselves derived from U.S. Census categories). The OMB categories include “five racial categories (White, Black, American Indian or Alaskan Native, Asian, and Native Hawaiian and other Pacific Islander) and one ethnic category (Hispanic),” which most biomedical researchers collapse into six categories of “self-­identified race” that respondents may choose from among (Williams et al. 2010, 70). When researchers view race as an individual characteristic, they often fall into the pattern of biologizing race in terms of ancestry or genetics. This in turn invites the tendency to view health disparities as a question of individual failure (for less healthy individuals) or individual success (for more healthy individuals) and to identify solutions to disparities that focus on the individual body, such as pharmaceutical solutions (see Kahn, this volume). But Roberts tells a very different story about the racial disparity in breast cancer survival in Chicago. She urges us to think about how racial discrimination

6

Laura E. Gómez

has structured all sorts of inequalities, including health inequities. In her narrative, the emphasis is on how Chicago’s existing pattern of residential segregation of African Americans produces tangible gaps at every stage of healthcare. In a similar way, Williams and colleagues offer a nuanced picture of how neighborhood segregation leads to poor health via six “pathways”: First, segregation limits socioeconomic mobility by limiting access to quality elementary and high school education, preparation for higher education and employment opportunities. Second, the conditions created by concentrated poverty and segregation make it more difficult for residents to adhere to good health practices. . . . Third, the concentration of poverty can lead to exposure to elevated levels of economic hardship and acute stressors at the individual, household and neighborhood level. Fourth, the weakened community and neighborhood infrastructure in segregated areas can also adversely affect interpersonal relationships and trust among neighbors. Fifth, the institutional neglect and disinvestment in poor, segregated communities contributes to increased exposure to environmental toxins, poor quality housing and criminal victimization. Finally, segregation adversely affects both access to care and the quality of care. (2010, 79)

An even broader context would situate current residential segregation as part of a larger history of government-­sponsored, legally enforced policies that created today’s urban spaces. In this broader narrative, today’s health disparities are inextricably intertwined with the past because that past has direct impacts today—­via laws that restricted Blacks, other non-­Whites (and sometimes Jews) from buying houses in certain neighborhoods (racially restrictive covenants); via federal housing policies and agencies that created White-­only suburbs, supported the redlining of Black neighborhoods by banks and insurance companies, and that destroyed vibrant minority neighborhoods with the placement of highways and public transportation axes; and via neighborhood-­ based public schools that created a system of “naturally” occurring segregated schools that themselves became a self-­fulfilling prophecy for marking neighborhoods as “good” (White) and “bad” (non-­White). The starting point for Dorothy Roberts’s analysis is the claim that “race is not a biological category that is politically charged [but rather it] is a political category that has been disguised as a biological one” (2011, 4). It is the cumulative, social meaning of race in particular times and places that has shaped and continues to shape both racial discrimination (and racism) and racial meaning. To put it another way, we should not consider racial disparities in

Introduction 7

any isolated sense but instead consider the social context of racial categories and racial discrimination. Thinking about race-­based health disparities, then, must also involve engaging the social meaning of race and racism—­the objective of this book.3 Mapping “Race” asserts that we must think critically about how ideas about race are used in making claims about health disparities. We focus on three common deficits in the health disparities literature: (1) the failure to adequately define and/or conceptualize “race;”4 (2) the frequent and uncritical use of race as a control variable; and (3) the analytical slippage that often results when scholars (who have often engaged in the first two problematic steps) overstate or misrepresent the effects of race. The combined effect of these three weaknesses in the current literature on race-­based health disparities is to lead researchers and policymakers to think of race as fixed and biologically rooted. And this, in turn, leads to two major flaws that have significant public health implications. First, as Pamela Sankar and colleagues have noted, there are real dangers in taking an overly simplistic view of race, including the tendency to privilege genetics rather than environmental factors and the tendency to “blame the victim” by attributing poor health outcomes to particular racial groups (2004, 2987–­88). Moreover, at the level of large-­scale public health interventions, misidentifying race as the culprit leads to misplaced government resources coupled with the failure to implement genuine solutions. Let me illustrate the problem by returning to the example of diabetes that opened the chapter. Today, our approach is largely one of education to encourage early medical intervention, testing, and treatment; in other words, we spend a great deal of money widely publicizing to African Americans, Latinos, and Native Americans that they are especially susceptible to diabetes. But if LaVeist and colleagues are correct that diabetes rates more likely reflect patterns of social class and neighborhood/environmental factors that often are linked to race and racial discrimination at the macro-­, meso-­, and micro levels, then we are not effectively addressing the problem. Instead, an overly simplistic focus on individual-­level race masks the actual causes of health disparities—­the racialized-­gendered social determinants of health (see López, this volume)—­ that policies should address. Our education efforts would be better spent targeting the populations most vulnerable to poor access to healthcare, rather than using racial group membership as a crude proxy. Moreover, this approach involves thinking about how racial status and racism interact with socioeconomic status and discrimination in a complex, mutually constitutive way (rather than thinking about social class as determining race, as has been historically prevalent in the literature; see Takeuchi and Gage 2003, 439).

8

Laura E. Gómez

Mapping “Race” asks students and scholars of health inequalities to make the following four commitments in their research. First, they should work hard to understand what it means to say that race is socially constructed as part of interrogating and explicating their own conception of race. Second, they should understand the substantial scientific costs of continuing to mismeasure race in their research; in other words, they must be accountable for the limitations of their findings, given their reliance on a cramped measurement of race rather than one that corresponds to a robust notion of socially constructed race. Third, no matter what toolbox of methods they deploy, researchers must become proactive in seeking ways to measure race that more accurately reflect their chosen conceptualization. Finally, scholars of health disparities should be cautious when they make claims about race-­based disparities. In other words, they should avoid making claims that attribute causation to race without sufficiently exhausting other causes—­particularly social and environmental causes and structural racism—­and they should acknowledge the ways in which race is inextricably intertwined with other causes of health inequalities. Flaws in the Current Approach

Much empirical research on race shares a common problem—­the tendency to conceptualize race narrowly as phenotype and to crudely measure race via subject self-­identification from a closed list of options. This is a weakness in health disparities research as well as in a variety of other subfields and disciplines.5 In essence, there are two problems. The first is that scholars use race without saying what they mean by the term, without articulating a particular conception of race, and without justifying why race matters in their analysis. The second is that, when they attempt to operationalize race in the context of their research, too many scholars fall into a default mode in which race is a “control variable” or, in qualitative studies, a background variable that has not been adequately conceptualized and specified (C. Lee 2009). For example, in their review of more than one thousand articles published in the American Sociological Review (ASR), Martin and Yeung found that, between 1937 and 1999, the number of ASR articles became increasingly quantitative (mostly relying on regression analyses) and that these studies increasingly used race as a control variable. The rise of regression methods increased the likelihood that scholars would take race into account, and yet scholars’ tendency to “simply add race as a control variable in a regression model” implicitly is a very narrow way of conceptualizing race. They conclude that this way of introducing race into the analysis “implies that, while race makes a difference, it is not a profound one, in that race does not affect the relationships

Introduction 9

between other variables” (Martin and Yeung 2003, 532). Anthropologists Clarence Gravlee and Elizabeth Sweet contend that similar problems characterize the health disparities literature, where many researchers reflexively (rather than self-­consciously) use race as a proxy “for some unspecified combination of environmental, behavioral, and genetic factors” (2008, 49). They identify two resulting problems as the tendency to obscure the actual causes of health inequities and the promotion (explicit or implicit) of the idea that racial differences are genetic and innate (Gravlee and Sweet 2008, 49). Studies that use race in this or similarly narrow ways have had several, perhaps unintended, consequences. For one thing, these studies have cumulatively contributed to the popularization of a simplistic measure of race as a dichotomous variable (Black/White or White/non-­White) usually based on either bureaucratic assignment (racial assignment by someone collecting government data, such as a county coroner) or subject self-­identification. This common approach to race has in turn contributed to the idea, accreted over time, that race is fixed and biologically rooted. In this respect, scholars’ frequent, but unremarked upon use of race (especially as an independent variable) mimics the social phenomenon of Americans being repeatedly asked to report their race in a variety of life contexts, as noted by sociologist Ann Morning (2011). Framing the question as a simple one (with which of these listed groups do you identify racially?) and asking it repeatedly have led Americans to see race “as a permanent and individual characteristic: something that is embedded within us and [that] does not change over time” (Morning 2011, 3–­4)—­in other words, to reify race, rather than to see race as a complex and dynamic set of social processes. The problem of failing to conceptualize and/or define race spans the methodological spectrum. Sociologist Edward Morris argues for “greater transparency in how race is measured in qualitative studies and increased reflection on this concept as it is socially situated” (2007, 411). It is not enough, he contends, to simply declare that race is a social construction; he urges scholars to go further by expressly acknowledging “how they choose to identify race as well as recognizing the limitations of this choice and being attentive to the enactment of race in a particular context” (Morris 2007, 422). In other words, scholars who include race as a facet of their studies (whether a major or minor facet) should deliberately conceptualize race, regardless of their chosen methods. While such problems certainly transcend methodological orientation, there appear to be particular limitations with the tendency, in quantitative research, to use race as a control variable or as a crude proxy for some other social fact or process. Political scientist Taeku Lee has characterized the fundamental problem succinctly: “Although we acknowledge that race, like ethnicity, is a

10

Laura E. Gómez

social construct marked by fluidity, multiplicity, and contingency, we continue to measure racial and ethnic identities as fixed, categorical variables” (T. Lee 2009, 113). He has been most critical of his own brand of quantitative political science—­“multivariate statistical models in which some political variable of interest is explained by including a dummy variable for a given racial/ethnic category” (T. Lee 2008, 462)—­due to the trio of methodological problems it presents. First, this logic assumes that self-­identified race influences the dependent variable (such as the health outcome), without any explanation or justification (T. Lee 2008, 462). Second, it makes the unwarranted assumption that self-­identified race does not covary with other independent variables in the model (T. Lee 2008, 462). Third, such research typically fails to consider how self-­identified race (or some other measure of race) could be unreliable because of such factors as how the subject’s racial self-­identification might vary depending on the race of the person asking the question (same or other race of subject), the place where the question was asked (home, work, or school), the language of the question, or myriad other circumstances (T. Lee 2008, 463). Some scholars make the mistake of assuming that what “race” means is obvious. Yet we should make the opposite assumption: what “race” means is highly contested in popular culture, politics, law, and, as a result, science. In her research on how laypeople and scientists conceptualize race, sociologist Ann Morning (2009) finds three popular, contemporary conceptions of race: a biological notion, a culture-­based notion, and the idea that race is socially constructed. Moreover, she finds that people do not hold one conception of race to the exclusion of others but move back and forth between these three conceptions in order to explain different situations they encounter where race is relevant. In a similar way, major institutions such as the federal and state courts move back and forth among several conceptions of race, even within a particular time period or legal area (see Gotanda 1991; Haney López 1996; Pascoe 2009). Without making a conscious effort to do so, scholars are no more able to put aside folk notions of race than are laypeople. As Morning contends, “Despite the special authority that scientists enjoy, their beliefs are by no means independent of the broader society in which they train and practice. If lay people are influenced by what ‘experts’ say about race, the reverse is true too: scientific notions of race are informed by the broader political and social currents of their times” (2011, 4; see also Almaguer and Jung 1999, 234). In the context of the contemporary assault on race-­conscious law and policy and the entrenchment of color-­blind ideology in law and politics, it becomes all the more important for scholars to make clear the conception of race that they employ. Research that fails to expressly define race implicitly endorses a notion of race as fixed

Introduction 11

and biologically rooted—­a position that does not fit comfortably with the conception of race as socially constructed. Benefits of Viewing Race as Socially Constructed

Thinking of race as socially constructed presents a stark alternative to this way of thinking about race (or, more accurately, the default mode of not thinking carefully about race). To say that race is socially constructed is to acknowledge that we use phenotype or other visible characteristics to sort people into social groups, that we impute qualities of good and bad to these groups, and that the resulting racial order structurally and ideological supports a system of racial stratification that is socially contingent and historically rooted (Omi and Winant 1994). Rather than conceiving of race as shorthand for some essential, biologically rooted human difference, the constructionist view of race foregrounds the use of racial categories to justify racial hierarchy and inequality in particular times and places. This should lead us to focus on the effects of racial categories and racial ideology, namely, racial stratification, racial discrimination, and racism. An important aspect of the constructionist account has to do with historically situating scientific racism (in the social, biological, and medical sciences) as having played a critical role by creating and promoting biological claims about race in order to justify Whites’ racial domination of people of color (Hartigan 2008, 185; Taekuchi and Gage 2003, 436). Anthropologist John Hartigan Jr. makes the point that from “the earliest developments of colonialism to the current, emergent operations of biocapital,” the effort to differentiate the world’s populations on the basis of usually racialized biological characteristics has been the basis for justifying the sorting of distinctly racial labor pools (2008, 185). According to him, “Current assertions that race is socially constructed do important work by keeping both this history and these contemporary misuses of race in view, while also challenging the evidentiary ground for making claims about linkages between race and genetics” (Hartigan 2008, 185). This is a point worth emphasizing: viewing race as socially constructed foregrounds the racist nature of past conceptions of race, including those embraced by scientists. From the sociological vantage point, the claim that race is socially constructed has salience at all three levels of analysis: micro (individual), meso (community and organizational), and macro (institutional and structural). At the micro level, a constructionist approach acknowledges that race is dynamic over the life course (as one ages and develops) and that race varies by situation (for example, one might describe one’s racial identity differently at work or at

12

Laura E. Gómez

school than one would in the context of one’s neighborhood or extended family). At the meso level, a constructionist approach treats race as the product of organizational policies and practices that institutionalize ideas about race and that form a racial hierarchy that, over time, become natural and taken for granted by members of the organization. At the macro level, a constructionist approach to race directs our attention to aspects of the social structure that help reproduce the racial order and that justify that racial order as both right and inevitable. Let me illustrate each with an example. Consider a study by sociologists Andrew Penner and Aliya Saperstein; they used a random sample of nearly 13,000 Americans to test whether racial perceptions were, in fact, fixed or fluid (Penner and Saperstein 2008; see also Saperstein and Penner 2010). Respondents were interviewed annually over two decades and were asked to self-­identify their race, and Penner and Saperstein used these responses to code the subjects as “White,” “Black,” or “Other” (2008, 19630). In addition to this measure of racial self-­identification in the dataset, interviewers were instructed to classify the subjects’ race at the end of the interview, using the same categories (“White,” “Black,” or “Other) (2008, 19629–­30). If race was fixed and rooted in objectively understood phenotype, we would expect little change in the interviewer-­ascribed race of the 13,000 respondents. Instead, 20 percent of the individuals experienced at least one change in how interviewers classified them (2008, 19628). Penner and Saperstein found that, even when controlling for a wide array of possible factors, three characteristics of the respondent stood out as statistically significant predictors of the interviewers’ switching a respondent’s racial category: incarceration, unemployment, and income below the poverty line. The study suggests, then, that knowing one of those facts (about incarceration history, unemployment status, or poverty status) changed whether the interviewer saw the study participant as more or less “White” or more or less “Black,” suggesting that racial status is far from fixed and uncontested in one-­on-­one interactions. Moreover, the magnitude of change in racial perception was significant, affecting two in ten study participants in this large, random sample. As an example of meso-­level dynamics, sociologist Nancy López’s ethnographic research on New York City schools illustrates how socially constructed race works—­via “formal and informal institutional practices within schools”—­to racialize students (2003, 41). For example, she found that hyper-­ segregated urban schools usually had “dumbed down curriculums” boring to students who responded by “engaging in willful laziness,” thereby perpetuating a cycle in which mostly White teachers thought of their students, all of whom were people of color in this school, as intellectually incapable (2003,

Introduction 13

42). Another way that schools harden racial and gender boundaries is under the guise of “discipline.” López spent extensive time at a Manhattan high school that was characterized by hyper-­policing despite the fact that security officials told her that most student-­to-­student disputes were fights about property and that there had not been a single gun incident in more than five years. For example, the school’s head of security explained that enforcement of the rule against wearing hats was motivated by his need to demand respect from students. The rule was enforced selectively, only against male students (all of whom were Black and Latino), and it became a site for hardening racial categories because the boys took as many opportunities as they could to resist the rule (2003, 76–­77). Overall, López concludes that the school’s discipline policies constitute a pipeline between over-­policed public urban high schools and the larger society’s criminalization of men of color via the prison industrial complex (2003, 76–­77). In this way, we see schools’ powerful work to reproduce racial hierarchy and racial ideology. One by-­product is that young men of color are primed to experience greater stress and racist microaggressions that will contribute to a lifetime of stress that will feed into various mental and physical health problems (for example, see the chapters by Geronimus and Helms and Mereish, this volume). The dynamic nature of racial categories and racial ideology is also observable at the macro, or societywide, level. For example, my analysis of how Mexican Americans were incorporated into the American racial hierarchy in the nineteenth-­century Southwest shows that time and place matter and that “race” is not the same across space, time, or groups (Gómez 2007). Largely because doing so served the interests of the national campaign to invade Mexico and take its northern territory in order to reach the Pacific (often justified by the rhetoric of Manifest Destiny), 115,000 Mexican Americans living in the former Mexico became naturalized American citizens overnight in 1848. This occurred despite the fact that, at the time, U.S. laws excluded all non-­Whites from becoming citizens; in this sense, Mexican Americans were recognized as legally White. At the same time, there were multiple ways in which Mexican Americans and Mexicans were treated as racially subordinate and non-­White between 1848, when the war with Mexico came to an end, and 1912, when Congress made New Mexico a state despite its majority-­Mexican, Spanish-­ speaking population (Gómez 2007). Mexican Americans came to play a role as what I term an off-­White wedge group on the national scene, between African Americans below them (who were both socially and legally marked as racially inferior) and Whites above them (who were both socially and legally marked as racially superior), as well as in the more specific racial order in New Mexico, in

14

Laura E. Gómez

which Mexican Americans were an intermediate group between Native Americans below them and Whites above them. To say that race is socially constructed means to acknowledge that racial status is dynamic and situational: rather than being fixed at birth, life has any number of feedback loops that can change one’s race at the level of individual interactions, at the level of how organizations and communities operate, and at the level of societywide structure and ideology. The view that race is socially constructed has become the dominant approach in the social sciences (American Anthropological Association 1998; American Sociological Association 2003; Hartigan 2008; C. Lee 2009), and the idea has gained traction in many other fields and even in popular discourse (Morning 2009, 1171).6 Indeed, the constructionist view of race has become so popular that it has been embraced by proponents of a color-­blind perspective in many domains, from popular culture to medicine and law. According to this color-­blind world view, the fact that race is socially constructed means that race is not “real,” and therefore it should never be the basis of government policies such as affirmative action. This position deeply misunderstands the basic premise of sociology and anthropology: the fact that how we collectively understand the world powerfully shapes how we interact in it and therefore reality as we know and experience it (Berger and Luckmann 1967). For example, proponents of the color-­blind view often emphasize the role of individual preferences and market behavior, and yet, these aspects of social life are themselves key aspects of the socially constructed social fabric: Ideas, norms, and rituals evolve at the group or societal level and help to constitute individual identities, needs, preferences, and behavior. Individual action cannot be understood apart from the social environment that gives meaning to that action. Both “preferences” and market behavior are governed by taken-­for-­granted notions of what is natural, right, and rational (Edelman 2004, 186). With respect to race, enduring notions about the biological basis for race support and interact with other racial ideas to create the taken-­for-­granted, natural world in which racial identity and racial categories persist, in which we routinely (and often without thinking) classify people whom we encounter into racial categories, and in which we make a host of decisions (conscious and unconscious) based on those categorizations. Or, as epidemiologist Jay Kaufman puts it, “Despite this widespread understanding that racial categories are a product of our cultural imagination, we are still no more able to dismiss them. Race is not in our heads because it is real, but rather it is real because it is in our heads” (1999, 101). The fact that biology—­or what we often use as proxies for biology, such as ancestry, phenotype, genes—­is seen as related to race and even, in many

Introduction 15

people’s contemporary understandings of race, what produces racial difference, is indeed an important part of the social meaning of race. Yet biology is no less socially constructed, as sociologist Troy Duster has noted, emphasizing that the social meanings of race and racial interactions themselves have “feedback loops” into the biochemical, neurophysiological, and cellular aspects of bodies that can just as readily be studied scientifically (Duster 2003; Duster 2004). In other words, when human beings define situations as real, they can and often do have real social and biological consequences, consequences that can be translated into social facts that we as researchers can attempt to study and understand. Thus, saying race is socially constructed is not the same as saying that biology is irrelevant to race. As Hartigan suggests, rather “than deploying ‘social construction’ to reassert a distinction between the ‘biological’ and the ‘social’ or to assail the return of ‘atavistic beliefs’ about race, the more important move is to establish the primacy of cultural dynamics at work in shaping not just the genetic evidence and its interpretation, but the very interests and desires related to race that inform how this controversy unfolds both within and outside of the lab” (2008, 186; see also Gravlee 2009). Anthropologist Michael Montoya (2011) does some of this work in the context of unpacking how scientists, Mexican-­origin DNA donors, and health professionals all came together in recent decades to construct diabetes as “a Mexican problem,” with the pharmacogentic implications that such a view entails. Montoya warns, What requires explanation are not the processes and practices that technoscientists reiterate in the making of the three or five “races of man.” Although necessary, this is an unsatisfactory level of resolution if we are to understand and transform those material and semiotic assemblages that perpetuate inequity. Rather, what demands explanations are why this occurs in spite of the scientific, ethical, social, and political consequences so carefully detailed by scholars and analysts from a spectrum of fields across the social and biological sciences and humanities. (Montoya 2011, 180) Implications for Health Disparities Research

Consider that, of nearly 136,000 articles published in biomedicine (human research studies only), a whopping 51,039 used the concept “Negroid,” while 37,044 used the concept “Caucasian,” and 20,656 used the term “Mongoloid” (Graves 2010, 43). As biologist Joseph Graves has pointed out, while we know today that, on the one hand, these terms do not correspond to any discrete

16

Laura E. Gómez

population groups (so they could not correspond to mutually exclusive categories that scientists could study), on the other hand, biomedical researchers (and laypersons) continue to deploy these terms ubiquitously and casually. Building on the just noted calls by Hartigan and Montoya, and in response to scientists’ continued use of these terms despite their lack of validity, our agenda should be to ask a different set of scientific questions: why do these terms persist (in science and in the popular imagination) as racial designations? What political work does continuing to talk in terms of “Negroids,” “Caucasians,” and “Mongoloids” do in the contemporary discourse about race and racial difference? As an intellectual inquiry, such an analysis would involve conceiving of race as socially constructed—­as a product of a particular place and time—­as well as considering how racism and racial subordination have been and are promoted by this kind of language and categorization that is anything but scientifically neutral. This way of thinking about race as principally about socially created (and politically sustained) hierarchy and subordination differs dramatically from an approach that emphasizes race as an essential individual characteristic. Consider Dr. Sally Satel, who has gained notoriety by describing herself as a doctor who engages in “racial profiling”: “In practicing medicine, I am not colorblind. I always take note of my patient’s race. So do most of my colleagues. We do it because certain diseases and treatment responses cluster by ethnicity. Recognizing these patterns can help us diagnose disease more efficiently and prescribe medications more effectively. When it comes to practicing medicine, stereotyping often works” (Satel 2002, 56). Satel’s comments in the New York Times deserve unpacking, for they have the potential to do serious damage. In an era when racial stereotyping is decried by almost everyone and when color-­ blindness has become the mantra of conservatives in politics, law, and popular culture (Bonilla-­Silva 2010), Satel makes the provocative claim that race matters.7 Yet Satel’s claim that race improves disease diagnosis and medication treatment is largely unsupported by the biomedical research. It is precisely the type of claim that Mapping “Race” seeks to counter: the evidence simply does not suggest that self-­identified race today maps onto disease and treatment models in meaningful ways. We must recognize that what Satel is really asking us to do is think about race-­based health disparities in a simplistic, scientifically crude way that has fundamental implications. This way of seeing “race” presupposes that race is an individual characteristic about a patient that tells a health care provider something meaningful, implying that race is an essential characteristic, a fixed characteristic, and a biologically rooted characteristic (conjuring notions, in Satel’s narrative, of genetic

Introduction 17

or ancestral significance). As biologists Joseph Graves and Michael Rose have noted, this view of race in the clinical context may be quite dangerous because, “in addition to fostering social inequality by underscoring racial classification, racial medicine might kill people by neglecting the substantial genetic variation within, and genetic overlap between, human populations” (2006, 492). Furthermore, and as many chapters in this book illustrate, it is no simple matter for a doctor to assign her patient to a racial category or even to ask a patient to self-­identify racially, making the act of racial categorization in the biomedical context challenging (see López, this volume). For example, research shows that physicians and other health care personnel often are embarrassed or otherwise unwilling to directly ask a patient to identify her or his race, so that a health care provider might well be guessing a patient’s race based on inappropriate criteria (see Lee, this volume). An alternative approach to health inequalities that departs from the constructionist viewpoint is founded on two different realities. First, “race” is complex and inherently hard for researchers to measure or operationalize. Second, health disparities researchers must be extremely cautious about attributing outcomes to self-­identified race because seeming racially inflected dynamics often reflect larger process of racism and other forms of social inequality. For example, best practices for doing research on health disparities should include avoiding the assumption that race is important in favor of the premise that race is hard to measure and might be significant (for other approaches to making race more complicated, see the chapters by Garcia, López, and Saperstein, this volume). Consider the strategy taken by anthropologist Clarence Gravlee, whose research team set out to explore the connections between hypertension and race and, specifically, the claim that Blacks are more likely than Whites to experience high blood pressure and associated health problems (see Gravlee and Dressler 2005; Gravlee et al. 2005). Once they picked their research site in southeastern Puerto Rico, they implemented a three-­stage, multi-­method research design: first, using ethnographic methods, they assessed social norms about color and race in Puerto Rico; second, using interviews and surveys, they measured those ideas in the subject sample; and, third, using reflectometry, they objectively measured subjects’ skin pigment. Gravlee and his colleagues conclude that: “both self-­rated and culturally ascribed color—­but not skin pigmentation—­were associated with blood pressure through an interaction with income and education” (Gravlee and Dressler 2005). In other words, middle-­and high-­income people perceived by others as Black were more likely than those with objectively darker skin tone (at least as measured by the objective reflectometry test) or than those who were perceived as Black but who were

18

Laura E. Gómez

low-­income, to have hypertension. This finding suggests many public health questions about how we ought to think about identifying those at risk for high blood pressure and treating the disease. The authors also suggest that more research needs to be done to explore the dynamics of race-­based microaggressions and structural racism that affect the mental and physical well-­being of people of color, including hypertension, in ways that we are only beginning to understand. But the caution that a constructionist account raises is, who is in what “racial” category? “Whether people from Syria or the Indian subcontinent are placed in the ‘white’ group, whether Haitains, West Africans, or Cubans are categorized as ‘black’—­these are questions for which there are no materially valid answers, only vagarious compromises of one kind or another” (Kaufman 1999, 102). Kaufman would revert to self-­identified race as “the gold standard”—­ “that people are who they say they are” (1999, 103), but this too has its limits, as I have noted. At the end of the day, one of the main reasons for thinking about race as socially constructed is so that we can recognize the ways in which race thinking operates to solidify stratification and racial subordination. In particular, in the health context, the focus on race as measured via static self-­identification obscures the ways in which racism operates to create inequities in health outcomes and access to health care. If we understand racial categories as socially constructed to serve political purposes, then we are better positioned to identify and combat racism. Epidemiologist and physician Camara Jones and colleagues counsel that “it is vitally important that we develop a detailed understanding of the characteristics and manifestations of racism,” including institutionalized racism, personally mediated racism, and internalized racism (1991; see also Williams et al. 2010). At least part of the solution lies in adjusting our research studies in order to make race more complex so that it looks more like race in the real world. Acknowledgments

I thank Aliya Saperstein for helpful comments on an earlier draft of this chapter. I am grateful to Nancy López for many conversations about race and health disparities (often over good food), as well as for her sharp editing skills, both of which made this chapter better. For research assistance, I thank the librarians and student assistants at the libraries at the UNM School of Law and UCLA School of Law. Thanks to Rusty Klibaner for help with manuscript production.

Introduction 19

Notes 1. A range of limitations plague federal data collection for non-­ Black minority group members, and especially for Latinos and Asian Americans. Williams and colleagues report that “national data on mortality are more accurate for blacks and whites than for Hispanics, Asians, and Native Americans. A major problem affecting the quality of mortality data is related to the undercount in the number of deaths because of misclassification of nontrivial proportions of Hispanics, Asians, and especially American Indians as white on death certificates” (2010, 71). Nationwide data for Latinos and Asian Americans was not systematically collected before 1980, when immigrants from Latin America and Asia became significant nationally after the lifting of pre-­1965 restrictions on immigration from these regions (see also Takeuchi and Gage 2003, 440 [noting that social scientists did not historically treat Latinos and Asians in a racial framework]). Moreover, since both groups combine many different national origin subgroups, the data is subject to more variation and interpretive debate (for two studies that explore intragroup differences with the Latino and Asian American categories, respectively, see Sanchez and Ybarra and Iwamato, Kindaichi, and Miller, this volume). There are a variety of structural reasons for ongoing misclassification, however, including the historical racial ambiguity of Mexican Americans (see Gómez 2007) and other non-­Black minority groups. In addition, data such as that on birth and death certificates is subject to substantial regional variation in the United States (Pascoe 2009), further adding to the possible misclassification by race. 2. As noted in the preface, the NIH Office of Research on Minority Health funded the 2011 workshop that was the catalyst for this book. 3. Note that this inequality or social justice view of the cause (and solution) to race-­ based health disparities is the minority view in both the scientific literature and the popular media. One study of almost four thousand articles on health disparities that appeared in forty major newspapers between 1996 and 2005 concluded that only rarely were race-­based health disparities framed by the media as a question of social injustice—­that theme appeared in less than 4 percent of the articles (Kim et al. 2010, S229). 4. Sociologist Catherine Lee makes the important point that the failure of researchers to define race is simply bad science because it is “antithetical to the tenets of scientific research, which, in its ideal form, demands that analytical variables be consistent and their categories mutually exclusive” (2009, 1183). 5. For example, these authors have identified this weakness in the following fields: Gómez 2012 (socio-­legal studies); Gravlee and Sweet 2008 (medical anthropology); Harrison 1999 (cultural anthropology); Helms 2007 (psychology); Jones et al. 1991 (epidemiology); C. Lee 2009 (biomedical research); T. Lee 2009 (political science); Martin and Yeung 2003 (sociology); Morris 2007 (sociology); Mukhopadhyay and Moses 1997 (cultural anthropology); Saperstein 2008 (demography). 6. At the same time, Morning contends that the social constructionist conception of race has by no means completely displaced essentialist views of race rooted in biology. In fact, she finds that, among scientists who teach at the university level, an essentialist view of race as rooted in biology remains alive and well (2011, 6, 38–­47, 221). 7. Of course, this claim has been made historically by those advocating for race-­ conscious legal remedies and government policies to alleviate racial discrimination.

20

Laura E. Gómez

References Abu El-­Haj, Nadia. 2007.“The Genetic Reinscription of Race.” Annual Review of Anthropology 36: 283–­300. Almaguer, Tomas, and Moon-­Kie Jung.1999. “The Enduring Ambiguities of Race in the United States.” In Sociology for the Twenty-­first Century: Continuities and Cutting Edges, edited by Janet L. Abu-­Kughod. Chicago: University of Chicago Press. American Anthropological Association. 1998. “Statement on ‘Race.’” Accessed December 29, 2009. http://www.aaanet.org/stmts/racepp.htm. American Sociological Association. 2003. Statement of the American Sociological Association on the Importance of Collecting Data and Doing Social Scientific Research on Race. Washington, DC: American Sociological Association. Berger, Peter L., and Thomas Luckmann. 1967. The Social Construction of Reality. Garden City, NY: Anchor Books. Bonilla-­ Silva, Eduardo. 2010. Racism without Racists: Color-­ Blind Racism and Racial Inequality in Contemporary America. 3d edition. Lanham, MD: Rowman and Littlefield. Carter, Janette S., Jacqueline A. Pugh, and Ana Monterrosa. 1996. “Non-­Insulin Dependent Diabetes Mellitus in Minorities in the United States.” Annals of Internal Medicine 125 (3): 221–­32. “Diabetes Disparities among Racial and Ethnic Minorities Fact Sheet.” 2012. Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services. Accessed July 4. http://www.ahrq/gov/research/diabdisp/htm. Duster, Troy. 2003. “Buried Alive: The Concept of Race in Science.” In Genetic Nature/ Culture: Anthropology and Science Beyond the Two-­Culture Divide, edited by Alan H. Goodman, Deborah Heath, and M. Susan Lindee. Berkeley: University of California Press. ———. 2004. “Feedback Loops in the Politics of Knowledge Production. In The Governance of Knowledge, edited by Nico Stehr, 139–­60. New Brunswick, NJ: Transaction Publishers. Edelman, Lauren. 2004. “Rivers of Law and Contested Terrain: A Law and Society Approach to Economic Rationality.” Law and Society Review 38: 181–­97. Epstein, Stephen. 2007. Inclusion: The Politics of Difference in Medical Research. Chicago: University of Chicago Press. Gómez, Laura E. 2007. Manifest Destinies: The Making of the Mexican American Race. New York: New York University Press. ———. 2012. “Looking for Race in All the Wrong Places: Presidential Address.” Law and Society Review 46: 221–­45. Gotanda, Neil. 1991. “A Critique of ‘Our Constitution Is Color-­Blind.’” Stanford Law Review 44: 1–­68. Graves, Joseph L., Jr. 2010. “Biological v. Social Definition of Race: Implications for Modern Biomedical Research.” Review of Black Political Economy 37: 43–­60. Graves, Joseph L., Jr., and Michael R. Rose. 2006. “Against Racial Medicine.” Patterns of Prejudice 40: 481–­93. Gravlee, Clarence. 2009. “How Race Becomes Biology: Embodiment of Social Inequality.” American Journal of Physical Anthropology 139: 47–­57. Gravlee, Clarence C., and William W. Dressler. 2005. “Skin Pigmentation, Self-­Perceived Color, and Arterial Blood Pressure in Puerto Rico.” American Journal of Human Biology 17: 195–­206. Gravlee, Clarence, and Elizabeth Sweet. 2008. “Race, Ethnicity, and Racism in Medical Anthropology, 1977–­2002.” Medical Anthropology Quarterly 22: 27–­51.

Introduction 21

Gravlee, Clarence C., William W. Dressler, and H. Russell Bernard. 2005. “Skin Color, Social Classification, and Blood Pressure in Southeastern Puerto Rico.” American Journal of Public Health 95 (12): 2191–­97. Haney López, Ian F. 1996. White by Law: The Legal Construction of Race. New York: New York University Press. Harrison, Faye V. 1995. “The Persistent Power of ‘Race’ in the Cultural and Political Economy of Racism.” Annual Review of Anthropology 24: 47–­74. ———. 1999. “Introduction: Expanding the Discourse on ‘Race.’” American Anthropologist 100: 609–­31. Hartigan, John, Jr. 2008. “Is Race Still Socially Constructed? The Recent Controversy over Race and Medical Genetics.” Science as Culture 17: 163–­93. Helms, Janet. 2007. “Some Better Practices for Measuring Racial and Ethnic Identity Constructs.” Journal of Counseling Psychology 54: 235–­46. Jones, Camara Phyllis, Thomas A. LaVeist, and Marsha Lillie-­Blanton. 1991. “‘Race’ in the Epidemiologic Literature: An Examination of the American Journal of Epidemiology, 1921–­1990.” American Journal of Epidemiology 134: 1079–­84. Kaufman, Jay S. 1999. “How Inconsistencies in Racial Classification Demystify the Race Construct in Public Health Statistics” (editorial). Epidemiology 10 (2): 101–­3. ———. 2008. “Epidemiologic Analysis of Racial/Ethnic Disparities: Some Fundamental Issues and a Cautionary Example.” Social Science and Medicine 66: 1659–­69. Kim, Annice E., Shiriki Kumanyika, Daniel Shive, UzyIgweatu, and Son-­Ho Kim. 2010. “Coverage and Framing of Racial and Ethnic Health Disparities in U.S. Newspapers, 1996–­2005.” American Journal of Public Health 100: S224–­31. LaVeist, Thomas, Roland J. Thorpe Jr., Jessica E. Galarraga, Kelly M. Bower, and Tiffany L. Gary-­Webb. 2009. “Environmental and Socio-­Economic Factors as Contributors to Racial Disparities in Diabetes Prevalence.” Journal of General Internal Medicine 24 (10): 1144–­48. Lee, Catherine. 2009. “‘Race’ and ‘Ethnicity’ in Biomedical Research: How Do Scientists Construct and Explain Difference in Health?” Social Science and Medicine 68: 1183–­90. Lee, Taeku. 2008. “Race, Immigration, and the Identity-­to-­Politics Link.” Annual Review of Political Science 11: 457–­78. ———. 2009. “Between Social Theory and Social Science Practice: Toward a New Approach to the Survey Measurement of ‘Race.’” In Measuring Identity: A Guide for Social Scientists, edited by Rawi Abdelal, Yoshiko M. Herrerra, Alastair Iain Johnston, and Rose McDermott, 113–­44. Cambridge: Cambridge University Press. López, Nancy. 2003. Hopeful Girls, Troubled Boys: Race and Gender Disparity in Urban Education. New York: Routledge. Martin, John Levi, and King-­To Yeung. 2003. “The Use of the Conceptual Category of Race in American Sociology, 1937–­99.” Sociological Forum 18: 521–­43. Mokdad, Ali H., Earl S. Ford, Barbara A. Bowman, David E. Nelson, Michael M. Engelgau, Frank Vinicor, and James S. Marks. 2000. “Diabetes Trends in the U.S.: 1990–­ 1998.” Diabetes Care 23 (9): 1278–­83. Montoya, Michael J. 2007. “Do Genes Explain Diabetes Health Disparities between Ethnic Groups?” Endocrine Today (online edition). http://www.healio.com/Endocrinology/ news/print/endocrine-today/ percent7B00295910-DDE1–4643-B3D7–3D3DC6D046C0 percent7D/Do-genes-explain-diabetes-health-disparities-between-ethnic-groups. ———. 2011. Making the Mexican Diabetic: Race, Science, and the Genetics of Inequality. Berkeley: University of California Press.

22

Laura E. Gómez

Morning, Ann. 2009. “Toward a Sociology of Racial Conceptualization for the Twenty-­ first Century.” Social Forces 87: 1167–­92. ———. 2011. The Nature of Race: How Scientists Think and Teach about Human Difference. Berkeley: University of California Press. Morris, Edward. 2007. “Researching Race: Identifying a Social Construction through Qualitative Methods and an Interactionist Perspective.” Symbolic Interaction 30: 409–­25. Mukhopadhyay, Carol C., and Yolanda T. Moses. 1997. “Reestablishing ‘Race’ in Anthropological Discourse.” American Anthropologist 99: 517–­33. Omi, Michael, and Howard Winant. 1994. Racial Formation in the United States: From the 1960s to the 1990s. New York: Routledge. Pascoe, Peggy. 2009. What Comes Naturally: Miscegenation Law and the Making of Race in America. Oxford: Oxford University Press. Penner, Andrew, and Aliya Saperstien. 2008. “How Social Status Shapes Race.” Proceedings of the National Academy of Sciences 105: 19628–­30. “Racial Disparities in Diabetes Prevalence Linked to Living Conditions.” 2009. Press release, Bloomberg School of Public Health, Johns Hopkins University, September 21. Roberts, Dorothy. 2011. Fatal Invention: How Science, Politics, and Big Business Re-­ create Race in the Twenty-­first Century. New York: New Press. Sankar, Pamela, Mildred K. Cho, Celeste M. Condit, Linda M. Hunt, Barbara Koenig, Patricia Marshall, Sandra Soo-­Jin Lee, and Paul Spicer. 2004. “Genetic Research and Health Disparities.” Journal of the American Medical Association 291 (24): 2985–­89. Saperstein, Aliya. 2008. “(Re)Modeling Race: Moving from Intrinsic Characteristic to Multidimensional Marker of Status.” In Racism in Post-­Race America: New Theories, New Directions, edited by Charles Gallagher. Chapel Hill, NC: Social Forces Publishing. Saperstein, Aliya, and Andrew M. Penner. 2010. “The Race of a Criminal Record: How Incarceration Colors Racial Perceptions.” Social Problems 57: 92–­113. Satel, Sally. 2002. “I Am a Racially Profiling Doctor.” New York Times Sunday Magazine, May 5. Takeuchi, David T., and Sue-­Je Lee Gage. 2003. “What to Do with Race? Changing Notions of Race in the Social Sciences.” Culture, Medicine, and Psychiatry 27: 435–­45. Williams, David R., Selina A. Mohammed, Jacinta Leavell, and Chiquita Collins. 2010. “Race, Socioeconomic Status, and Health: Complexities, Ongoing Challenges, and Research Opportunities.” Annals of the New York Academy of Sciences 1186: 69–­101.

Chapter 2

Jonathan Kahn

The Politics of Framing Health Disparities Markets and Justice

The increase of using race as a biological construct in biomedical research and practice raises concerns over the dangers of reifying race in a manner that leads to new forms of discrimination. The most prominent example of racialized medicine is BiDil, the first drug ever approved by the FDA with a race-­specific indication on its label—­for use in “Black” patients. BiDil, however, is part of a much larger dynamic in which the purported “reality of race” as genetic is used to obscure the social reality of racism (Kahn 2008; Kahn 2012). To the extent that this dynamic succeeds in reductively reconfiguring health (and other types of) disparities in terms of genetic difference, it casts personal responsibility and the market as the appropriate arenas for addressing differential outcomes. It also undermines the rationale for deliberate state or institutional interventions that could more effectively address discrimination (Kahn 2005a; Kahn 2012). This is not to advocate “color-­blind” medicine. To the contrary, there are very real health disparities in the United States that correlate with race. African Americans suffer a disproportionate burden of a number of diseases, including hypertension and diabetes. Like heart failure, these are complex conditions caused by an array of environmental, social, and economic as well as genetic factors. Central among these is the fact that African Americans experience discrimination, both in society at large and in the health care system specifically. The question, once you identify these disparities in health outcomes, is how to address the underlying causes. Of course, outcomes can have multiple causes, both social and genetic. But health disparities are not caused by an absence of “Black” drugs. As studies by the Institute of Medicine (IOM) among others make clear, they are caused by social discrimination and economic inequality (Smedley, Stith, and Nelson 2003). The

25

26

Jonathan Kahn

problem with marketing race-­specific drugs is that it becomes easier to ignore the social realities and instead focus on the molecules. Frames for “Disparities” and “Difference”

Contemporary debates over the meaning and significance of racial disparities in health have deep roots (Washington 2006; Williams and Jackson 2005). Tracing their development over the past century, particularly in relation to the legal construction of racial difference, elucidates a history of tension between two competing frames for characterizing social understandings and responses to disparities. I characterize these frames here as a distinct binary, but they really exist as a continuum with one side or another gaining different degrees of prominence in different eras or contexts. One the one side I place “markets,” on the other, “justice.” First, when identifying a race-­ based health disparity, one can locate its source or cause in human bodies or in social conditions. The former tends to biologize race, marking racialized bodies as somehow defective, weak, or diseased, often at the molecular level; examples would include characterizations of heart failure as a “different disease” in African Americans (Yancy 2002), framing the resulting health disparities as due more to biology than social or historical context. The latter tends to racialize social dynamics, marking society as somehow discriminatory or unjust; examples would include many of the findings of the IOM report Unequal Treatment: Confronting Racial and Ethnic Disparities in Healthcare that explicitly examined diverse social, historical, and economic factors as contributing to health disparities (Smedley, Stith, and Nelson 2003). Second, having identified the source of the problem, the next step is to frame a locus of responsibility for addressing it. If the source has been located in racialized human bodies, then the tendency is to situate responsibility for addressing the problem in the individuals whose bodies are affected. This often takes the form of calls for “personal responsibility” in taking care of oneself. Examples would include action in 2003 by the Department of Health and Human Services (DHHS) to emphasize personal responsibility when it reported to Congress on the IOM study (Kahn 2005a). DHHS also pushed to replace multiple references to “disparities” with the term “difference,” thereby also focusing more on a biological basis for disparities. If the source of the problem is located in society, then the tendency is to situate responsibility for addressing the problem in the polity. Examples would include characterizations of health equality as a civil right, going back at least to Dr. Martin Luther King Jr.’s 1966 declaration that “of all the forms of inequality, injustice in health care is the most shocking and inhumane” (King 1966).



The Politics of Framing 27

Third, having identified the source or the problem and located responsibility for it, the final step is to formulate an approach to solving it. Going down one track, if you have located the source of the problem in human bodies, perhaps at the genetic level, and situated responsibility in the individuals whose bodies are affected, then the tendency is to formulate privatized, market-­based approaches to address the problem. Examples would include BiDil, a drug formulated to address purported disparities in heart failure at the molecular level, to be purchased from a pharmaceutical corporation by individual consumers taking responsibility for their condition (Kahn 2005a). This approach was evident in a 2006 statement by Alan Levine, secretary of Florida’s Agency for Healthcare Administration, declaring that “in the case of BiDil, we have clearly identified a product targeted toward closing the racial disparity gap” (AHCA 2006). Going down the other track, if you have located the source of the problem in social conditions, and situated responsibility in the political community as a matter of justice, then the tendency is to formulate government-­based policy initiatives to address the problem. Examples would include any of an array of civil rights–­ related laws that seek to improve the social or economic conditions of minority populations thereby improving health (Epstein 2007; Omi and Winant 1994). Of course, many responses reflect different mixes of these frames of health disparities. Thus, for example, a federal program may encourage market-­based mechanisms to improve minority access to drugs in the marketplace. The disparities frame of “access” often emerges as an area around which Democrats and Republicans share a measure of agreement as it implicates both government initiatives to reduce disparities and personal responsibility for individualized health care choices in the market. The “access” approach involves the political community in formulating policy responses to a problem, but it also ultimately locates the problem of disparities in the individual bodies of the affected people. Calls for improved access to health care, then, may involve federal initiatives to help historically disadvantaged groups, but they are double-­edged insofar as they promote market solutions to health disparities that characterize health care as a consumer good rather than as a civil right. Moreover, diverse well-­intentioned federal initiatives have both directly and indirectly promoted the framing of health disparities in terms that locate the problem in the bodies of individual members of geneticized racial groups (Sankar 2004). In the context of using social identities such as race in public health research, Ellison and Jones have expressed the concern that recent “expansion in genetic technology will lead to a further focusing of the medical gaze onto individual risk factors and away from social, environmental and ecological factors” (Ellison and Jones 2002). The critical point, then, is to be sensitive to which frame is

28

Jonathan Kahn

being deployed at any given point in a debate over the source, responsibility, or response to health disparities. This chapter will proceed by exploring the historical background to framing issues of racial inequality and justice with particular attention to legal developments manifest in key Supreme Court decisions, from Plessy v. Ferguson in 1896, to Brown v. Board of Education in 1954, to the 2007 case of Parents Involved in Community Schools. I show how arguments about racial inequality over the course of more than a century have been critically shaped according to how they locate responsibility for such inequality at the individual or social level, with consequent implications for constructing (or ignoring) appropriate remedies to the situation. The chapter considers how this long-­standing dynamic is manifest today in the politics of health disparities. I argue that during the first decade of the twenty-­first century, concerted efforts were made to cast health care as a consumer good and health disparities as a function of biological difference best addressed through market-­oriented, rather than political, means. Markets and Justice: The Historical Background

At the turn of the twentieth century, Fredrick Hoffman published Race Traits and Tendencies of the American Negro for the American Economic Association (Hoffman 1896). Hoffman, a statistician at the Prudential Life Insurance Company, wrote the article for Prudential in response to a wave of state legislation banning discrimination against African Americans. Hoffman’s aim was to establish the biological inferiority of the Negro as a basis for justifying Prudential’s decision to exclude African Americans from access to insurance coverage. Prudential had begun cutting back on providing life insurance to African Americans as early as 1881 on the grounds that they suffered higher rates of mortality. Other insurance companies soon followed suit. State legislatures, particularly in the North, were wary of the growing power of the insurance industry and soon enacted new statutes to regulate the industry; some of these included antidiscrimination provisions. Prudential decided to resist the new laws by asserting a “natural” biological basis for their discrimination. Leslie Ward, the vice president of Prudential, declared “We are quite sure that mortality, even amongst the best of colored lives, would not compare favorably with the mortality amongst whites” (Wolff 2006, 89). Hoffman’s task was to provide rigorous scientific analysis to undergird such claims. In his 330-­page treatise, Hoffman compiled statistics, anecdotal observations, and eugenic theories in order to argue that “it is not in the conditions of life, but in the race traits and tendencies that we find the causes of excessive mortality” (Hoffman 1896, 95). Hoffman’s work is replete with



The Politics of Framing 29

charts, measurements, and statistical observations all purporting to establish the biological inferiority of the “American Negro” not only in basic mortality but in such physiological traits as “chest expansion,” “physical strength,” or resistance to disease. Such observations built on Hoffman’s prior conclusion that “the time will come, if it has not already come, when the negro, like the Indian, will be a vanishing race” (Hoffman 1892, 542). Of course, academics and expert commentators had been declaring the inferiority of the “negro race” for decades before Hoffman came on the scene (Braun 2002, 160–­63). What is distinctive about his work for Prudential is that it demonstrates the use of actuarial data to provide a gloss of objective scientific rigor to the construction of biological racial difference in a legal context in order to gain economic advantage in a competitive marketplace. This is the flip side of the BiDil story, where the drug’s developers used a purported race-­based biological difference to include African Americans in a patent (Kahn 2004). Here, Prudential was using biological difference to exclude African Americans from insurance coverage. Both strategies involve constructing race as biological in order to serve underlying commercial interests. Hoffman’s work did not go unchallenged. W.E.B. Du Bois, in particular, presented a powerful critique, questioning Hoffman’s methodology and noting that the health outcomes of African Americans were comparable to those of immigrant groups with similar economic resources. Acknowledging “that in certain diseases the Negroes have a much higher rate than the whites,” Du Bois asked, “the question is: Is this racial? Mr. Hoffman would lead us to say yes, and to infer that it means that Negroes are inherently inferior in physique to whites” (Du Bois 1906, 275). Du Bois asserted, however, that such differences “can be explained on other grounds than upon race” (Du Bois 1906, 275). Examining the data behind various differentials in morbidity and mortality, Du Bois concluded “that the Negro death rate and sickness are largely matters of [social] condition and not due to racial traits and tendencies” (Du Bois 1906, 276). Hoffman’s thesis was published in 1896, the same year that the U.S. Supreme Court issued its infamous “separate but equal” decision in the case of Plessy v. Ferguson (163 U.S. 537 [1896]). Homer Plessy, a resident of New Orleans, was seven-­eighths White and one-­eighth Black—­an “octoroon.” On June 7, 1892, he bought a first-­class ticket on the East Louisiana Railway for passage from New Orleans to Covington, Louisiana, and took a seat reserved for White passengers. He was arrested for violating the Separate Accommodations Act of 1890, which required “all railway companies carrying passengers on their trains, in this state, to provide equal, but separate, accommodations for the white and colored races” (11 So. 948 [La. 1892]).

30

Jonathan Kahn

The Supreme Court ultimately upheld the Louisiana law, rejecting Plessy’s argument “that the enforced separation of the two races stamps the colored race with a badge of inferiority,” concluding that “if this be so, it is not by reason of anything found in the act, but solely because the colored race chooses to put that construction upon it” (163 U.S. 537 [1896]). Writing for the majority, Justice Henry Brown declared “Legislation is powerless to eradicate racial instincts, or to abolish distinctions based upon physical differences, and the attempt to do so can only result in accentuating the difficulties of the present situation. . . . If one race be inferior to the other socially, the constitution of the United States cannot put them upon the same plane” (163 U.S. 537). Just as Du Bois offered a trenchant critique of Hoffman’s work that placed social forces front and center in analyzing mortality differentials, Justice John Harlan’s eloquent dissent in Plessy argued that the social consequences of state action must be accorded constitutional recognition. “The arbitrary separation of citizens, on the basis of race,” Harlan declared, “while they are on a public highway, is a badge of servitude wholly inconsistent with the civil freedom and the equality before the law established by the constitution. It cannot be justified upon any legal grounds” (163 U.S. 562). Legal scholar J. Allen Douglas argues that Plessy and its progeny can be understood as attempts “to locate racial identity in the body in the form of an object or property—­an immutable, natural ‘thing’ possessed—­to ensure a means for ‘quieting title’ in whiteness” (Douglas 2003, 889). In its attempt to locate racial identity in the “immutable, natural” object of Plessy’s body, Brown’s brief opinion echoed Hoffman’s massive compilation of statistics, which similarly grounded its defense of a “segregated” approach to life insurance underwriting by locating racial difference in the bodies of African Americans. Brown also located responsibility for stigma that might result from the law in the social realm, beyond the purview of the Court. As Brown declared, “Legislation is powerless to eradicate racial instincts, or to abolish distinctions based upon physical differences” (163 U.S. 551 [1896]). Louisiana could draw legal distinctions based upon such physical differences without violating the Equal Protection clause of the Fourteenth Amendment, but if African Americans felt degraded by the law, that was their problem. The state could act on biological difference but had no responsibility to redress harms related to social difference. The Supreme Court’s opinion in Plessy thus legitimated the racial segregation of physical space by denying the legal significance of social differences that might be engendered by the law. Hoffman similarly tried to justify the social (that is, non-­state-­sponsored) segregation of insurance coverage by asserting the legal significance of purported biological differences among races. Each of these foundational documents of twentieth-­ century race theory characterized the relative legal significance of social and biological difference



The Politics of Framing 31

in commercial contexts. In justifying racial subordination, they deny or obscure the significance of social or political forces while elevating biological difference as a justification for state-­sanctioned discrimination. This schema naturalizes racial difference and places any responsibility for addressing its implications beyond the purview of the state, more appropriately addressed by individual action or structured by commercial market considerations. For the past century, this characterization of racial difference has existed in tension (and often direct conflict) with the approach of those like Du Bois and Harlan who have sought to foreground the social basis of racial difference and assert a public, state responsibility for both creating and redressing inequalities embedded in such differences. The 1954 case of Brown v. Board of Education repudiated Plessy’s concept of “separate but equal” as “inherently unequal” (347 U.S. 483 [1954]). It did so by upending the logic that relegated the “social” beyond legal consideration, noting that “to separate [school children] from others of similar age and qualifications solely because of their race generates a feeling of inferiority as to their status in the community that may affect their hearts and minds in a way unlikely ever to be undone” (347 U.S. 494). The opinion also marked a transition in the social sciences, away from Hoffman’s biologization of racial difference toward Du Bois’s focus on the impact of social forces on the status and situation of racial groups. The Warren Court famously cited the work of social scientists such as Kenneth Clark and Gunnar Myrdal in supporting its legal recognition of the impact of segregation upon the hearts and minds of African American children (347 U.S. 494 fn. 11; Hovenkamp 1985, 671–­72). Where Hoffman and Justice Brown foregrounded race as a biological concept in order to locate the responsibility for inequality in Black bodies and minds, Clark, Myrdal, and the Warren Court foregrounded the social dynamics of race to locate responsibility for inequality in the state and the polity. From Brown to the passage of the Civil Rights Act in 1964 and the Voting Rights Act of 1965, the emergent civil rights revolution focused on securing government guarantees of fundamental citizenship rights for African Americans. Journalists Thomas and Mary Edsall, in their influential book Chain Reaction, have argued that the politics of this first era focused on equality of opportunity in a manner that galvanized a majority of the country and overcame conservative forces of racial resistance. After 1965, however, the civil rights agenda shifted its focus to “broader goals of emphasizing equal outcomes or results for Blacks” (Edsall and Edsall 1992, 7). Statistics were necessary for measuring such outcomes, and it is around this time that race-­specific data from the Census Bureau became essential for gauging social, economic, and political discrimination in this country (Robbin 2000b, 435; Robbin 2000a, 129). It is in

32

Jonathan Kahn

this context, speaking before the Second National Convention of the Medical Committee for Human Rights that Dr. Martin Luther King Jr. made his statement about injustice in health care being shocking and inhumane (King 1966). The civil rights shift in focus from access to fundamental rights of citizenship to outcomes secured through affirmative government action also gave rise to a conservative backlash, operationalized via Richard Nixon’s “southern strategy” that used race as a wedge issue to “divide voters over values, and to isolate one disproportionately poor segment of the population from the rest of the electorate” (Edsall and Edsall 1992, 5). The new Republican strategy, carried to its acme by Ronald Reagan in the 1980s, focused on the concept of “equal opportunity” as opposed to “affirmative action” and included among its supporters “radical conservatives, including those who were in fact anti-­black, as well as much of the moderate center and the ideological right” (Edsall and Edsall 1992, 143). As the Edsalls note, “The changing politics of civil rights permitted the Republican Party to achieve its central goal—­the establishment of a putatively egalitarian, ideologically respectable, conservatism. In 1980, Reagan and the GOP portrayed opposition to central elements of civil rights enforcement—­opposition to the use of race and sex preferences in hiring and in college admittance, to court-­ordered busing, and to the introduction of means-­tested programs for the poor—­as deriving from a principled concern for fairness: as a form of populist opposition to the granting of special privilege” (Edsall and Edsall 1992, 144). The language of equal opportunity—­often framed in terms of access—­was clearly more acceptable than the language of affirmative action to the “emerging Republican majority” (Phillips 1969). This was part of a broader shift from a “‘liberal’ paradigm centered on ‘social responsibility,’ in favor of a conservative paradigm centered around ‘legitimate self-­interest’” (Edsall and Edsall 1992, 97). In the legal arena, this approach manifested itself in the Reagan Department of Justice (DOJ) position on school desegregation and affirmative action, characterized as follows by the Edsalls: “the function of government intervention was not to correct contemporary employment, or contracting, or student assignment patterns that grew out of historic discrimination . . . Instead, the function of government was to protect individuals from specific acts of discrimination” (Edsall and Edsall 1992, 188). The Reagan DOJ’s assault on the 1971 Supreme Court decision in Griggs v. Duke Power (401 U.S. 424 [1971]) was emblematic of this shift. In Griggs, the Court found that employment policies that are facially neutral with regard to race—­such as aptitude test scores—­that nonetheless resulted in a highly disparate impact on a minority racial group could violate Title VII of the Civil Rights Act. To justify such tests, an employer would have to meet the high standard that there was a “business necessity for such test” (401 U.S. 431).



The Politics of Framing 33

The critique was framed in terms of individual merit being subordinated to racial quotas. For example, Linda Gottfredson of the University of Delaware argued that the logic of Griggs ignored “the whole issue of individual rights, and individual fairness . . . there is no provision for the constitutional rights, the civil rights of whites” (qtd. in Edsall and Edsall 1992, 252). Ultimately the Reagan DOJ’s approach made headway with the Supreme Court, as more conservative appointees shifted the Court’s ideological balance to the right. Later cases, beginning with City of Richmond v. J. A. Croson Company (488 U.S. 469 [1989], decided under Chief Justice William Rehnquist) and culminating in Parents Involved in Community Schools v. Seattle School District No. 1 (551 U.S. 701 [2007], decided under Chief Justice John Roberts), effectively rolled back affirmative mandates in employment and school desegregation reaching back to Brown. A particular focus of these cases was on individuals as a locus for constitutional protection. In cases such as Richmond v. Croson, this meant requiring institutions wishing to implement affirmative action programs to show particularized injury suffered by identifiable minority groups in the recent past. In her opinion for the majority, Justice Sandra Day O’Connor expressed the concern that “to accept [a] claim that past societal discrimination alone can serve as the basis for rigid racial preferences would be to open the door to competing claims for ‘remedial relief’ for every disadvantaged group. . . . Those whose societal injury is thought to exceed some arbitrary level of tolerability then would be entitled to preferential classification. We think such a result would be contrary to both the letter and the spirit of a constitutional provision whose central command is equality” (488 U.S. 505–­6 [1989]; emphasis added). O’Connor’s dismissal of the legal significance of the “societal injury” “thought” to be experienced by “minority groups” echoes Justice Brown’s similar dismissal, in Plessy, of claims about the stigma of segregation by asserting “if this be so, it is not by reason of anything found in the act, but solely because the colored race chooses to put that construction upon it” (163 U.S. 551 [1896]). He might just as easily have said that African Americans’ claims to “societal injury” resulting from enforced segregation were merely “thought to exceed some arbitrary level of tolerability” and hence did not deserve constitutional consideration. Despite writing nearly a century apart, both Justices Brown and O’Connor characterized the social realities of racial discrimination as troublesome intrusions into legal analysis. Brown wants to ignore them, O’Connor to transcend them. Both saw race as obscuring a more proper focus on individual rights and identities. Similarly, in Parents Involved, Chief Justice Roberts used a rhetoric of individual rights to strike down race-­specific school assignments meant to achieve the ideal of desegregation called for in Brown v. Board of Education. Roberts

34

Jonathan Kahn

struck down the assignment plans because they did not “provide for a meaningful individualized review of applicants but instead rely on racial classifications in a ‘nonindividualized, mechanical’ way” (551 U.S. 701 [2007]). The state, here, was forbidden from considering race as a primary basis for assigning students to public schools. At first blush, this might seem to be the logical culmination of Brown’s repudiation of Plessy’s state-­mandated racial segregation. On closer examination, however, it is revealed to be quite the opposite. In mandating a “color-­blind” approach to remedying the past effects of racial discrimination, Roberts effectively locates responsibility for addressing the consequences of racial discrimination in the private sphere of social life, beyond the purview of the state—­exactly where Justice Brown located it in his Plessy opinion. This “privatization” of discrimination is fully in line with the Reagan-­era legacy of neoliberal (or neoconservative) valorization of private, market-­based solutions as the appropriate means to address social problems. To the extent that the Rehnquist or Roberts courts accord any legal recognition to the social conditions of racial discrimination, they have done so as an interim concept that is to be transcended. Thus, in City of Richmond v. Croson, Justice O’Connor juxtaposed group-­based racial remedies against “the dream of a Nation of equal citizens in a society where race is irrelevant to personal opportunity and achievement” (488 U.S. 505–­6 [1989]). O’Connor reiterated her dream of a world where race was “irrelevant” in her 2003 opinion in Grutter v. Bollinger, where she anticipated that “25 years from now, the use of racial preferences will no longer be necessary to further the interest approved today” (539 U.S. 306, 343 [2003]). O’Connor’s dream of a country where “race is irrelevant” resonates with contemporary promises by biomedical researchers and clinicians yearning for a time when genetic information will similarly render race irrelevant. O’Connor, though, was speaking of making race socially and legally irrelevant, whereas pharmacogenomic experts now speak of making race biologically irrelevant. The two sentiments are connected, though, by their common disdain for the social realities of group-­based racial harms. Both view social concepts of race as impediments to progress, to be transcended and left to the dustbin of history. Both also valorize the individual as the ultimate target of intervention, whether legal or medical. Neoconservative race theorists tolerate the use of race only where such schemes retain a primary focus on considering the individual attributes of persons being classified. As Justice Scalia wrote in his concurring opinion in the 1995 case of Adarand v. Pena, “Individuals who have been wronged by unlawful racial discrimination should be made whole; but under our Constitution there can be no such thing as either a creditor or a debtor race” (515 U.S. 200, 239 [1995]). Similarly, biomedical researchers using race



The Politics of Framing 35

as an interim surrogate on the road to the promised land of personalized pharmacogenomic medicine focus on the individual bodies of patients to address what are often broad-­based social determinants of health. Race-­based pharmacogenomic practices ironically obscure or elide the social significance of group-­based harms to health, much as Scalia denied the significance of group-­ based harms to the civic standing of racial minorities. This approach aligns with a dynamic in law and politics identified by legal scholar Dorothy Roberts in which “race consciousness is decreasing in government social policy at the very moment it is increasing in biomedicine” (Roberts 2008, 538). All of this fits with the Edsalls’ analysis of the ideology engendered by the Reagan revolution which rejected “a ‘liberal’ paradigm centered on ‘social responsibility,’ in favor of a conservative paradigm centered on ‘legitimate self-­ interest’” (Edsall and Edsall 1992, 97). Applied to the framing of health disparities, this shift also marks the frames invoked to characterize health disparities in contemporary policy debates. Today’s Politics of Health Disparities

The touchstone for current debates on disparities remains the 2003 IOM report Unequal Treatment: Confronting Racial and Ethnic Disparities in Healthcare. This report was at the heart of the controversy involving DHHS Secretary Tommy Thompson’s suppression of its findings and the subsequent critiques from the likes of Sally Satel and Richard Epstein. The report itself defined disparities in health care as “racial or ethnic differences in the quality of healthcare that are not due to access-­related factors or clinical needs, preferences [defined as patients’ informed choices regarding health care], and appropriateness of intervention” (Smedley, Stith, and Nelson 2003, 3–­4). From the outset, then, the IOM report excluded issues of access and personal preferences from its consideration of disparities. This frame directed attention away from both market-­driven issues (such as facilitating the provision of pharmaceutical products to minority consumers) and individualizing issues (such as changing personal preferences or behavior). The report focused, instead, on (1) “the operation of healthcare systems and the legal and regulatory climate in which health systems function” and (2) “discrimination at the individual, patient provider level” (Smedley, Stith, and Nelson 2003, 4). The frame for the first item clearly located disparities in social structures that were the responsibility of the political community to address. The frame for the second item was more individualizing but still focused on the behavior of health care providers instead of patients, thereby implicating concerns for group-­based prejudice—­whether explicit or implicit—­in the provision of health care.

36

Jonathan Kahn

Much of the criticism of the report targeted its focus on structural issues, arguing for concentrating more on individual behavior and responsibility. The move from “disparities” to “difference” aimed to shift the frame of reference from social and political to individual and biological bases for the disparities chronicled by the report. This was accomplished, in part, through a blurring of the dual senses of the individual—­social and biological—­implicated by health disparities. Specifically, in terms of the initial frame of causation, the concept of “difference” (exemplified by Satel’s geneticization of racial difference with respect to BiDil) locates the source of disparities in racialized, individual biology. Second, in terms of the frame for responsibility in addressing the problem of disparities, the critics’ focus on such issues as individual behavior and access to medical care (or drugs) locates responsibility in racialized, individual social conduct. The first frame constructs a pathology of racialized bodies and has roots going back to Hoffman’s “Race Traits and Tendencies of the American Negro.” The second frame constructs a pathology of racialized social behavior that has similarly deep roots but is more directly traceable to the Moynihan Report on “The Negro Family” that suggested that many poor Black families were caught in a “tangle of pathology” that essentially caused their social and economic problems (Edsall and Edsall 1992, 52–­55). A focus on genes and personal responsibility combines concepts of biological and social pathology to locate responsibility for disparities firmly in individual members of minority groups. The blurring of the biological and the social in order to pathologize entire minority groups relates more broadly to recent, conservative approaches to health disparities that address individual behavior and “empowers” individuals by giving them increased access to “products” (including health care and drugs) in the medical “marketplace.” Thus, for example, in 2005, then Republican Senate Majority Leader (and medical doctor) Bill Frist prominently called for approaches to disparities that “promote dignity and personal responsibility” (Frist 2005, 447). While acknowledging historical and social forces that shape disparities, Frist nonetheless argues for a primary focus on programs that “will help decrease individuals’ risky behavior” (Frist 2005, 447). He went on to say that “the major causes of death among African Americans, for instance, are heart disease, cancer, stroke, accidents, and diabetes. Most of these are chronic diseases rather than acute illnesses, and all of these causes of death are at least arguably preventable” (Frist 2005, 447). By prevention, Frist means action by individuals who take personal responsibility for their poor health outcomes. In this formulation, Frist acknowledged social determinants of health but then omitted them from his actual framing of the problem that located responsibility primarily in the pathologies of



The Politics of Framing 37

“individuals’ risky behavior.” Having individualized the problem of disparities, Frist then called for fostering “competition” and “empower[ing] patients” by adopting market and consumer-­oriented reforms in place of government policies to address disparities (Frist 2005, 447–­48). Each of these remedies is firmly grounded in market-­based approaches that promote individual “choice” and “access.” In this model, health care is cast as a consumer good, not a civil right; and health itself becomes a matter of privatized risk-­management.1 Note 1. It echoes an earlier request, in 2000, from Republicans in the National Governors Association for federal permission to “design [Medicaid] benefit packages to look more like commercial models,” to “promote personal responsibility” and “to encourage choice through private health insurance” (Quadagno and McKelvey 2008, 19). References Adarand Constructors, Inc., v. Pena, 515 U.S. 200, 239 (1995). AHCA (Florida Agency for Health Care Administration). 2006. “Agency for Health Care Administration Announces the Addition of BiDil to Medicaid Preferred Drug List.”Accessed March 5, 2010. http://ahca.myflorida.com/Executive/Communications/ Press_Releases/archive/2006/05-01_BiDilFINAL.pdf. Braun, L. 2002. “Race, Ethnicity, and Health: Can Genetics Explain Disparities?” Perspectives in Biology and Medicine 45: 160–­63. Brown v. Board of Education Topeka, Shawnee County, Kan., et al., 347 U.S. 483 (1954). City of Richmond v. J. A. Croson Company, 488 U.S. 469 (1989). Du Bois, W. E. Burghardt, ed. 1906. “The Health and Physique of the Negro American.” Reprinted in “Voices from the Past: The Health and Physique of the Negro American by W. E. Burghardt Du Bois.” American Journal of Public Health 93 (2003): 272–­76. http://ajph.aphapublications.org/cgi/content/full/93/2/272. Douglas, J. Allen. 2003. “The ‘Most Valuable Form of Property’: Constructing White Identity in American Law, 1880–­1940.” San Diego Law Review 40: 881–­946. Edsall, Thomas Byrne, and Mary D. Edsall. 1992. Chain Reaction: The Impact of Race, Rights, and Taxes on American Politics. New York: Norton. Ellison, G., and I. Jones. 2002. “Social Identities and the New Genetics: Scientific and Social Consequences.” Critical Public Health 12: 265–­82. Epstein, Steven. 2007. Inclusion: The Politics of Difference in Medical Research. Chicago: University of Chicago Press. Ex parte Plessy, 11 So. 948 (La. 1892). Frist, William H. 2005. “Overcoming Disparities in U.S. Health Care.” Health Affairs 24: 445–­51. Griggs v. Duke Power Company, 401 U.S. 424 (1971). Grutter v. Bollinger, 539 U.S. 306, 343 (2003). Hoffman, Frederick L. 1892. “Vital Statistics of the Negro.” Arena 29: 542. ———. 1896. Race Traits and Tendencies of the American Negro. Vol. 11. New York: Publications of the American Economic Association. Hovenkamp, Herbert. 1985. “Social Science and Segregation before Brown.” Duke Law Journal 1985: 671–­72.

38

Jonathan Kahn

Kahn, Jonathan. 2004. “How a Drug Becomes ‘Ethnic’: Law, Commerce, and the Production of Racial Categories in Medicine.” Yale Journal of Health Policy, Law, and Ethics 4: 1–­46. ———. 2005a. “From Disparity to Difference: How Race-­Specific Medicines May Undermine Policies to Address Inequalities in Health Care.” Southern California Interdisciplinary Law Review 15 (2005): 105–­29. ———. 2005b. “Controlling Identity: Plessy, Privacy, and Racial Defamation.” DePaul Law Review 54: 755–­82. ———. 2008. “Exploiting Race in Drug Development: BiDil’s Interim Model of Pharmacogenomics.” Social Studies of Science 38: 737–­58. ———. 2012. Race in a Bottle: The Story of BiDil and Racialized Medicine in a Post-­ Genomic Age. New York: Columbia University Press. King, Martin Luther, Jr. 1966. Presentation at the Second National Convention of the Medical Committee for Human Rights, Chicago, March 25. Omi, Michael, and Howard Winant. 1994. Racial Formation in the United States: From the 1960s to the 1990s. 2d ed. New York: Routledge. Parents Involved in Community Schools v. Seattle School District No. 1, 551 U.S. 701 (2007). Phillips, Kevin. 1969. The Emerging Republican Majority. New York: Arlington House. Plessy v. Ferguson, 163 U.S. 537 (1896). Quadagno, Jill, and J. Brandon McKelvey. 2008. “The Transformation of American Health Insurance.” In Health at Risk: America’s Ailing Health System—­and How to Heal It, ed. Jacob Hacker, 10–­31. New York: Columbia University Press. Robbin, Alice. 2000a. “Classifying Racial and Ethnic Group Data in the United States: The Politics of Negotiation and Accommodation.” Journal of Government Information 27: 129–­56. ———. 2000b. “The Politics of Representation in the U.S. National Statistical System: Origins of Minority Population Interest Group Participation.” Journal of Government Information 27: 431–­53. Roberts, Dorothy E. 2008. “Is Race-­Based Medicine Good for Us?: African American Approaches to Race, Biomedicine, and Equality.” Journal of Law, Medicine, and Ethics 36: 537–­45. Sankar, P., et al. 2004. “Genetic Research and Health Disparities.” Journal of the American Medical Association: 2985–­89. Smedley, Brian D., Adrienne Y. Stith, and Alan R. Nelson, eds. 2003. Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington, DC: National Academies Press. Washington, Harriet. 2006. Medical Apartheid: The Dark History of Medical Experimental on Black Americans from Colonial Times to the Present. New York: Doubleday. Williams, D. R., and P. B. Jackson. 2005. “Social Sources of Racial Disparities in Health.” Health Affairs 24: 325–­34. Wolff, Megan J. 2006. “The Myth of the Actuary: Life Insurance and Frederick L. Hoffman’s Race Traits and Tendencies of the American Negro.” Public Health Reports 121 (January–­February): 89–­93. Yancy, Clyde. 2002. “The Role of Race in Heart Failure Therapy.” Current Cardiology Reports 4: 218–­25.

Chapter 3

Joseph L. Graves Jr.

Looking at the World through “Race”-­Colored Glasses The Fallacy of Ascertainment Bias in Biomedical Research and Practice

The mortality pattern difference between African Americans and Euro-­ Americans in the twentieth century was staggering. Figure 3.1 shows the ratios of age-­specific mortality from all sources for these two populations from 1963, 1980, 1996, and 2004. African American infant mortality was between two to two and a half times higher than that of European Americans, and the age-­ specific mortality of African Americans was always in excess of that of European Americans except at the oldest ages (greater than eighty-­five years). The data from 2004 show significant reduction of the differential in age-­specific mortality occurring between twenty-­five and fifty-­five years old. Yet, the robustness of these patterns has led some to suggest that racialized genetic differences are the best explanation of the disparity. The question still remains, however, whether either the theory or the available experimental data support that assertion. We shall see that the answer is no, and the belief that these differentials arise mainly from genetic differences results from an elementary logical fallacy called ascertainment bias. The working hypothesis of many within the medical community is that disease incidence differentials are heavily genetic and result from the reality of “race.” This assumption is made while actively ignoring what is known about the nature of human genetic variability. In short, anatomically modern humans are a young species with very little genetic variation compared to similar large-­ bodied mammals. Most of the variation we do have occurs in sub-­Saharan Africans, such that 85 percent of human variation exists within these populations. Human genetic variation is best explained through isolation by distance (Barbujani and Colonna 2010). This means that population groups geographically

39

40

Joseph L. Graves Jr.

2.75 2.50

Mortality Ratio

2.25 2.00 1.75 1.50 1.25 1.00

0

10

20

30

40

50

60

70

80

90

100

Age Age-Specific B/W Mortality Ratio in 1963 Age-Specific B/W Mortality Ratio in 1981 Age-Specific B/W Mortality Ratio in 1996 Age-Specific B/W Mortality Ratio in 2004 Figure 3.1   Age-­Specific Mortality, Selected Years (1963–­2004), Black/White Americans

closest to each other are most genetically alike. If we look at genetic variation on a continental basis (such as Africa, Asia, Europe, and so forth), we find that there is more genetic variation within such populations than there is between them. A case in point is the genetic variation at the angiotensin-­ converting enzyme locus (ACE). This locus has two common alleles created by the presence or absence of an ALU insertion. The insertion allele (I) ranges from approximately 0.10–­0.52 in Africa; 0.20–­0.85 in Western Asia; 0.25–­0.80 in Eastern Asia; 0.15–­0.54 in Europe; 0.40–­0.85 in North America; 0.30–­0.90 in South America; and 0.45–­1.00 in Oceania (Brutsaert and Parra 2006). Besides the large amount of genetic overlap in world populations, it is also known that Europeans contain more deleterious genetic mutations than sub-­Saharan Africans do (Lohmueller et al. 2008); thus a strictly genetic hypothesis of human disease prevalence would suggest Europeans and European Americans should be “sicker” than Africans or African Americans; yet higher African American



Looking at the World through “Race”-­C olored Glasses 41

age-­adjusted mortality in twenty-­two of twenty-­four categories is not considered problematic by this school of thought. An elementary grasp of how natural selection and genetic drift combine to produce genetic variation in all species including humans should seriously question the notion that African Americans are genetically more likely to have diseases than European Americans. In other words, what is the probability that the combination of natural selection and genetic drift produced such a one-­ sided distribution of mortality and morbidity in one population, one that, by the way, contains significant admixture from the other two? Thus, the fact that African Americans show higher rates of biological mortality in several infectious diseases (tuberculosis, pneumonia, and ulcer), hypertensive diseases, diabetes, heart diseases, cerebrovascular diseases, chronic obstructive pulmonary disease, chronic glomerulonephritis, and renal failure cannot simply be explained by differences in gene frequencies. Despite the overwhelming logic against such a position, biomedical researchers continue to proceed from that premise (see Chang et al. 2009). In any other field, an extremely powerful evidentiary base would be needed to support this sort of claim. However, health disparities research associated with diseases found disproportionally among African Americans seems to operate by different (and inferior) logical standards. Why We Suffer from Disease and How the Answer Vitiates Racial Medicine

The new field of “Darwinian” or “evolutionary” medicine has emerged over the last two decades (Gluckman, Hanson, and Beedle 2009; Graves 2011; Nesse and Williams 1996). It unifies and explains seemingly unrelated medical phenomena thus providing definite explanations of why humans suffer from disease. Evolutionary medicine proposes that disease results from one or a combination of five basic causes: infection (acute and stealth); genes; novel environments; design compromises; evolutionary legacies Infectious disease is a primary cause of health disparity. Outside of economically prosperous nations, the largest causes of death and sickness are still acute and infectious disease. Over 3.3 billion people are at risk for death from respiratory infections, HIV/AIDS, diarrheal diseases, tuberculosis, and malaria. For example, there are 300 to 500 million clinical cases of malaria yearly, accounting for 1.5 to 2.7 million deaths. Over half of the deaths of African children (age five or less) are due to malaria (Centers for Disease Control 2011). This is an example of a health disparity between the industrialized, wealthy nations and the underdeveloped, impoverished ones. With regard to health disparities in the industrialized world, stealth infections are still

42

Joseph L. Graves Jr.

quite important. The bacterium Heliobacter pylori can cause a stealth infection associated with stomach cancer. In the United States, East Asians (Japanese, Chinese, and Koreans) show higher incidence of this disease. However, human genetic factors are not the primary cause of this disparity. The risk factors include H. pylori infection, tobacco smoking, high salt intake, low intake of fruits and vegetables, high intake of pickled foods, high intake of nitrates and nitrate-­related compounds, and family history (genetics). Indeed, studies indicate that genetic variation in the bacterium (CagA-­+ strains) is highly associated with non-­gastric carcinoma in Taiwanese patients. In Taiwan, 95 percent of H. pylori strains are CagA+. In short, you cannot assume that any cancer disparity you observe results from genetic differences among human populations, since so many other factors are at play. Recent work on the prevalence of the triple negative tumor (TNT) breast cancer phenotype shows that it is related to a number of life-­history attributes and behaviors that are not shared equivalently by African American and European American women. TNT breast cancer is particularly aggressive and does not respond well to the existing cancer drugs. A number of life-­history variables are associated with a greater risk of these basal-­like breast cancers: lower age at menarche, increased risk for parity, younger age at first full-­term pregnancy, shorter duration of breastfeeding, lower number of children breastfed, decreased number of months of breastfeeding, multiple live births, not breastfeeding, and use of lactation medication. In addition, body mass index and elevated waist-­to-­hip ratio were associated with greater risk of basal-­like breast cancer in pre-­and post-­menopausal women. Other variables that are associated with greater risk of basal-­like breast cancers include diet, duration of smoking, and poverty (Lund et al. 2009; Millikan et al. 2008). Complex diseases are those determined by a combination of genetic, environmental, and chance effects. These always show an age-­associated pattern. Evolutionary biology has demonstrated that the age-­associated pattern of complex disease in metazoans (eukaryotic, multicellular organisms such as ourselves) results from the declining force of natural selection with age. In other words, natural selection is primarily concerned with reproductive success at younger age, thus it does not have the power to eliminate deleterious mutations that are expressed after an individual’s net future reproductive value is zero. In modern humans, this generally occurs between ages forty-­five to fifty-­five years old, but one must remember that our patterns of aging evolved in previous human societies with shorter life spans. For example, in sixteenth-­century Europe, the average life span was only about thirty-­five years.



Looking at the World through “Race”-­C olored Glasses 43

The declining force of natural selection means that the pattern of aging will be determined by the action of two population genetic mechanisms: antagonistic pleiotropy and mutation accumulation. Antagonistic pleiotropy refers to genes that have a beneficial effect at early age and are therefore favored by natural selection despite the fact that they have a deleterious impact at later age. Mutation accumulation refers to genes that have no impact on early life fitness, so they can accumulate in populations via random chance (genetic drift) causing a variety of late-­life pathologies. In the terms of evolutionary medicine, mutation accumulation results from an evolutionary legacy (for example, the inability of natural selection to eliminate it). As humans age, antagonistic pleitropy mechanisms are shared widely within our species, although individual and population variation exists around them. Good examples of antagonistic pleitropy–­associated aging diseases are the various cancers experienced at late ages. They are widespread because the genes responsible for them have advantageous effects on early-­life fitness (specifically development). Mutation accumulation age-­associated mechanisms are more variable among individuals and populations. An example of this is Alzheimer’s disease. Genetic variation at the apolipoprotein locus is highly associated with Alzheimer’s disease, particularly the ε4 allele. In Europe, the frequencies of ε4 in large samples are close to 0.150–­0.110; in Chinese samples, 0.049–­0.081; Amerindian samples, 0.112–­0.193; and sub-­Saharan African and African American samples, ~0.250 (ALFRED 2011). Since mutation accumulation results from genetic drift, small populations should be even more variable in mutation accumulation mechanisms. Consistent with this claim is the variation that occurs in ε4 frequency in small populations. The frequency for the Khoi-­San of Southern Africa was 0.370, the Biaka pygmies 0.407, and five Brazilian Amerindian groups ranged from 0.423 to 0.043. Most people have little exposure to evolutionary medicine or biological anthropology. For this reason, even my most adept students readily accept the notion that races differ in disease predisposition and prevalence. Almost all of them cite the example of sickle cell anemia as a “Black” disease. These students do not realize that the sickle cell anemia allele provides resistance to malaria and therefore is in high frequency in malarial zones. For this reason, Greeks have a higher frequency of the allele than African Americans or native South Africans. There are large frequency differences even within the nation of Kenya, with high-­altitude Kenyans displaying little to no sickle cell anemia, as opposed to low-­altitude Kenyans showing much higher frequencies of the disease. There is little sickle cell anemia at high altitudes around the world, since these environments do not support the mosquitoes that carry malaria. In terms

44

Joseph L. Graves Jr.

of evolutionary medicine, the allele that causes sickle cell anemia results from adaptation to resist acute infection by malaria. Genes, Evolutionary Mechanisms, and Disease

Evolutionary medicine predicts that genes can be a source of disease. Mutation selection balance, antagonistic pleitropy, and mutation accumulation combine to produce genetic predisposition for disease. Natural selection acts to reduce the frequency of alleles that cause some of the worse inborn errors of metabolism. These alleles will be found at a frequency balancing the force of natural selection against them and their mutation frequency. An extreme example of this is illustrated by the dominant mutations that cause progeria. Progeria results from a dominant mutation in the Lamin A gene found on chromosome 1q. Lamin A is a protein that plays an essential function in the nuclear membrane, regulates chromatin, and helps with nuclear integrity and shape. These mutations are called pleiotropic because this protein impacts so many physiological features (one mutation impacting many traits is called pleiotropy). These mutations create a phenotype that appears to accelerate aging. Most people who carry this mutation do not live past thirteen years old and so never reproduce. This means that the only way these alleles can appear is by spontaneous mutation (error in replicating the genetic code). The frequency of progeria ranges between one in 8 million and one in 4 million live births. At the time of this writing, there were only fifty people in the world with this disease, and one form was so rare that only six of the fifty carried that specific mutation. Clearly, no serious discussion of health disparity can be enjoined around these rare mutations. Despite this, some biomedical researchers conflate differences in the frequency of rare mutations (such as Crohn’s disease) with the more common mutations that play roles in complex diseases. Antagonistic pleitropy and mutation accumulation operate on alleles found at higher frequencies, which contribute to the eventual manifestation of (or resistance to) disease. The frequencies of these alleles are caused by a combination of selection along gradients (clines) of environmental factors and genetic drift resulting from peculiarities of population history. These processes are responsible for the principle of discordance. Discordance means that physical traits are not consistently correlated with each other and this means that human populations cannot be unambiguously classified into racial groups. For example, differences in skin color result from weak selection along the gradient of solar intensity from the tropics to the arctic. Thus, tropical populations have darker skin that protects them from higher solar intensity and skin color shows gradual lightening progressing toward higher latitudes. Such change is called a



Looking at the World through “Race”-­C olored Glasses 45

cline, which means a gradient of continuous change in either a physical character or genetic composition within a species. Clines can exist for several reasons. A cline could result if natural selection favored slightly different features along a gradient. Also a cline can exist if there was adaptation to features of two different environments and gene flow between them. In the eastern hemisphere, skin color shows a clear cline related to solar intensity, which was established over a relatively long time period. However, populations in the western hemisphere have not lived there long enough to reestablish the cline (arriving between 35,000 to 11,000 years ago). The distribution of alleles affecting melanin in the skin follows the cline, and there is a well-­established relationship between population variation and resistance to diseases associated with sun exposure. All populations living in the tropics show high melanin concentration in their skin. This includes groups that were classified by nineteenth-­century anthropologists as “Caucasians,” “Australoids,” “Amerindians” in Central America, “Mongoloids” in Indochina, and “Negroids” in Central Africa. Despite the concordance between skin color and solar intensity, the incidence of skin melanomas does not map directly onto skin color. For example, Australians, who have very high melanin content, show some of the highest rates of melanoma, perhaps from cultural practices that lead to greater lifelong sun exposure. Of all the evolutionary mechanisms of disease, novel environment is probably the most explanatory with regard to health disparities. The rate of cultural and technological change is orders of magnitude larger than the rate of genetic change within our species. To get a sense of novel modern environments consider the following events in the life of the human species condensed into a calendar year as a heuristic device (table 3.1). For example, radiation sickness could never have been a major problem to humans until December 31 on the calendar of our existence. It required the development of nuclear weapons in the 1940s and the nuclear power industry in the 1950s. While nuclear materials are a clear example, their impact on health disparities at this time is limited. Far more subtle and pervasive examples of the impact of novel environments exist. One of the most significant novel environments for our species resulted from the development of agriculture, the Industrial Revolution, and resultant food surpluses experienced by those living in economically affluent nations. For the majority of our existence as a species, we consumed fruits, vegetables, underground storage organs (roots, tubers, and bulbs), insects, and meat. None of this was available in excess, and gathering these foods required significant amounts of daily activity. Now, grains, refined fats, nonhuman milk, refined

46

Joseph L. Graves Jr.

Table 3.1   Life of Homo sapiens Condensed into a Calendar Year Event

Date

Origin of species

January 1

Some of us leave Africa

June 2

Some arrive in New Guinea

August 22

Some arrive in Europe

September 17

Ice man dies in the Alps

December 1

Agriculture starts in the Middle East

December 4

Black Death in Europe

December 29 morning

Trans-­Atlantic slave trade

December 29 afternoon

Industrial Revolution

December 29 evening

Cardiovascular disease prevalent

December 30

Nuclear age begins

December 31 morning

Obesity epidemic in the United States of America

December 31 afternoon

sugars, and salts make up the bulk of caloric and mineral intake in most countries (Lindeberg 2010). For those living in affluent nations, little exercise is required to acquire these foods in excess. Not surprisingly, when we look at the list of diseases most associated with poor health (diabetes, heart disease, hypertensive diseases, cerebrovascular diseases, thrombosis, chronic obstructive pulmonary disease, and so on), all are impacted by excess calories, lack of exercise, and too much salt intake. For example, there are excellent physiological reasons why high-­fructose diets (particularly high-­fructose corn syrup) contribute to endocrine and metabolic imbalance (Stanhope and Havel 2008). Yet it can be argued that if these are all diseases of modern civilization and food excess, why are there disparities in prevalence and mortality associated with them? The answer lies in another aspect of novel environment, social subordination. The preeminent role of social subordination in producing health disparities is borne out by both global and within-­nation patterns of morbidity and mortality (Wermuth 2003). The major factor determining whether individuals will live longer lives and experience less life-­threatening illness is the availability of living conditions that meet basic needs. Affluent countries have higher life expectancies than poor countries. Improvements in living conditions among populations below the epidemiological threshold result in significant improvements in health and longevity. Across societies, there is an association between the degree of social inequality and the rate of disease and death. Social inequality impacts health in direct and indirect ways. For example, psychosocial factors due to one’s position in the social hierarchy



Looking at the World through “Race”-­C olored Glasses 47

may impact trust and participation in the society. The higher social status gives greater individual health and longevity chances by direct (nutrition, access to health care) and indirect mechanisms (psychosocial; Kawachi and Kennedy 2002; Wermuth 2003). All of these patterns point to the role of social subordination on health outcomes and health disparities in particular. One powerful example of the novel social environment effect on health is given by the change that occurs in the morbidity profiles of immigrants of African descent when they come to live in the United States. For example, compared to U.S.-­born persons of African descent, African-­descended immigrants from minority European-­ descended societies (Asia, South America) and racially mixed regions (West Indies) have superior health, while those from majority European-­descended societies (Europe) regions fare no better than native-­born African Americans. A similar gradient exists among African-­ descended immigrants, with Africans faring the best, followed by South Americans, then West Indians, and African-­descended individuals living in Europe showing the poorest health (Read, Emerson, and Tarlov 2005). The authors of this study conclude that the negative health effect results from being a “racial” minority in a White majority society. Design compromises and evolutionary legacies are characteristics of all species. For example, in our ancestral lineage, the digestive tube and trachea cross each other (evolutionary legacy). The epiglottis closes when we are moving food down the esophagus and opens when we are breathing. Due to this fact, it is possible for food to get caught in the trachea and individuals may choke to death. This possibility occurs in all population groups so it is not likely to be a cause of a health disparity; although a case could be made for other forces acting on an evolutionary legacy and thus causing a health disparity. For example, in the United States, African American children drown at three to four times the rate of European American children. The legacy here is lungs (not gills); but this is a problem due to social forces, which have made it more difficult for African American children to learn to swim (Brenner, Smith, and Overpeck 1994; Saluja et al. 2006). Walking upright or the contradiction between pelvic girdle width and human infant size is a design compromise. Again, all humans share this compromise; so it is not likely to contribute to health disparities unless other factors intervene; women without proper access to appropriate medical care could show greater rates of mortality in childbirth. Final Philosophical Difficulties of Race and Medicine

The latter half of the twentieth century saw major advances in molecular genetics. This began with the determination that DNA was the molecule of heredity,

48

Joseph L. Graves Jr.

followed by the eventual sequencing of whole organism genomes (including humans), improved techniques for the extraction of DNA from fossils, and greater bio-­computational capacity allowing interpretation of genetic sequence data. However, these advances have not occurred without a price. The rise of molecular reductionism resulted in part from the necessary growth of specialization as the knowledge base grew within each subdiscipline. Specialization almost always entails a lack of communication with allied fields. The lack of communication between evolutionary and molecular genetics has caused failures in both disciplines. Ignoring the significance of molecular mechanisms fostered a naïve allegiance to adaptationism in evolutionary biology, while underestimating the significance of evolutionary genetics contributed to misinterpretation of the origin and significance of genetic variation in molecular biology. The failure of molecular genetics to properly integrate evolutionary thinking has contributed to the spread of reductionism within the discipline. Molecular reductionism assumes that each phenotype can be simply traced backward to a specific DNA sequence. It ignores all the complex interactions that occur between gene and phenotype. In statistics, ascertainment bias occurs when some members of a population are less likely to be sampled than others. If you draw conclusions on a biased sample, then your conclusions may be flawed. If the sampling is purposefully biased (even if unconsciously), then racial ascertainment bias results in seeing the results that you wished to see; in a sense, you are viewing the world through “race-­colored glasses.” In the case of race and biomedical research, using socially defined racial categories as if they were legitimate biological divisions is just such an example. Examining genetic differences that occur in such groups and assuming that these are responsible for the differential morbidity and mortality in these groups ignores all the other relevant variables that are different between the groups. This is a form of ascertainment bias. That is, when researchers differentially pursue one set of hypotheses while ignoring equally likely hypotheses, the results are by necessity incomplete. In the case of the United States, there is a clear and well-­documented history of environmental differences between socially defined groups (Olden and White 2005). These environmental differences work on the existing genetic variation within population groups. Thus it can be asked whether it is simpler to address the genetic differences via medical intervention, or to address the ongoing physical and social differences. The environmental differences can also be addressed via medical intervention; although it seems more logical, as well as more ethical, to address social and environmental differences via less technical and invasive techniques.



Looking at the World through “Race”-­C olored Glasses 49

The following exaggerated example makes the point. Kistner et al. (2007) found that African American boys grades three through five reported significantly more depressive symptoms on the Child Depression Inventory than did European American boys. No differences emerged between African American and European American girls in this study. Lower academic achievement scores were associated with increased depressive symptoms for the African American boys. It is clear that you could simply treat the depression symptoms of these children via drug therapy, or you could provide this group with drugs that boost mental function addressing the lower academic achievement and thus ameliorating the depression. However, neither of these strategies would be effective in the long run, free from long-­term physiological harm, nor considered ethical in our society. It would seem that addressing the social/environmental differences impacting young African American males would be more likely to produce long-­term and sustainable gains, which would also be less likely to cause harm and therefore be more ethical. This leaves us with some monumental elephant-­in-­the-­room sorts of questions. Why do many biomedical researchers assume that measured genetic differentials necessarily drive health disparities? Why are these assumptions made when current population genetic and evolutionary theory so strongly argues against their utility? The answer cannot simply lie in the ignorance, by researchers, of evolutionary theory or of discussions explaining the ambiguity of socially defined racial ideas (see Caulfield et al. 2009). Indeed, if they used the standard strong inference reasoning (Platt 1964) expected of normal science, they would clearly see through their fallacies. Indeed, it can be suggested that the problem resides in the social character of the biomedical profession itself. The world’s existing social hierarchy means that the vast majority of this research is carried out within the borders of economically prosperous nations (who were made prosperous by the very racial enterprise critiqued in this discussion). Within those nations, those from the dominant socially defined racial groups are far more likely to enter careers in the field, have more success, and achieve prestigious academic appointments. For example, Ginther et al. (2011) reported that between the years 2000 and 2006, of the investigators applying for R01 grants from the National Institutes of Health only 1.4 percent were black, 3.2 percent were Hispanic, and 0.05 percent were Native American. Indeed, these figures overestimate members of racially socially subordinated groups in the United States because the NIH uses the term “Black” to include persons from Africa and the Caribbean who are of African descent but who are not African Americans. Moreover, it is also difficult to interpret the term “Hispanic” in NIH lingo, which can mean anyone who primary language is Spanish,

50

Joseph L. Graves Jr.

including those who may have grown up in aristocratic luxury outside of the United States. Yet even with these broad categories in play, Blacks were shown to be least likely to have a R01 grant funded by NIH compared to Hispanics, Asians, and Whites. It is highly likely that the bulk of the biomedical research community, which originated in racially favored positions, has spent a great deal of time problematizing the racial ethos of this society. Add to this their inadequate training and exposure to evolutionary medicine, and finally the runaway freight train of translational research (see Maienschein et al. 2008), and we should not be surprised by the ongoing dominance of socially defined racial confusion in the field. These components suggest actionable items to help us take off the race-­ colored spectacles. First, universities must implement instruction that exposes all students to the fallacies of the American racial mythology. No university-­ educated person should graduate without course work in this area. Second, evolutionary biology should be inculcated throughout the biology major and premedical curriculum. Indeed, medical schools should be offering courses in evolutionary medicine (the author is presently part of a working group supported by the National Evolutionary Synthesis Center devoted to just these tasks). Third, more serious efforts must be made at the societal level to address underrepresentation of socially subordinated racial groups in science. This is a task that must also be championed by these underserved communities themselves. Finally, it is imperative that this society rethink translation and halt the inappropriate influence that private industry has over basic research. This is systematically biasing research in the direction of reductionism. I fear that if these actions are not taken, and soon, that not only will the health of the society continue to deteriorate (and health disparities widen) but our democracy may also be at risk. References ALFRED—­Allele Frequency Database. 2011. National Science Foundation. Accessed July 24, 2012. http://alfred.med.yale.edu. Barbujani, Guido, and Vincenza Colonna. 2010. “Human Genome Diversity: Frequently Asked Questions.” Trends in Genetics 26: 285–­95. Brenner, Ruth, Gordon S. Smith, and Mary D. Overpeck. 1994. “Divergent Trends in Childhood Drowning Rates, 1971 through 1988.” Journal of the American Medical Association 271: 1606–­8. Brutsaert, Tom, and Esteban Parra. 2006. “What Makes a Champion? Explaining Variation in Human Athletic Performance.” Respiratory Physiology and Neurobiology 151: 109–­23. Caulfield, Timothy, Stephanie M. Fullerton, Sarah E. Ali-­Khan, et al. 2009. “Race and Ancestry in Biomedical Research: Exploring the Challenges.” Genome Medicine 1 (1): 1–­8.



Looking at the World through “Race”-­C olored Glasses 51

Chang, Manhuei, et al. 2009. “Prevalence in the United States of the Selected Candidate Gene Variants: Third National Health and Nutrition Examination Study, 1991–­1994.” American Journal of Epidemiology 169: 54–­66. Centers for Disease Control. 2011. “Malaria Impact.” Accessed February 23, 2012. http:// www.cdc.gov/malaria/malaria_worldwide/impact.html. Cosa, Aldo, Antonio J. Silva, Nuni D. Garrido, et al. 2009. “Association between ACE D Allele and Elite Short Distance Swimming.” European Journal of Applied Physiology 106: 785–­90. Ginther, Donna K., Waler T. Schaffer, Joshua Schnell, et al. 2011. “Race, Ethnicity, and NIH Research Awards.” Science 333: 1015–­19. Gluckman, Peter G., Mark Hanson, and Alan S. Beedle. 2009. Principles of Evolutionary Medicine. New York: Oxford University Press. Graves, Joseph Jr. 2005. The Race Myth: Why We Pretend Race Exists in America. New York: Dutton Books. ———. 2006. “What We Know and What We Don’t Know: Human Genetic Variation and the Social Construction of Race.” In Is Race Real?, solicited by the Social Science Research Council, edited by Craig Calhoun, president, Social Science Research Council, published on June 7, 2006. Accessed February 12, 2012. http://raceandgenomics .ssrc.org/Graves/. ———. 2010. “Biological vs. Social Definitions of Race: Implications for Modern Biomedical Research.” Review of Black Political Economy 37 (1): 43–­60. Accessed February 23, 2012. DOI: 10.1007/s12114-­009-­9053-­3. ———. 2011. “Evolutionary versus Racial Medicine: Why It Matters.” In Race and the Genetic Revolution: Science, Myth, and Culture, ed. Sheldon Krimsky and Kathleen Sloan. New York: Columbia University Press. Graves, Joseph L., Jr., and Michael R. Rose. 2006. “Against Racial Medicine.” Patterns of Prejudice 40 (4–­5): 481–­93. Henderson, J. Neil, Richard Crook, Julia Crook, et al. 2002. “Apolipoprotein E4 and Tau Allele Frequencies in Choctaw Indians.” Neuroscience Letters 324 (1): 77–­79. Kawachi, Ichiro, and Brian P. Kennedy. 2002. The Health of Nations: Why Inequality Is Harmful to Your Health. New York: New Press. Kistner, Janet A., Corinne F. David-­Ferdon, Christina M. Lopez, and Stephanie B. Dunkel. 2007. “Ethnic and Sex Differences in Children’s Depressive Symptoms.” Journal of Clinical Child and Adolescent Psychology 36: 171–­81. Lindeberg, Staffan. 2010. Food and Western Disease: Health and Nutrition from an Evolutionary Perspective. West Sussex, UK: Wiley-­Blackwell. Liu, Hsiu-­Chih, et al. 1999. “ApoEGenotype in Relation to A.D. and Cholesterol: A Study of 2,326 Chinese Adults.” Neurology 53: 962. Lohmueller, Kirk E., Amit R. Indap, Steffen Schmidt, et al. 2008. “Proportionally More Deleterious Genetic Variation in European than in African Populations.” Nature 451: 994–­98. Lund, Mary Jo, Katrina F. Trivers, Peggy Porter Peggy, et al. 2009. “Race and Triple Negative Threats to Breast Cancer Survival: A Population-­Based Study in Atlanta, GA.” Breast Cancer Research and Treatment 113 (2): 357–­70. Maienschein, Jane, Mark Sunderland, Rachel Ankeny, and Jason Scott Robert. 2008. “The Ethos and Ethics of Translational Research.” American Journal of Bioethics 8 (3): 43–­51. Marin, Guilherme B., Marli H. Tavella, Joao F. Guerreiro, et al. 1997. “Short Communication: Absence of the E2 Allele of Apolipoprotein in Amerindians.” Brazilian

52

Joseph L. Graves Jr.

Journal of Genetics 20 (4). Accessed February 23, 2012. http://dx.doi.org/10.1590/ S0100-84551997000400029. Millikan, Robert C., Beth Newman, Chiu-­Kit Tse, et al. 2008. “Epidemiology of Basal-­Like Breast Cancer.” Breast Cancer Research and Treatment 109 (1): 123–­39. Neese, Randolph M., Cart T. Bergstrom, Peter T. Ellison, et al. 2009. “Making Evolutionary Biology a Basic Science of Medicine.” Proceedings of the National Academy of Sciences: September. Accessed February 23, 2012. http://www.pnas.org/cgi/doi/10.1073/ pnas.0906224106. Nesse, Randolph M., and George C. Williams. 1996. Why We Get Sick: The New Science of Darwinian Medicine. New York: Vintage. Olden, Kenneth, and Sandra White. 2005. “Health-­Related Disparities: Influence of Environmental Factors.” Medical Clinics of North America 89: 721–­38. Platt, John R. 1964. “Strong Inference: Certain Systematic Methods of Scientific Thinking May Produce Much More Rapid Progress than Others.” Science 146 (3642): 347–­53. Read, Jen’nan Ghazal, Michael O. Emerson, and Alvin Tarlov. 2005. “Implications of Black Immigrant Health for U.S. Racial Disparities in Health.” Journal of Immigrant Health 7 (3): 205–­12. Saluja, Gitanjali, Ruth A. Brenner, Ann C. Trumble, et al. 2006. “Swimming Pool Drownings among U.S. Residents Aged 5–­24 Years: Understanding Racial/Ethnic Disparities.” American Journal of Public Health 96 (4): 728–­33. Smith, William B. 1905. The Color Line: A Brief in Behalf of the Unborn. New York: McClare, Phillips. Stanhope, Kimber, and Peter Have1. 2008. “Endocrine and Metabolic Effects of Consuming Beverages Sweetened with Fructose, Glucose, Sucrose, or High-­Fructose Corn Syrup.” American Journal Clinical Nutrition 88 (Suppl.): 1733–­37S. Van der Flier, W. M., Y. A. Pijnenburg, S. N. Schoonenboom, et al. 2008. “Distribution of APOE Genotypes in a Memory Clinic Cohort.” Dementia and Geriatric Cognition Disorder 25 (5): 433–­38. Wermuth, Laurie. 2003. Global Inequality and Human Needs: Health and Illness in an Incredibly Unequal World. New York: Pearson Education. Wu, I-­Chen, Deng-­Chyang Wu, Fang-­Jung Yu, et al. 2005. “Association between Heliobacter pylori Seropositivity and Digestive Tract Cancers.” World Journal of Gasteroenterology 15 (43): 5465–­71.

Chapter 4

Jay S. Kaufman

Ethical Dilemmas in Statistical Practice The Problem of Race in Biomedicine

Most discussions of ethics in statistical practice revolve around declaring conflicts of interest due to funding sources and engaging in honest descriptions of data and procedures (American Statistical Association 1983, 1999; International Statistical Institute 1985; Jowell 1981). This is all well and good, but such a narrow focus on greed and fabrication ignores some of the more interesting and intractable ethical problems necessarily involved in inference and prediction, especially when social groups are transformed into statistical categories. Racial comparisons therefore provide an interesting case study of some of these problems, as they are frequently the targets of statistical analysis in clinical medicine and public health, and they are imbued with all kinds of baggage due to our understandings about racial groups in the social world (Marshall 1993). The history of modern statistics is intimately intertwined with the quantitative demonstration of racial differences, with obvious implications for the justification and maintenance of existing social hierarchies. For example, the increasingly intricate quantifications of late nineteenth-­century craniology propelled the development of regression techniques by Francis Galton, while the nascent field of intelligence testing gave rise to factor analysis (Gould 1981). The development of psychometrics as a quantitative discipline was motivated by the task of reifying intelligence as a measurable trait and arraying racial or ethnic groups in a ranked fashion across this singular dimension of innate capacity. The ethical implications of this work can be seen in their application to a wide variety of social policies, including immigration restrictions and forced sterilizations (Lewontin, Rose, and Kamin 1993).

53

54

Jay S. Kaufman

The problems we face in such comparisons are more fundamental than those related to the assumptions and technical limitations of specific statistical models, however. Even the simplest descriptions of disparity invoke conundrums in the choice of contrasts, which can paint dramatically different pictures of the state of the world (Harper et al. 2010; Scanlon 2006). For example, the simple technique of epidemiologic standardization is a ubiquitous feature of surveillance because it facilitates “fair” comparisons between populations that are imbalanced by another factor (for example, age). The procedure involves taking a weighted average of the stratum-­specific measures, where the weights are taken from some population that is defined as the common standard. The seemingly innocuous choice of the standard population, however, can lead to drastically different impressions. An illustration of this phenomenon occurred when the U.S. government switched from the 1970 census to the 2000 census population standard for its official statistics, with the consequence that the magnitude of racial disparity in mortality decreased overnight. This occurred because the year 2000 population was older, and because ratio measures of disparity are more modest in older age, so a heavier weighting at the older end of the population reduced the magnitude of disparity overall (Krieger and Williams 2001). This is not to argue that the 1970 standard population is any more or less valid, only that any standard is a necessary fiction; the real world remains stubbornly unadjusted. Categorization

Just as standard populations are fictional representations of our world, so are virtually all categorized variables in statistical analyses. For example, while there really is something called height, which can be measured objectively along a continuum, there are no objective standards for what constitutes a “tall” or a “short” person. Human variability is similarly continuous or near continuous in most respects, and so categorizations of study subjects into “Black” and “White,” or “poor” and “non-­poor,” or “hypertensive” and “normortensive,” are all to some extent arbitrary, and often inflected by myths and traditions that end up being imposed on individuals as essential characteristics that come to define them (Little 1998). Even the categories “male” and “female,” which obviously have some biological significance in the natural world, are nonetheless carefully tidied into a dichotomy when Mother Nature is not so abhorrent of ambiguity (Dreger 1998). Epidemiologists and biostatisticians have a long tradition of favoring categorized variables, much more so than is common practice in psychology or economics. For the left-­hand-­side (response) variable in a regression model, this is



Ethical Dilemmas in Statistical Practice 55

probably due to the fact that we traditionally study “disease,” which is some kind of clinical categorization of health status. Some disease outcomes are reasonably concrete, such as a fractured bone or infection by a parasite. Others, however, are closer to utter fabrication. “End stage renal disease” for example, emerged as a disease through the application of a billing code (Peitzman 2007), whereas “female sexual dysfunction” has recently been established as a medical entity in response to a perceived marketing opportunity (Tiefer 2006). Biomedicine has generally been equally fond of categorization on the right-­ hand (independent) variable side of the regression equation as well, and this is how study subjects end up being binned together in one way or another—­by race, age, sex, and so forth. An important ethical component of such categorization naturally arises around questions of who gets to make such distinctions, and on what basis. Ethnic groups are overtly political in their formation and classification, such as the creation of the “Yoruba” ethnicity by nineteenth-­century British colonials (Rotimi 2003), but racial groups no less so, as evidenced, for example, by the recently announced decision that Chinese in South Africa are now to be considered as “Black” (BBC News 2008). The history of many social groupings has been similarly politicized, with various American groups lobbying at different times to be permitted into the “White” category, for example (Nobles 2000). What is important in ethical terms is whether or not subjects have some autonomy in determining their own racial or ethnic categorization. Since no objective “gold standard” exists, mustn’t we simply accept that people are who they say they are? Most health studies do indeed rely on self-­reported ethnic or racial identity, but there are two moments in the lifespan—­its beginning and its end—­when people have no self-­identity, and where methods for making assignment therefore become less logically consistent (Hahn 1999). Ironically, most routinely collected demographic information comes from exactly these two singular moments when self-­identity is absent. Furthermore, because race and ethnicity are just as much about social recognition as they are about self-­ identity, there are limits on what we can assert ourselves to be; others will judge us by physical appearance, language, and behavior, and there is a rich American tradition of clandestine identity (Kennedy 2001; Malcomson 2000). Biomedical researchers therefore commonly invoke several imbedded layers of mythology involving bins. They imagine that the bins exist, both on the exposure and the outcome sides of the regression equation, and then they imagine that we can sort people into these bins. To say that such categorization involves misclassification on both sides of the equation is overly credulous, because “misclassification” implies an unknown truth, and the fact is that, for classifications that have no objective reality, there can be no appeal to this kind

56

Jay S. Kaufman

of anchor. For example, to say that “Black Americans” are at higher risk of “schizophrenia” is to conflate all manner of historical, sociological, and biological issues, since there is no necessary nor stable meaning to the boundaries for either of those two categories (Breslau et al. 2006). Moreover, the observation becomes a self-­fulfilling prophesy, since subsequent clinical judgments are affected by this common knowledge (Loring and Powell 1988). This last point warrants some further explanation, because diagnosis is an estimation procedure and therefore has a well-­described statistical logic. The clinician observes some data (for example, history, test results, signs and symptoms, and so on) and uses this information to assign the patient to a disease category. If definitive confirmation is available at some later time, then we can think of this as an assignment that has some quantifiable performance in terms of error. Thus, various specific clinical observations entail their own distinctive diagnostic characteristics, for example, test sensitivity (the proportion of truly diseased who are labeled as having disease) and specificity (the proportion of truly undiseased who are labeled as not having the disease). In the ongoing probabilistic game of medical diagnosis, however, a given test result serves as data to update a prior probability, providing a posterior probability (Sox et al. 2007). The starting point of this process is generally taken to be the population prevalence of the condition. For example, if I am testing to see if a patient is infected with syphilis, my laboratory test is applied to a baseline probability of syphilis in the relevant population. There are two ethical problems with this paradigm. The first is that the clinician must decide what is the relevant population to which the patient belongs. The American medical tradition in this regard is to describe the patient in terms of age, sex, and race, with the privileging of race in this context being a somewhat indefensible historical artifact (Finucane and Carrese 1990). The obvious ethical discomfort here results from the clinician choosing which groupings are salient for the purpose of defining the referent population. But a second dilemma arises from the social process of science, because the available information on baseline prevalence includes whatever previous biases existed and then reproduces them. For example, suppose that police at some time in the past were racist and differentially stopped the cars of Black motorists in order to search for illegal drugs (Meeks 2000). This practice would generate flawed data that would show Blacks to be more likely to have drugs in their vehicles even if the true underlying prevalence were uncorrelated with race. Now imagine that today’s police have shed all previous irrationalities and are interested simply in searching cars that have a high probability of containing illegal drugs. The data will show dispassionately that cars driven by



Ethical Dilemmas in Statistical Practice 57

Blacks have a higher probability of containing illegal drugs, leading them to be targeted more frequently for search, in turn leading them to maintain a higher prevalence of illegal drugs in the available statistics. Differential case finding by race in the process of clinical care seems all too often to follow this general scenario (Garcia 2003). Attribution of Blame

The previous paragraphs have discussed how disparities are established in biomedical research and clinical practice. Once such disparities are described, however, then they must be classified according to some metric of moral outrage. We are naturally troubled that Black Americans live five years less on average than White Americans, for example, but not so indignant about shorter life expectancies for heavy cigarette smokers. While a disparity is nothing more than a contrast of two risks or rates, a disparity in the sense of a public health priority must be something that involves a perceived injustice, which is the ethical judgment inherent in the distinction between inequalities and inequities (Asada 2005). How offended we ought to be over a given disparity is not often gauged according to any established criteria, but clearly rests on some notions of fairness and responsibility. The death of a young bungee jumper, for example, might be perceived as a terrible tragedy but not a mortal outrage, and the same logic is projected onto populations (Sharkey and Gillam 2010). If obesity is a lifestyle choice, for example, then the health problems of the corpulent are shrugged off as little more than the wages of gluttony. When seen as victims of a changing nutritional and physical environment over which they little control, however, the same disparities in illness and death take on the hue of a moral emergency (Brownell and Warner 2009). It is with a savvy eye for these gut reactions, therefore, that political conservatives often argue against being overly concerned by racial/ethnic health disparities. It is not discrimination that explains the premature morbidity and mortality of African Americans, argue neoconservative apologists like Sally Satel, because discrimination would be an injustice. Rather, African Americans simply have innate physiologic defects that make them more likely to become sick and die young (Satel 2001). Similarly, faced with overwhelming documentation of differential access to medical treatment by race, Satel argues that it is simply the case that African Americans choose to live in areas of the country with inferior medical care (Klick and Satel 2006). The implication that the disparity arises as a consequence of some volitional behavior, like the bungee-­jumping example, helps to reassign the blame and thus quell any sense of unfairness that might generate calls for redress.

58

Jay S. Kaufman

After group membership is assigned and disparities in some health outcome are noted and prioritized, then statistical analyses that attempt to “explain” these disparities generate even greater difficulties related to the choices and assumptions involved in causal inference, decomposition of effects, model specification, and the testing and interpretation of estimated model coefficients (Duan et al. 2008; Kaufman 2008). Descriptions of “high-­risk” populations necessarily invoke connotations of dangerous contagion or inherent defects that articulate with popular notions of racial attributes for many groups (Plant and Rushworth 1998). Associations between racial/ethnic group and disease create stigma, which for many outcomes will exacerbate the disparity in a treacherous feedback loop. This can lead to discrimination on the basis of group membership as well as stereotyping of the group members as unhealthy or inherently defected. One need only recall the blanket of suspicion cast over Haitians early in the HIV/AIDS epidemic (Frank et al. 1985), or the medical literatures on hypertension in Black Americans (Weir and Hanes 1996), or diabetes in Mexican Americans (Montoya 2011), all of which relentlessly portray the darker-­skinned body as essentially pathological. This stigmatization in turn leads to fatalism on the part of the labeled group. After all, if Native Americans are prone to alcoholism, then why fight it (Ehlers 2007)? If Australian Aborigines are destined to get diabetes, what’s the point of a healthy lifestyle (Paradies et al. 2007)? And yet in the fatalistic acquiescence to one’s stereotypical group health profile, others can find blame and resentment (Kim et al. 2010): “Why can’t those people take care of themselves?” Reification

Out of this cycle of risk foretold and disease bestowed, there arises an even more powerful consequence, which is the reificiation of the racial grouping as something natural and objective. If hypertension and diabetes are so endemic in certain ethnic groups, for example, doesn’t that just prove that the delineations are not historically arbitrary after all, but rather some important biological contours of human variation? Throughout the biomedical literature, the fact that pathology was seen to vary so markedly from one category to another was used to justify their continued use as classifications that were substantively important, and the ubiquitous presence of these categories in biomedical studies lent them the imprimatur of respectability as valid scientific objects (Kaufman and Hall 2003; Montoya 2007). Statistical tools are employed not only to link groups to biomedical outcomes but also to justify the groupings in the first place as some valid demarcation of human variability (Rosenberg et al. 2002). The use of “ancestry informative



Ethical Dilemmas in Statistical Practice 59

markers” (AIMs) in genetic studies promises to provide a quantification of continental ancestry, for example, which has been linked in an overly facile way with racial or ethnic ancestry (Bolnick et al. 2007). This creates a potentially fundamental tension with the autonomy of an individual to define themselves in terms of social affiliations that ultimately have nothing to do with biologic ancestry. One telling story of this disconnect is that of Wayne Joseph, a high school principal from Chino, California, who had a history of Black advocacy, for example writing essays on race for national news magazines. In a moment of idle curiosity about ten years ago, Joseph submitted a biological sample to be tested for AIMs in order to apportion his continental origins. The test results (apparently provided to him with no measures of precision or qualifications concerning error rates) arrived with the news that he was 57 percent Indo-­European, 39 percent Native American, 4 percent East Asian, and 0 percent African. Despite a lifetime of self-­identified Blackness, he was told that he lacked even the “single drop of blood” that is presumably the basis for America’s distinctive racial dichotomy (Kaplan 2003). After some reflection, however, Joseph realized that his tastes and affiliations had not changed at all, and that he was just as firmly self-­identified as a Black American in every meaningful sense as he always had been. He concluded that the gold standard for such a definition was not biologic ancestry, even if the test results were in some sense accurate (although they are necessarily arbitrary in a historical sense, since going back in time far enough would locate all human beings as 100 percent African). The broad ethical question that arises, therefore, is the extent to which statistics is the handmaiden of existing prejudices, the reflection and amplification of innumerable social conventions as a quantified object, thereby lending to such conventions a whiff of scientific legitimacy. The linking of racial/ethnic group with disease could conceivably yield important etiologic insights, but on the other hand, it might merely serve as a self-­fulfilling prophesy that bins people according to a capricious taxonomy and then promulgates this taxonomy on the grounds of its success in this very exercise. And yet similarly difficult issues come into play when attempting to demonstrate discrimination statistically, for example in the receipt of medical diagnoses or procedures (Blank et al. 2004). For the selectively skeptical analyst who wishes to doubt the existence of discrimination, there are always alternative explanations that can be invoked and potential statistical improprieties that can be decried (Heckman 1998). The practice of statistical inference is therefore remarkably malleable, a tool that most often fashions data on the anvil of our prejudices and presuppositions (Lewontin and Levins 2000). From this inherent subjectivity derives the obligation for ethical self-­examination.

60

Jay S. Kaufman

Prediction

The final example of statistical practice that has important (yet oft overlooked) ethical implications is the task of prediction. Many established statistical procedures that are based on investigator-­specified groupings, such as James-­Stein shrinkage and multilevel (hierarchical) modeling, promise ensemble advantages with respect to prediction error (Efron and Morris 1977). Indeed, under the Bayesian paradigm of “prior” knowledge for such groupings, the procedures necessarily require the admission of social facts into the analytic process. However, the way that the a priori groupings and the loss function are specified have important ramifications for the predicted values for individuals within the population, potentially changing how groups or individuals are described or treated, and therein lay the intriguing ethical implications of such maneuvers. As a simple example of this phenomena, consider performance on an academic test. Suppose that every tested student has a latent (that is, unobserved) score that is a function of their intrinsic talent and knowledge, but that this quantity is measured with random error on any one test that is administered. If I were to give a test to a group of students, therefore, the student with the very best score is most likely to have done somewhat better than her true latent potential, either by just happening to know more of the answers or by having been lucky in guessing. When next tested, this student is likely to do somewhat worse. Similar reasoning applies to the student who scores at the bottom of the class, who will be expected to do a bit better the next time she is tested. This phenomenon was described by Francis Galton in 1889 as “regression to the mean” and more recently was referred to by Howard Wainer as “Kelly’s Paradox” (2000). The innovation of the James-­Stein estimator, which has since served as the basis for the development of many statistical procedures including multilevel models, is that we can always do better if we “shrink” an observed value toward a group mean. For example, suppose there are two groups in the class, boys and girls. The boys have a mean verbal ability test score of 80, and the girls have a mean test score of 120. If we were to force a student to take many iterations of the test, we would be better able to estimate his or her true latent score, but most often we have only a single test score to work with. For a student who gets a single test score of 100, this is an unbiased estimate of their true latent score, but we can do better. If the student is a girl, we can bet that her true score will lie between 100 and 120. If the student is a boy, we can bet that his true score will lie between 80 and 100. Betting in this way, we will always outperform the predictions of someone using the observed scores alone without incorporating information on the gender of the student (Efron and Morris 1977).



Ethical Dilemmas in Statistical Practice 61

The ethical dilemma arises because the statistical procedure is agnostic about how units are categorized; we always do better in our prediction by “shrinking” toward the group mean, no matter what grouping is chosen. For example, the mean scores of boys and girls may be 80 and 120, but say that the mean scores for Blacks and Whites are also 80 and 120. One student, a Black girl named Mary, takes the test and scores 100 points. If we decide that the groups of interest are boys and girls, we would assign her a higher value than her achieved score, and in aggregate (for example, summing the squared error over all the students in the class), we are guaranteed to do better in our predictions; that is, we would accumulate a lower mean square error than if we had simply used the measured scores. On the other hand, if we decide that the groups of interest are Blacks and Whites, we would assign Mary a lower value than her achieved score, and in aggregate, we would still do better in our predictions; we would also accumulate a lower mean square error than if we had simply used the measured scores (figure 4.1). As analysts, we win either way, in the sense of reduced total (ensemble) error, and so choosing gender or race as the salient grouping is of little practical concern to us. But it makes a big difference to Mary, whose corrected score rises or falls depending on whether we happen to consider gender or race to be the most salient way to slice the population. The first reaction to such an example might be to discount it as too artificial, since we rarely reveal individual predicted scores back to the test takers when we are aiming for an ensemble advantage in prediction. But in a very real-­ world example of this sort of dilemma, Wainer (2000) recounts a program at the Educational Testing Service (ETS) in which high-­performing students from predominantly low-­income schools were labeled “strivers” and their standardized college admissions scores were indeed augmented, on the argument that they performed well even under adverse conditions. Because of the “regression to the mean” phenomenon, however, their scores might be better considered overestimates, which should be “shrunk” back to their group means. “It stands to reason,” wrote sociologist Nathan Glazer in 1999, “that a student from a materially and educationally impoverished environment who does fairly well on the [Scholastic Aptitude Test] and better than other students who come from a similar environment is probably stronger than the unadjusted score indicates” (Glazer 1999). Similar arguments were advanced by Anthony Carnevale, then ETS vice president. “When you look at a Striver who gets a score of 1000,” he said to the Wall Street Journal in 1999, “you’re looking at someone who really performs at 1200” (Marcus 1999). But Wainer argued instead for shrinkage to the group mean: “When you look at a Striver who gets a score of 1000, you’re probably

62

Jay S. Kaufman

prior

?

?

posterior

Figure 4.1   Mary Gets a Score of 100 on a Verbal Ability Test

Mary’s observed verbal ability score is 100, but she is a member of one group (girls) that has a mean score of 120 and a member of another group (Blacks) that has a mean score of 80. Whether we improve our prediction by increasing or decreasing her score therefore depends on which grouping we judge to be more salient.

looking at someone who really performs at 950. And, alas, a Striver is probably weaker than the unadjusted score indicates” (Wainer and Brown 2004). Whether one adjusts the score up or down depends therefore on the mental model one has of the meaning of group membership. If the low-­income school is a handicap like an extra weight that the student must bear, then the higher adjusted score is the better predictor of future performance when the weight is removed. On the other hand, if the low-­income school is the best representation of the aggregate tendency of students from this environment, then the lower adjusted score is the better predictor (Zaslavsky 2000). The choice of raising or lowering the score says more about us (the analysts) and our social assumptions than it does about the students and their true potential. This is the ethical component inherent in all such statistical analysis, because we, as researchers and analysts, decide which people belong together in a group and which trait is most salient as a basis for such categorizations. In a seemingly objective world of data, such choices are constantly made out of necessity and are always problematic. Since we tend to rely on our existing notions for what categorizations are most obvious to us, this represents yet another opportunity for science to be contaminated by our social prejudices and ideological traditions (Lewontin and Levins 2000). Conclusion

Using the case of race and ethnicity in biomedicine, I have described a few examples in this essay in an effort to demonstrate the general point that statistical analysis cannot be discussed without regard to some basic ethical principles.



Ethical Dilemmas in Statistical Practice 63

Quantitative techniques are widely held to be more “objective” than qualitative techniques, and yet human judgment and subjective decision making play a huge role in data collection, analysis, and interpretation. Indeed, in both major schools of statistical inference, greater reliance on subject matter knowledge is extolled as a virtue: Bayesians formalize this in terms of prior distributions on parameters, whereas frequentists promote the merits of deep subject matter knowledge in data exploration and model fitting (Tukey 1977). While the necessity of substantive knowledge is clear from both practical and philosophical positions, judgments made on the basis of such knowledge invoke beliefs, prejudices, and irrationalities from the social world that work their way inevitably and insidiously into our science. With subjective judgments, therefore, come ethical obligations to consider the humanity of research subjects and the effects that research has on them and on their communities. I have considered ways in which the handling of data on racial groups can consciously or unconsciously perpetuate injustices that exist in the social and political spheres. This is not a question of fraud or abuse, but rather acknowledgment that all analytic choices involve trade-­offs between conflicting values. When people are the units of observation, the implications of simple decisions, such as who is collapsed into the cell of a table, or what is the choice of the null in a heterogeneity test, will inevitably lead to winners and losers that reflect political power, stereotypes, or other social facts. Statistics, as a human endeavor, is therefore intrinsically linked to our social conventions and institutions, both reflecting and reinforcing them (Best 2002). Acknowledgments

This research was undertaken, in part, thanks to funding from the Canada Research Chairs program. Earlier incarnations of this paper were presented at “Conference on Unhealthy Professional Boundaries,” Goodenough College, London, UK, December 5, 2007, and “Mapping ‘Race’ and Inequality: Best Practices for Theorizing and Operationalizing ‘Race’ in Health Policy Research,” Albuquerque, NM, April 28–­29, 2011. Nicholas King provided helpful feedback on the manuscript. References American Statistical Association. 1983. “Ethical Guidelines for Statistical Practice: Historical Perspective. Report of the ASA Ad Hoc Committee on Professional Ethics; and Discussion.” American Statistician 37 (1): 1–­19. ———. 1999. “Ethical Guidelines for Statistical Practice. Prepared by the Committee on Professional Ethics and Approved by the Board of Directors, August 7, 1999.” Accessed June 26, 2012. http://www.amstat.org/committees/ethics/index.html.

64

Jay S. Kaufman

Asada, Y. 2005. “A Framework for Measuring Health Inequity.” Journal of Epidemiology and Community Health 59 (8): 700–­705. BBC News. 2008. “South Africa Chinese ‘Become Black.’” Accessed June 26, 2012. http:// news.bbc.co.uk/2/hi/africa/7461099.stm. Best, J. 2002. “People Count: The Social Construction of Statistics. Presentation at American Statistical Association, Joint Statistical Meetings, August 2002.” Accessed June 26, 2012. http://www.statlit.org/pdf/2002BestASA.pdf. Blank, R. M., M. Dabady, and C. F. Citro. 2004. Measuring Racial Discrimination. Washington, DC: National Academies Press. Bolnick, D. A., D. Fullwiley, T. Duster, R. S. Cooper, J. H. Fujimura, J. Kahn, J. S. Kaufman, J. Marks, A. Morning, A. Nelson, P. Ossorio, J. Reardon, S. M. Reverby, and K. TallBear. 2007. “Genetics: The Science and Business of Genetic Ancestry Testing.” Science 318 (5849): 399–­400. Breslau, J., S. Aguilar-­Gaxiola, K. S. Kendler, M. Su, D. Williams, and R. C. Kessler. 2006. “Specifying Race-­Ethnic Differences in Risk for Psychiatric Disorder in a USA National Sample.” Psychological Medicine 36: 57–­68. Brownell, K. D., and K. E. Warner. 2009. “The Perils of Ignoring History: Big Tobacco Played Dirty and Millions Died. How Similar Is Big Food?” Milbank Quarterly 87 (1): 259–­94. Dreger, A. D. 1998. “Ambiguous Sex—­Or Ambivalent Medicine? Ethical Issues in the Treatment of Intersexuality.” Hastings Center Report 28: 24–­35. Duan, N., X. L. Meng, J. Y. Lin, C. N. Chen, and M. Alegria. 2008. “Disparities in Defining Disparities: Statistical Conceptual Frameworks.” Statistics in Medicine 27 (20): 3941–­56. Efron, B. and C. Morris. 1977. “Stein’s Paradox in Statistics.” Scientific American 236 (5): 119–­27. Ehlers, C. L. 2007. “Variations in ADH and ALDH in Southwest California Indians.” Alcohol Research and Health 30 (1): 14–­17. Finucane, T. E., and J. A. Carrese. 1990. “Racial Bias in Presentation of Cases.” Journal of General Internal Medicine 5 (2): 120–­21. Frank, E., S. H. Weiss, J. C. Compas, J. Bienstock, J. Weber, A. Bodner, and S. H. Landesman. 1985. “AIDS in Haitian-­Americans: A Reassessment. Cancer Research 45 (9 Suppl.): 4619s–­20s. Garcia, R. 2003. “The Misuse of Race in Medical Diagnosis.” Chronicle of Higher Education 49: B15. Glazer, N. 1999. “Should the SAT Account for Race?—­Yes.” New Republic 221 (13): 26–­28. Gould, S. J. 1981. The Mismeasure of Man. New York: Norton. Hahn, R. A. 1999. “Why Race Is Differentially Classified on U.S. Birth and Infant Death Certificates: An Examination of Two Hypotheses.” Epidemiology 10: 108–­11. Harper S., N. B. King, S. C. Meersman, M. E. Reichman, N. Breen, and J. Lynch. 2010. “Implicit Value Judgments in the Measurement of Health Inequalities.” Milbank Quarterly 88 (1): 4–­29. Heckman, J. J. 1998. “Detecting Discrimination.” Journal of Economic Perspectives 12 (2): 101–­16. International Statistical Institute (ISI). 1985. “Declaration on Professional Ethics.” Accessed June 26, 2012. http://www.isi-web.org/about-isi/professional-ethics. Jowell, R. 1981. “A Professional Code for Statisticians? Some Ethical and Technical Conflicts.” Bulletin of the International Statistical Institute (Proceedings of the 43rd Session) 49 (1): 165–­209.



Ethical Dilemmas in Statistical Practice 65

Kaplan, E. A. 2003. “Black Like I Thought I Was.” LA Weekly. Accessed June 26, 2012. http://www.alternet.org/story/16917. Kaufman, J. S. 1999. “How Inconsistencies in Racial Classification Demystify the Race Construct in Public Health Statistics.” Epidemiology 10 (2): 101–­3. ———. 2008. “Epidemiologic Analysis of Racial/Ethnic Disparities: Some Fundamental Issues and a Cautionary Example.” Social Science and Medicine 66 (8): 1659–­69. Kaufman, J. S., and S. A. Hall. 2003. “The Slavery Hypertension Hypothesis: Dissemination and Appeal of a Modern Race Theory.” Epidemiology 14 (1): 111–­18. Kennedy, R. L. 2001. “Racial Passing.” Ohio State Law Journal 62: 1145–­93. Kim, A. E., S. Kumanyika, D. Shive, U. Igweatu, and S. H. Kim. 2010. “Coverage and Framing of Racial and Ethnic Health Disparities in U.S. Newspapers, 1996–­2005.” American Journal of Public Health 100 (Suppl. 1): S224–­31. Klick, J., and S. Satel. 2006. The Health Disparities Myth: Diagnosing the Treatment Gap. Washington, DC: American Enterprise Institute Press. Krieger, N., and D. R. Williams. 2001. “Changing to the 2000 Standard Million: Are Declining Racial/Ethnic and Socioeconomic Inequalities in Health Real Progress or Statistical Illusion?” American Journal of Public Health 91 (8): 1209–­13. Lewontin, R., and R. Levins. 2000. “Let the Numbers Speak.” International Journal of Health Services 3 (4): 873–­77. Lewontin, R. C., S. Rose, and L. Kamin. 1993. “IQ: The Rank Ordering of the World.” In The “Racial” Economy of Science, edited by S. Harding, 142–­60. Bloomington: Indiana University Press. Little, M. 1998. “Assignments of Meaning in Epidemiology.” Social Science and Medicine 47: 1135–­45. Loring, M., and B. Powell. 1988. “Gender, Race, and DSM-­III: A Study of the Objectivity of Psychiatric Diagnostic Behavior.” Journal of Health and Social Behavior 29 (1): 1–­22. Malcomson, S. L. 2000. One Drop of Blood: The American Misadventure of Race. New York: Farrar, Straus, and Giroux. Marcus, A. D. 1999. “To Spot Bias in SAT Questions, Test Maker Tests the Tests.” Wall Street Journal, B1. Marshall, G. A. 1993. “Racial Classifications: Popular and Scientific.” In The “Racial” Economy of Science, edited by S. Harding, 116–­27. Bloomington: Indiana University Press. Meeks, K. 2000. Driving While Black. New York: Broadway Books. Montoya, M. J. 2007. “Do Genes Explain Diabetes Health Disparities between Ethnic Groups?” Endocrine Today (online edition). http://www.healio.com/Endocrinology/ news/print/endocrine-today/ percent7B00295910-DDE1–4643-B3D7–3D3DC6D046C0 percent7D/Do-genes-explain-diabetes-health-disparities-between-ethnic-groups. ———. 2011. Making the Mexican Diabetic: Race, Science, and the Genetics of Inequality. Berkeley: University of California Press. Nobles, M. 2000. Shades of Citizenship: Race and the Census in Modern Politics. Stanford, CA: Stanford University Press. Paradies, Y. C., M. J. Montoya, and S. M. Fullerton. 2007. “Racialized Genetics and the Study of Complex Diseases: The Thrifty Genotype Revisited.” Perspectives in Biology and Medicine 50 (2): 203–­27. Peitzman, S. J. 2007. Dropsy, Dialysis, Transplant: A Short History of Failing Kidneys. Johns Hopkins Biographies of Disease. Baltimore: Johns Hopkins University Press.

66

Jay S. Kaufman

Plant, A. J., and R. L. Rushworth. 1998. “‘Death by Proxy’: Ethics and Classification in Epidemiology.” Social Science and Medicine 47 (9): 1147–­53. Rosenberg, N. A., J. K. Pritchard, J. L Weber, H. M. Cann, K. K. Kidd, L. A. Zhivotovsk, and M. W. Feldman. 2002. “Genetic Structure of Human Populations.” Science 298 (5602): 2381–­85. Rotimi, C. N. 2003. “Genetic Ancestry Tracing and the African Identity: A Double-­Edged Sword?” Developing World Bioethics 3 (2): 151–­58. Satel, S. 2001. “Medicine’s Race Problem.” Policy Review 110. Accessed June 26, 2012. http://www.hoover.org/publications/policyreview/3462856.html. Scanlon, J. P. 2006. “Can We Actually Measure Health Disparities?” Chance 19 (2): 47–­51. Sharkey, K., and L. Gillam. 2010. “Should Patients with Self-­Inflicted Illness Receive Lower Priority in Access to Healthcare Resources? Mapping out the Debate.” Journal of Medical Ethics 36 (11): 661–­65. Sox, H. C., M. A. Blatt, M. C. Higgins, K. I. Marton. 2007. Medical Decision Making. Philadelphia: American College of Physicians. Tiefer, L. 2006. “Female Sexual Dysfunction: A Case Study of Disease Mongering and Activist Resistance.” PLoS Medicine 3 (4): e178. Tukey, J. W. 1977. Exploratory Data Analysis. Reading, MA: Addison-­Wesley. Wainer, H. 2000. “Visual Revelations: Kelly’s Paradox.” Chance 13 (1): 47–­48. Wainer, H., and L. M. Brown. 2004. “Two Statistical Paradoxes in the Interpretation of Group Differences: Illustrated with Medical School Admission and Licensing Data. American Statistician 58 (2): 117–­23. Weir, M. R., and D. S. Hanes. 1996. “Hypertension in African-­Americans: A Paradigm of Metabolic Disarray.” Seminars in Nephrology 16: 102–­9. Zaslavsky, A. 2000. “On Kelly’s Paradox.” Chance 13 (3): 3–­4.

Chapter 5

John A. Garcia

A Holistic Alternative to Current Survey Research Approaches to Race

Social scientists have examined and used measures of race for much of the period marking the advent of empirical social science. The world’s largest social science data archive, the Inter-­University Consortium for Political and Social Research, maintains an archive of more than 70,000 data sets of research in the social sciences. Hosting sixteen specialized collections of data in education, aging, criminal justice, substance abuse, terrorism, and other fields, more than 19,000 of its data sets include at least one variable labeled as race. An initial review of these studies indicates two major patterns. First, virtually all of the data sets compiled prior to the mid-­1980s identified only two racial categories, Black or White (with a small proportion including the category “other”). After the mid-­1980s, many studies included more racial categories, such as Asian Americans, Native American and Pacific Islanders, and Latinos. The second pattern is the overwhelming reliance on an individual’s racial self-­ identification as the basis for racial categorization. My contention is that these typical approaches to measuring race in survey research are limited, both in terms of their conceptual basis and in terms of measures, and thus they warrant reformulation and replacement. Survey researchers should incorporate the premise that race is a complex, multidimensional construct that includes institutional racism and internalized racism and that has consequences for racial identification and health outcomes, as well as other important sociopolitical phenomena (Ford and Kelly 2005; Lin and Kelsey 2000). I advocate a holistic concept of race that reflects a complex set of relationships and social processes, all of which are set upon the foundation that race is a social category rather than a biological one (Williams 1997).

67

68

John A. Garcia

As a political scientist whose career has focused largely on survey research, my goal is to translate that holistic conception into practical tools for measuring race more effectively. I propose the creation of multidimensional indicators of race that match the major dimensions of the suggested holistic conception. Perhaps the most important and concrete implication of my contribution for survey researchers is the departure from the standard reliance on a single item (or perhaps two) to categorize survey respondents into racial groupings. At the same time, considerations of the efficiency and other costs of implementing an approach to measuring race require parsimonious survey items. For example, Sellers et al. developed a multilayered racial identity measure based on four major dimensions and using more than sixty items (1997). My goal is to develop an approach that establishes the key dimensions of race and then prudently develops items that could realistically be used in large-­scale, probability surveys. For the rest of this chapter, I develop the conceptual basis for a more holistic approach to race. Next, I provide a heuristic presentation of possible dimensions and discuss additional components of this reconceptualization. I end with a description of the linkages of multidimensional measures of race and ethnicity to health disparities research. Conceptualizing Race as Socially Constructed and Multidimensional

Researchers have begun to provide models that reconceptualize race and identify measures that capture its full scope (Bonilla-­Silva and Biaoccho 2001; Burton et al. 2010; Williams 1997). Historically, the concept of race has been tied to heredity, physical characteristics, and biological varieties to identify discrete populations (Williams 1997). While the primary emphasis has been on observable biological distinctions, race in fact implicates an individual’s social identity and how one negotiates that identity within different environments. These subjective attachments and common historical experiences have served to establish social-­cultural boundaries, which can result in issues of equity, exclusion, and differential outcomes (James 2001). The processes of race category creation, racialization, and racism are dynamic ones in which individuals and groups negotiate their identities and interests when resources and opportunities are limited. Importantly, White racial group members have historically been the most active in negotiating resources and privileges tied to their race and excluding non-­Whites from those resources and privileges. For the most part, race today is viewed as a social construct such that racial categorization can change over time (LaVeist 1994), thus the need to re-­conceptualize race as a social experience that is both dynamic and situational (Espiritu 1992; Omi and Winant 1994). This dynamism and fluidity is



A Holistic Alternative 69

understood better by linking it to social processes. Angela James (2001) illustrates this perspective by differentiating those who use race in their research as opposed to those who study race. In the case of the former, the perspective tends to treat race as primordial or as a fixed characteristic that can be measured as a dichotomous control variable (James 2001; Zuberi 2001). In the case of the latter, a broadly grounded conceptualization of race places it within a social and historical context connected to the effects of racial status (James 2001). The emphasis on reconceptualizing and measuring race more comprehensively has a direct bearing on how researchers apply a multidimensional construct of race for analysis and interpreting social justice, health disparities, and inequities. In short, we must move away from the treatment of race as a “fixed characteristic” or as a “demographic, control variable” (James 2001, 245–­46). Broadening the scope and including the relational nature of the race concept adds a more socially and historically based perspective that emphasizes the sociopolitical underpinnings of race. My proposed conceptualization of race has the following eight components. First, race is a social construct based on social and political context, rather than any essential biological difference between groups. Second, individuals have considerable agency in placing themselves within racial categories (for example, racial identity). Third, racial self-­identification is a cognitive dimension of one’s self-­concept and is a developmental process (see Iwamoto et al. and Helms and Mereish, this volume). Fourth, the individual does not “choose” racial identity or group membership in isolation and, in fact, is heavily influenced by a host of externalities including social interactions, historical context and patterns, legal status and constraints, and other factors too numerous to list here. Fifth, race is dynamic, meaning that the understanding and/or expression of one’s race can change over time. Sixth, race is an element of multiple social identities in which the “constellation” of self includes race among other salient social identities. Seventh, physical features (phenotypical characteristics) serve as a basis for racial classification, both by the individual and especially in terms of how others identify her or him. Finally, the “separate but related” (Garcia 2009) notions of race and ethnicity require more focused examination. Table 5.1 presents a heuristic tool to illustrate my proposed indicators of holistic race that move us beyond a single classification measure. Our discussion and analysis of the concept of race has emphasized its complexities. It has relational, structural, dynamic, experiential, and individually based determinants. My chart identifies variables that need to be part of a multidimensional set of indicators to capture race. This examination takes the position that a broader construction of race can enable more extensive analysis of the

70

John A. Garcia

consequences of racial status and categories and their meaning, as well as how race is connected to individuals’ or groups’ experiences, status, and opportunity structures. In a similar vein, the theoretical nature of disparities or social inequality demands clarity and coherency. Focusing greater attention to reconceptualizing race has been discussed in this policy area to encourage the development of measures and analytical techniques that “capture” more complex analyses and interpretation of race and inequality. The recognition of race as socially constructed has been accompanied, over the past four decades, by the transition from other-­ defined to self-­ identified racial categorization. One direct consequence for the understanding of race is that an individual’s selects her/his racial group, therefore making self-­ identification the “gold standard” for racial classification. Based upon the concepts of self-­and social identity (Cerulo 1997; Owens, Robinson, and Smith-­Lovin 2010), individuals “place themselves” within established racial categories based on criteria the individual chooses. Self-­identification is “determined” by any range of factors. For example, cognitive research conducted by the Census Bureau conducts follow-­up interviews after a person has filled out

Table 5.1   Heuristic Chart of Holistic Race Measures Source of Variables

Individual Perspective

Group Perspective

Response

Racial identification

Respondent (R) self-­ identification

Respondent’s perception how others classify him/her racially

Respondent

Skin tone

5-­point light to dark scale based on self-­ report

R places herself on 5-­point scale based upon perceptions of how others see R

Observer, face-­to face, places R on 5-­point scale

Saliency of racial identification

Likert-­type scale indicating importance of race relative to other identities

Discrimination: perceived and experiential

R reports of both perception, actual experience with discrimination

Perceptions for own group members

Respondent

Multiracial status

Range of racial identities selected

Ask R how others view her racially; more than one race option

Respondent

Structural racism

Institutional, cultural, and organizational

Access, equity rights accorded due to group membership

Respondent



A Holistic Alternative 71

the census short form (de la Puente and McKay 1995). Individuals were asked why they selected a specific racial category on the form. On many occasions, their answers reflect notions of physical features and/or colorism, ancestry, societal cues, public policies, and other external cues (de la Puente and McKay 1995). With the reliance on preset categories, the underlying meaning of an individual’s racial identification is hard to know, but the end result is a count of how many persons fall within a racial category. Latinos complicate this process more than other groups because a substantial portion of this community does not place itself into the current racial categories of the Office of Management and Budget Directive 15 (OMB 1997). The revised standard has five minimum categories for data on race: American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Other Pacific Islander, and White. There are two categories for data on ethnicity: “Hispanic or Latino” and “Not Hispanic or Latino.” Almost two-­fifths (36.7 percent) of Latinos marked the “some other race” option in the 2010 decennial census (Humes et al. 2011, 6). Interpretations cover the gamut of explanations from respondents’ confusion as to the meaning and use of race in the United States to a desire to categorize themselves as a distinctive race in this country. The contextual and situational nature of race can generate a response based upon homeland experiences. That is, if the race question were to be asked outside the United States, the understanding and descriptions of racial options would be different and vary across geopolitical borders. Many social science surveys separate race and ethnicity as two distinctive concepts, while others combine race and ethnicity as configuring social groups and/or racialized groups. The 1995 Current Population Survey had a race and ethnicity supplement section in which respondents were asked their understanding of race, ethnicity, and national origin. There was significant confusion among the respondents as whether race and ethnicity were different or interchangeable (McKay et al. 1996). In cognitive testing conducted by the Census Bureau, some foreign-­born interviewees would comment about the race question by indicating that they would have different responses if they were answering the question in the home country or as a resident of the United States. For example, a Peruvian-­born individual indicated that his answers would be different if he was still living in Peru, as opposed to having lived in the United States for over ten years (Bates et al. 2006). That is, the prevailing Black-­White racial paradigm would guide his response after many years in the United States, while a Peruvian-­based response would be more descriptive of skin tone and indigenous origins.

72

John A. Garcia

Institutional Racism and Discrimination

An additional dimension of race that deserves to be accurately and effectively measured in surveys has to do with its inextricable link to ideas about differential treatment, unequal status, and differential power relations. For example, in the context of health disparities research, race is viewed as a complex and multidimensional construct that includes institutional racism and internalized racism, and affects health outcomes and usage, as well as other important sociopolitical phenomena (Ford and Kelly 2005; Lin and Kelsey 2000). In short, we cannot talk about race without talking about discrimination and racism. In the field of political science, a variety of measures have linked racial identity to experiences of discrimination and racism. These include the concepts of group affiliation, affinity, and “linked fate” that political scientists believe serve to heighten racial group membership and affect life chances, health outcomes, and the like (Braveman 2012; Camara et al. 2011; Chae et al. 2012; Fiscella et al. 2002). Experiences with discriminatory behaviors and attitudes as well as perceived discrimination toward one’s racial group have been shown to vary significantly with White versus non-­White racial status (Williams et al. 2010). From a measurement perspective, an important issue is that the primary measures have been developed relative to African Americans and only later applied to other groups (for an example of how this plays out with regard to the racial identity development model, see Iwamoto et al., this volume). Questions about the validity and reliability of measures across groups come into consideration. For the most part, the items used have been either a “one-­or two-­stage” protocol (Shariff-­Marco et al. 2011). The one-­stage item places race and ethnicity as the basis for experiencing unfair treatment or discrimination. The two-­stage items ask about any experiences with discrimination, then follows up as to the respondent’s understanding about the bases of such behavior. If race/ethnicity is seen as the primary factor, then this is evidence of racial discrimination. The structure of the items affects the level of reporting of discrimination and indication of multiple acts of discrimination (for example, unfair treatment in public places, stereotypes, harassment, and so forth). They serve to differentiate the types of behaviors and attitudes more common to a racial group. For example, Latinos might be more prone to cite discriminatory behaviors/attitudes regarding having an accent and not being viewed as smart as others; while African Americans might cite experiencing poor treatment in public spaces and experiencing reactions of fear on the part of others (Shariff-­ Marco et al. 2011).



A Holistic Alternative 73

The construction of items that tap discrimination can be affected by the wording used, the structure, and how one interprets discriminatory experiences filtered by race (Ahmed and Williams 2007). Integrating discrimination as one of the elements that capture race requires multidimensional measures, as well as indicators that are valid and reliable across racial groups. The other important element regarding discrimination is the distinction between perceived versus actual experience. Some research indicates that the linkage between these two aspects is modest and each produces independent effects onto attitudes, behaviors, and health status (Borrell et al. 2006; Williams 1999). Some studies have sought to move to a more direct measurement of racial subordination as power relations or institutional racism. That is, they wish to find a way to tap into the view that the underlying basis for the existence of racial categories stems from power, control, status, access, and inequality. Structural racism represents a macro-­level system, social forces, and institutions, ideologies, and processes that interact with one another to generate and reinforce inequities among racial/ethnic groups. Gee and Ford (2011) and Harrell et al. (2010) identify four forms of racism: structural (for example, poverty, underemployment); cultural (for example, devaluing non-­dominant cultures); institutional (for example, institutions of mobility like schools, labor markets), and individual (for example, discrimination and prejudice). The traditional stress and coping model holds that racism constitutes a source of aversive experience that leads to poor health outcomes (for example, see Geronimus and Helms and Mereish, this volume). Current evidence points to psycho-­physiological pathways linking facets of racist environments with physiological reactions that contribute to disease (Harrell et al. 2010). The alternative pathways emphasize prenatal experiences, subcortical emotional neural circuits, conscious and preconscious emotion regulation, controlled repetition or continuation of cognitions, and negative affective states stemming from racist cognitive schemata. The structural forms of racism and their relationship to health inequities remain under-­studied (Gee and Ford 2011). The work of Gee and Ford isolates three domains from which structural racism operates. They include social segregation (primarily residential segregation) and its outcomes (for example, quality of schools, access to labor markets); immigration policy (for example, bases for admission, criteria for citizenship, selected punitive/ restrictive policies); and “intergenerational drag” (for example, the persistence over time of limited or downward intergenerational mobility). Analytically, the inclusion of structural racism strongly suggests the use of multilevel modeling to incorporate the different individually relevant factors as well as structural/ contextual components.

74

John A. Garcia

Race as a Dynamic Concept

Another noteworthy aspect about the multidimensionality of race is the dynamic nature of racial identification. When surveys include an item(s) tapping race, it is assumed, generally, that the response is a permanent one. That is, the response given is the one that would have been given in the past and will be in the future. Research on early childhood social identity demonstrates how social identities, especially racial identification, occur at the preschool or early elementary grade levels (Azmitia et al. 2008). The developmental process among children and youth indicates changes and how different experiences and perspectives affect one’s racial identification. For example, work on dissimilation by Hayes-­Bautista illustrates how institutional practices (for example, affirmative action policies) can cause individuals to reevaluate their self-­concept and incorporate more of a racial/ethnic identity as a result (Hayes-­ Bautista 1974). Consider how the wave of anti-­immigrant public opinion and legislative actions in recent years, along with massive public protests, might motivate native-­born Latinos who did not previously have a strong sense of identity to incorporate a sense of group affiliation with their foreign-­ born “counterparts” that accentuates their ancestral origin (Barreto et al. 2009). As a result, a changing, situational racial self-­identification and identification by others must be assumed and integrated into the conception and measurement of race. Multiracial Identity

By relying on self-­identification as the basis for racial classification, a singular response is an expected reply. Yet post-­2000 federal data allows the expression of race/ethnicity in multiple racial categories (OMB 1997). This option reinforces the notion that race and/or ethnicity comprise a “constellation” of social identities in which centrality or salience may separate one social identity from others. This sense of multiple racial identities is compounded with other social identities (for example, gender, religion, national origin, and so on) that can comprise a multiplicity of coexisting identities. For biracial and multiracial persons, their selections can vary by situation/circumstance, especially for younger people (Entwisle and Astone. 1994; Lopez 2003). The overall point is that the formation and expression of racial identity and identities is a function of socialization, experiences, and historical legacies. Individuals who embrace multiple racial identities present a methodological complication. For the most part, the racial self-­identification approach prevalent in survey research “allows” the respondent to indicate the number



A Holistic Alternative 75

of racial categories he/she identifies with. Usually, the individual will then specify their primary racial identification. From the 2000 and 2010 censuses, frequency of reporting “multiracial status” is relatively uncommon (from 2.4 percent to 2.9 percent). Yet the basis for indicating such a “racial background” depends, in part, on both knowledge and internalization of such multiple racial origins. Research by Woo et al. (2011) provides three alternative ways to measure multiracial status. They are mixed ancestry, self-­ identification as multiracial, and socially assigned multiracial status. In the first case, knowledge of or information about the racial background of the person’s parents is used to indicate mixed ancestry or mixed race. While multiracial self-­identification is an alignment with racial identity, it also entails incorporating an identity that emphasizes multiple origins rather than a single racial category. It is not uncommon in health surveys to allow for multiple racial responses followed by a question that asks which racial category “best” describes you the respondent. Miller and Willson (n.d.) raise important questions about the latter question and interpreting responses to it. The central problem is that the of the respondent’s criterion for identifying the “best” category could be based upon any of the following: how she thinks others perceive her, her sense of cultural belonging, her perception of what category is most appropriate for administrative records, or her estimate of which racial category is dominant in the sense of mathematical ancestry. Miller and Willson suggest that for the identification of “primary racial identity, a context be indicated for multi-­racial respondents. That is, ‘best’ applies to: a) group you belong to; b) how identified by other people; c) identity for official purposes; and d) in terms of your ‘blood relatives.’” Their intent is to reduce response error, item non-­response, and minimize the respondent’s “burden” of interpreting the intent of the question. This illustration expands our understanding of race and its origins that extends beyond the individual’s internal cognitive domain. Finally, the last dimension of Woo et al. (2011) is the public or external basis for multi-­racialness. The respondent is asked to indicate “how others would classify you in the U.S.” This recognizes the “societal” or structural basis for “classifying” persons into racial groups even though the individual may not see him or herself in the same manner. The extant research identifies a variety of factors that affect a multiracial identity (Bratter et al. 2011; Brunsma 2005). They include social class, social networks and their composition, the nexus of race and gender, phenotype and/or appearances, and parent’s racial background. The context of the home and school environments and the respondent’s experience with discrimination come into play as well. The

76

John A. Garcia

incorporation of multi-­racialness as part of the general concept of race captures the breadth of the meaning of race and the range of factors that influence the categorization of persons into racial groupings. Two underlying themes are the multidimensional nature of this concept and the relational impact of characterizing oneself as multiracial. Race and Phenotype

Issues of racial self-­identification and how others perceive one’s race also relate to phenotype or the appearance that each person has. Any discussion of race involves colorism and the role of physical appearances in categorizing racial groups. The White-­over-­Black paradigm that has defined U.S. race relations is anchored in physical characteristics, especially skin color. The infamous one-­ drop rule established sociopolitical and cultural boundaries between Whites and African Americans (Davis 2001). Not only were opportunity structures affected, as well as stereotypes, but internal (within group) differentiation was practiced between lighter and darker skinned persons (Allen et al. 2000; Murguía and Telles 1996; Tafoya 2004). Today, the changing demographics of the United States have taken research on colorism—­discrimination or privilege based on skin color within a racial group—­well beyond variations among Blacks. Studies over the last decade have placed greater emphasis of the continuing significance of skin color as marker of status, stratification, and racial distinctions (Gomez 2000; Harris and Sim 2002; Hochschild, Weaver, and Burch 2011). The examination of skin color has been extended beyond the African American community and indicators of distinctive phenotypical characteristics now include Latinos, Asian Americans, and Native Americans. While skin color (in terms of darker versus lighter skin tones) dominates the analytical attention, other traits such as indigenous versus “European” features, shape of eyes and the elliptical fold, hair texture, and skin bleaching represent different markers that are associated with race. The inclusion of skin color as an indicator of race produces measurement challenges that have been addressed in several different ways. Most commonly, the respondent is placed in a skin tone category (ranging from lighter to darker skin ratings) based on her or his self-­report or the report of an external observer. Obviously, the respondent’s basis for such self-­assignment (as well as the third party’s) will be based on the individual’s experiences, self-­concept, social identity/identities, social networks, as well as other external influences. For the most part, a five-­point scale is used to place him/herself on a light to dark continuum. Social meaning and value place more positive value on being lighter, so the full range of the scale is seldom used; instead,



A Holistic Alternative 77

the distribution is skewed toward the “light” end of the scale (Garcia 1991). A related approach has been the interviewer’s placement of the respondent on the skin color scale. This has resulted in “race-­of-­interviewer effect” in which interviewers perceive greater variations in the skin tones of same-­race respondents than among other-­ race respondents (for example, Black interviewers categorized White respondents as lighter than their White interviewer counterparts and vice versa; Hill 2002). More recently, Chan et al. (2004) has included both self-­reported skin “phenotype” as well as readings from a narrow-­band reflectance spectrometer. The latter has greater precision of differentiating skin pigmentation and a more reliable measure. The primary purpose of this approach was to make “precise” measures of skin color to be used in the evaluation of patients and expand the clinical information collected (Chan et al. 2004). However, the accuracy and precision of this measurement tool does not capture the social meaning, status, and context in which skin color and race are interconnected. Another method to measure skin color lies in the response how others would place the respondent on a skin color scale. Perceptions of others’ classifications represent the sociopolitical and structural underpinnings of the complexity and multidimensionality of race. While some researchers (Hochschild and Weaver 2007; Hochschild et al. 2011; Prewitt 2005) have proposed the use of skin color over the traditional category of race, there is greater consensus that a broader conceptualization of race needs to include the phenotypic measure of skin tone/color (LaVeist 1994; Williams 1997; Woo et al. 2011). Untangling Race and Ethnicity

An integral part of this examination of race and treating it as a more complex, holistic concept is the relationship between race and the concept of ethnicity. In previous work, I have used the descriptor of “separate but related” to portray the complex relationship between race and ethnicity (Garcia 2009). For the most part, the critical components of ethnicity include ancestry, cultural traditions, religious affiliation/beliefs, language, and national origin. Race is differentiated by an emphasis on physical characteristics or phenotypical traits. At the same time, the connections of both of these concepts with social status, institutional arrangements, power and influence, and mobility have the effects of enjoining these concepts. Race has been distinguishable from ethnicity due to the perceived intransigence and centrality of race in daily life. Ethnicity is perceived as more ephemeral or evolving to a more symbolic function, even though there is growing evidence of the persistence of ethnicity across

78

John A. Garcia

generations parallels the saliency and function of race (Arrighi 2007; Hans and Martinez 1994; Telles and Ortiz 2009). Another dimension of the ethnicity issue is the manner in which contemporary national origin and religious immigrant populations are being racialized. As a result, racial attributes, prejudice, and discriminatory behaviors take on patterns associated with racism. This process impacts immigrants, especially Latino immigrants, by being stigmatized into a low status associated with U.S. racial hierarchies. The sense of being different and/or a sense of “otherness” are a reproduction of inequality and affect access to better life opportunities (Viruell-­Fuentes 2011; Viruell-­Fuentes et al. 2012). Viruell-­Fuentes’s research explores lines of inquiry regarding immigrants’ experiences with day-­ to-­day discrimination as well as the role place and immigration policies have in shaping immigrant health outcomes. Another contemporary example is the post-­9/11 racialization of Muslim Americans and Middle Eastern immigrants and their descendants in the United States (Senzai 2012). While these individuals can be categorized into existing racial groupings, immigrants as a class of people have also become racialized. Considerations of national origin, language, phenotype, and culture are important aspects to incorporate into any analysis involving social status and race. The concepts of minority groups, disadvantaged populations, and marginalized groups are intertwined with the notions of race and ethnicity. This intersection acknowledges the relational and dynamic dimensions of the concept of race. Status, powerlessness, sociopolitical circumstances, and socioeconomic intergenerational mobility/stagnancy are also associated with racial categories and structural/institutional relations. It has been suggested that culture is relevant to conceptualizing race (and it is taken for granted that culture characterizes ethnicity), yet it remains relatively unexamined empirically how we might explore the intersection of culture with conventional notions of race. Culture relates to beliefs and behaviors (religious, folk, acculturation, migration experiences, and so on). How one incorporates the concept and measures of culture remains a serious challenge for researchers. For example, Ulmer et al. (2009) suggested inclusion of ethnic information such as language use, proficiency, and primary language of communication as well as granular ethnicity (for example, ancestry and national origins) to be part of its racial protocol (for example, items for a battery on race). Implications for Health Disparities Research

In the public health context, there is pressure on the public health industry and researchers to develop consistent and useful approaches to racial and ethnic



A Holistic Alternative 79

categories. Some researchers (James 2001; Krysan and Lewis 2004; LaVeist 1994) have been critical of the extent of “thought and care” that is needed before data habitually categorized by race is included as variables in research (Sheldon and Parker 1992). Sociologists like Bonilla-­Silva and Biaoccho (2005), Hirschman et al. (2000), and Anderson and Feinberg (2000) have examined the changing bases for racial classification by governmental agencies and social institutions. Part of that discussion lies with shifting American population demographics and official policies that establish racial/ethnic categories, as well as the ways in which individuals identify themselves. Other chapters in this collection relate race to matters of social justice, inequalities, health disparities, and important considerations for specific racial/ethnic groups. These represent the relational nature of race/ethnicity and sociopolitical and health consequences. The impact of race and ethnicity along with language, culture, and socioeconomic status on health status, behaviors, morbidity, and so forth has been linked with health disparities (Braveman 2012). Continuing work in addressing disparities not only requires the collection and use of data on race, ethnicity, and language in health care data systems (Ver Plog and Perrin 2004) but also the design of measures that more fully capture the concept of race. Such multidimensional measures can provide more opportunities to monitor and analyze disparities (for example, different aspects of health) as well as “sorting out” the magnitude of the elements of race affecting health-­related variables. There is considerable interest in understanding group differences and their causes, which calls for the availability and quality of individual-­level data on race, ethnicity, socioeconomic status, acculturation, and language (for example, language use, place of birth, generation status). These kinds of data are critical to documenting the nature of disparities in health care and to developing strategies to eliminate disparities (Ulmer et al. 2009). The collection of data on race, ethnicity, and language will, in principle, have the greatest impact if it is done according to standards that allow for comparison of data across organizations, sharing of individual-­level data from one organization to another, and combining of data from multiple sources (Ver Pleog and Perrin 2004). My emphasis on incorporating race measures parallels other researchers’ calls for recognizing the complexities and relational nature of race and the need for the development of comprehensive measures that reflect these dimensions. For example, socioeconomic status serves as an important mediator of racial and ethnic disparities and a further source of disparities (Braveman 2012; Ulmer et al. 2009). Acculturation (and its proxy measures language, place of birth, years in the United States, and generational status)

80

John A. Garcia

is also related to health status. Mismatches between the language spoken by health care providers and by patients can be a limiting factor in health care interactions and health information exchange. This latter point highlights my earlier comments regarding the intersection of race and ethnicity. Additionally, collection and retention for analysis of specific multiple-­race combinations rather only offering the more general category of “multiracial” is critical (Lopez 2003; Ulmer et al. 2009). Another dimension is evaluating how to include language as part of the race/ethnicity milieu. For example, Ulmer et al. (2009) conclude that spoken language needs can best be assessed by asking two questions: one aimed at determining whether an individual speaks English less than very well and a second aimed at identifying the individual’s preferred spoken language during a health care encounter. In any event, capturing the breadth and scope of race presents serious challenges for researchers in terms of reconceptualization and measurement, but it also provides added opportunities to understand health disparities. References Ahmed, A., S. Mohammed, and D. R. Williams. 2007. “Racial Discrimination and Health: Pathways and Evidence.” Indian Journal of Medical Research 126: 21–­30. Allen, Walter, Edward E. Telles, and Margaret Hunter. 2000. “Skin Color, Income, and Education: A Comparison of African Americans and Mexican Americans.” National Journal of Sociology 12 (1): 129–­80. Anderson, Margo, and Steven Fienberg. 2000. “Race and Ethnicity and the Controversy Over the U.S. Census.” Current Sociology 48 (3): 87–­110. Arrighi, Barbara, ed. 2007. Understanding Inequality: The Intersection of Race/Ethnicity, Class, and Gender. Lanham, MD: Rowman and Littlefield. Azmitia, M., M. Syed, and K. Radmacher. 2008. “On the Intersection of Personal and Social Identities: Introduction and Evidence from a Longitudinal Study of Emerging Adults.” In The Intersections of Personal and Social Identities: New Directions for Child and Adolescent Development, ed. M. Azmitia, M. Syed, and K. Radmacher, 120, 1–­16. New York: John Wiley and Sons. Barreto, Matt, Sylvia Manzano, Ricardo Ramirez, and Kathy Rim. 2009. “Mobilization, Participation, and Solidaridad: Latino Participation in the 2006 Immigration Protest Rallies.” Urban Affairs Review 44: 736–­64. Bates, Nancy, Elizabeth A. Martin, Theresa J. DeMaio, and Manuel de la Puente. 2006. “Questionnaire Effects on Measurements of Race and Spanish Origin (Research Report Series-­Survey Methodology # 2006–­12).” Washington, DC: U.S. Census Bureau. Bonilla-­Silva, Eduardo. 1999. “The Essential Social Fact of Race.” American Sociological Review 64 (6): 899–­906. Bonilla-­Silva, Eduardo, and Gianpalo Biaoccho. 2008 “Anything but Racism: How Sociologists Limit the Significance of Racism” in White Logic, White Methods: Racism and Methodology, ed. T. Zuberi and Eduardo Bonilla-­Silva. Lantham, MD: Rowman and Littlefield.



A Holistic Alternative 81

Borrell, L. N., C. I. Kiefe, D. R. Williams, A. V. Diez-­Roux, and P. Gordon-­Larsen. 2006. “Self-­Reported Health, Perceived Racial Discrimination, and Skin Color in African Americans in the CARDIA Study.” Social Science and Medicine 63: 1415–­27. Bratter, Jenifer L., and Bridget K. Gorman. 2011. “Does Multiracial Matter? A Study of Racial Disparities in Self-­Rated Health.” Demography 48: 127–­52. Braveman, Paula. 2012. “Commentary: Health Inequalities by Class and Race in the U.S.: What Can We Learn from the Patterns?” Social Science and Medicine 74: 665–­67. Brunsma, David L. 2005. “Interracial Families and the Racial Identification of Mixed-­ Race Children: Evidence from the Early Childhood Longitudinal Study.” Social Forces 84 (2): 1131–­57. Burton, Linda, Eduardo Bonilla-­Silva, Victor Ray, Rose Bucklew, and Elizabeth Hordge Freeman. 2010. “Critical Race Theories, Colorism, and the Decade’s Research on Families of Color.” Journal of Marriage and Family 72 (June): 440–­59. Camara, Jules P. Harrell, Tanisha I. Burford, Brandi N. Cage, Travette McNair Nelson, Sheronda Shearon, Adrian Thompson, and Steven Green. 2011. “Multiple Pathways Linking Racism to Health Outcomes.” Du Bois Review 8 (1): 143–­57. Cerulo, Karen A. 1997. “Identity Construction: New Issues, New Directions.” Annual Review of Sociology 23: 385–­409. Chae, David H., Amani M. Nuru-­ Jeter, Karen D. Lincoln, and Darlene D. Francis. 2011. “Conceptualizing Racial Disparities in Health Advancement of a Socio-­ Psychobiological Approach.” Du Bois Review 8 (1): 63–­77. Chae, David H., Amani M. Nuru-­Jeter, Karen D. Lincoln, and Kimberly R. Jacob Arriola. 2012. “Racial Discrimination, Mood Disorders, and Cardiovascular Disease among Black Americans.” Annals of Epidemiology 22: 104–­11. Chan, Joanna, Allison Ehrlich, Reva Lawrence, Alan Moshell, Maria Turner, and Alexa Kimball. 2004. “Assessing the Role of Race in Quantitative Measures of Skin Pigmentation and Clinical Assessments of Photosensitivity.” Journal of American Academy of Dermatology 52 (4): 609–­15. Davis, F. James. 2001. Who Is Black? One Nation’s Definition. 10th anniversary edition. University Park: Penn State University Press. De la Puente, Manuel, and Ruth McKay. 1995. “Developing and Testing Race and Ethnic Origin Questions for the Current Population Survey Supplement on Race and Ethnic Origin.” Proceedings of the Section on Survey Research Methods. Alexandria, VA: American Statistical Association. Entwisle, Doris R. and Astone, Nan. 1994. “Some Practical Guidelines for Measuring Children’s Race/Ethnicity and Socioeconomic Status.” Child Development 65: 1521–­40. Espiritu, Y. L. 1992. Asian American Pan-­Ethnicity: Bridging Institutions and Identities. Philadelphia: Temple University Press. Fiscella, Kevin, Peter Franks, Mark P. Doescher, and Barry G. Saver. 2002. “Disparities in Health Care by Race, Ethnicity, and Language among the Insured: Findings from a National Sample.” Medical Care 40 (1): 52–­59. Ford, Marvella, and P. Adam Kelly. 2005. “Conceptualizing and Categorizing Race and Ethnicity in Health Services Research.” HSR: Health Services Research 40 (5, part 2): 1658–­75. Garcia, John A. 1991. “Notes from the Training Session of Interviewers for the Latino National Political Survey.” Philadelphia, PA. ———. 2009. “Examining Notions of Race among Latinos and the Intersection of Culture and Language.” Pluralism in the Americas Conference. Bielefeld University, Bielefeld, Germany.

82

John A. Garcia

Gee, Gilbert C., and Chandra L. Ford. 2011. “Structural Racism and Health Inequities: Old Issues, New Directions.” Du Bois Review 8 (1): 115–­32. Gomez, Cristina. 2000. “The Continual Significance of Skin Color: An Exploratory Study of Latinos in the Northeast.” Hispanic Journal of Behavioral Sciences 22: 94–­103. Hans, Valerie P., and Ramiro Martinez Jr. 1994. “Intersections of Race, Ethnicity, and the Law.” Cornell Law Faculty Publications 18 (3): 211–­21. Harrell, Camara Jules P., Tanisha I. Burford, Brandi N. Cage, Travette McNair Nelson, Sheronda Shearon, Adrian Thompson, and Steven Green. 2010. “Multiple Pathways Linking Racism to Health Outcomes.” Du Bois Review 8 (1): 143–­57. Harris, David, and Jeremiah Sim. 2002. “Who Is Multi-­Racial? Assessing the Complexity of Lived Race.” American Sociological Review 67 (4): 614–­27. Hayes-­Bautista, David Emmett. 1974. “Becoming Chicano: A Dis-­assimilation Theory of Transformation of Ethnic Identity.” Ph.D. diss., University of California, San Francisco. Hill, Mark. 2002. “Race of the Interviewer and Perception of Skin Color: Evidence from the Multi-­ City Study of Urban Inequality.” American Sociological Review 66 (1): 99–­108. Hirschman, Charles, Richard Alba, and Reynolds Farley. 2000. “The Meaning and Measurement of Race in the U.S. Census: Glimpses into the Future.” Demography 37 (3): 381–­93. Hochschild, Jennifer, and Vesla Weaver. 2007. “The Skin Color Paradox and the American Racial Order.” Social Forces 86 (2): 643–­70. Hochschild, Jennifer L., Vesla M. Weaver, and Traci Burch. 2011. “Destabilizing the American Racial Order.” Dædalus: The Journal of the American Academy of Arts and Sciences 140 (2). Humes, Karen, Nicholas Jones, and Roberto Ramirez. 2011. “Overview of Race and Hispanic Origin (C2010BR-­02).” Washington, DC: U.S. Bureau of the Census. James, Angela. 2001. “Making Sense of Race and Racial Classification.” Race and Society 4: 235–­47. Krysan, M., and A. Lewis. 2004. The Changing Terrain of Race and Ethnicity. New York: Russell Sage. LaVeist, Thomas. 1994. “Beyond Dummy Variables and Sample Selection: What Health Services Researchers Ought to Know about Race as a Variable.” Health Services Report 29 (1): 1–­16. Lin, Scarlett, and Jennifer Kelsey. 2000. “Use of Race and Ethnicity in Epidemiologic Research: Concepts, Methodological Issues, and Suggestions for Research.” Epidemiologic Reviews 22 (2): 187–­202. Lopez, Alejandra M. 2003. “Collecting and Tabulating Race/Ethnicity Data with Diverse and Mixed Heritage Populations: A Case-­Study with U.S. High School Students.” Ethnic and Racial Studies 26 (5): 931–­61. McKay, Ruth B., Linda L. Stinson, Manuel de la Puente, and Brian A. Kojetin. 1996. “Interpreting the Findings of the Statistical Analysis of the CPS Supplement on Race and Ethnicity.” Washington, DC: Bureau of Labor Statistics, U.S. Bureau of the Census. Miller, Kristen, and Stephanie Willson. n.d. “Asking about Race: Survey Question Design for Respondents with Multiple Race Identities.” Washington, DC: National Center for Health Statistics. Murguía, Edward, and Edward E. Telles. 1996. “Phenotype and Schooling among Mexican Americans.” Sociology of Education 69 (October): 276–­89. Office of Management and Budget. 1997. “Revisions to the Standards for the Classification of Federal Data on Race and Ethnicity.” Federal Register (notice, October 30).



A Holistic Alternative 83

Omi, M., and H. Winant. 1994. Racial Formation in the United States: From 1960s to 1990s. New York: Routledge. Owens, Dawn, T. Robinson, and Lynn Smith-­Lovin. 2010. “Three Faces of Identity.” Annual Review Sociology 36: 477–­99. Prewitt, Kenneth. 2005. “Racial Classification in America: Where Do We Go from Here?” Dædalus: The Journal of the American Academy of Arts and Sciences (Winter): 4–­17. Sellers, R. M., S. A. Rowley, T. Chavous, N. Shelton, and M. Smith. 1997. “Multidimensional Inventory of Black Identity: Preliminary Investigation of Reliability and Construct Validity.” Journal of Personality and Social Psychology 73 (4): 805–­15. Senzai, Farid. 2012. Engaging American Muslims: Political Trends and Attitudes. Clinton, MI: Institute for Social Policy Understanding. Shariff-­Marco, Salma, Nancy Breen, Hope Landrine, Bryce B. Reeve, Nancy Krieger, Gilbert C. Gee, David R. Williams, Vickie M. Mays, Ninez A. Ponce, and Margarita Alegría. 2011. “Measuring Everyday Racial/Ethnic Discrimination in Health Surveys: How Best to Ask the Questions, in One or Two Stages, across Multiple Racial/Ethnic Groups?” Du Bois Review 8 (1): 159–­77. Sheldon, Trevor, and Hilda Parker. 1992. “Race and Ethnicity in Health Research.” Journal of Public Health Medicine 14 (2): 104–­10. Snipp, C. Matthew. 2003. “Racial Measurement in the American Census: Past Practices and Implications for the Future.” Annual Review of Sociology 29: 563–­88. Tafoya, Sonya. 2004. “Shades of Belonging.” Pew Hispanic Report (December 6). Washington, DC: Pew Hispanic Center. Telles, Edward, and Vilma Ortiz. 2009. Generations of Exclusion: Mexican Americans, Assimilation, and Race. New York: Russell Sage. Tucker, Clyde, Brian Kojetin, and Roderick Harrison. 1996. “A Statistical Analysis of the Current Population Survey Supplement on Race and Ethnic Origin.” Proceedings of the Bureau of the Census, 1996 Annual Research Conference, Rosslyn, VA. Ulmer, Cheryl, Bernadette McFadden, and David Nerenz, eds. 2009. Race, Ethnicity, and Language Data: Standardization for Health Care Quality Improvement (Institute of Medicine). Washington, DC: National Academies of Sciences. Ver Ploeg, Michele, and Edward Perrin, eds. 2004. Eliminating Health Disparities: Measurement and Data Needs. Washington, DC. National Research Council. Viruell-­Fuentes, Edna A. 2011. “‘It’s a Lot of Work’: Racialization Processes, Ethnic Identity Formations, and Their Health Implications.” Du Bois Review 8 (1): 37–­52. Viruell-­Fuentes, Edna A., P. Y. Miranda, and S. Abdulrahim. 2012. “Racialization Processes, Intersectionality Theory, and Immigrant Health: Moving Beyond Cultural Explanations.” Social Science and Medicine. DOI: 10.1016/j.socscimed.2011.12.037. Williams, David. 1997. “Race and Health: Basic Questions, Emerging Directions.” Annals of Epidemiology 7 (5): 322–­33. ———. 1999. “Race, Socioeconomic Status, and Health: The Added Effects of Racism and Discrimination.” Annals of the New York Academy of Sciences 896: 173–­88. Williams, D. R., S. A. Mohammed, J. Leavell, and C. Collins. 2010. “Race, Socioeconomic Status, and Health: Complexities, Ongoing Challenges, and Research Opportunities.” Annals of the New York Academy of Sciences 1186: 69–­101. Woo, Meghan, S. Austin, D. Williams, and Gary Bennett. 2011. “Reconceptualizing the Measurement of Multi-­Racial Status for Health Research in the United States.” Du Bois Review 8: 25–­36. Zuberi, T. 2001. Thicker than Blood. Minneapolis: University of Minnesota Press.

Chapter 6

Simon J. Craddock Lee

Organizational Practice and Social Constraints Problems of Racial Identity Data Collection in Cancer Care and Research

This chapter focuses on the pragmatics of how race and ethnic identity data are collected in the course of clinical care, drawing from my own interactions with the staff and procedural mechanisms of a university medical center and county safety-­net hospital. This chapter differs from others in this volume because I set aside topics such as the legitimacy of the race construct, its social construction, race as a biological fallacy, and the reification of race.1 I want to bracket those uncertainties. I also do not directly engage with the methodological challenges of how “race” or “ethnicity” may be engaged in the discipline of anthropology. Instead, by turning to everyday processes, I show the challenges of such categorical thinking even as we try to use these labels to identify, prevent, and remedy health disparities. The Institute of Medicine (IOM) has recognized that standardized data collection is critical to the understanding and elimination of racial and ethnic disparities in health care (IOM 1999). However, these data are collected inconsistently in the course of care, impacting patient-­provider relations, quality control, and even the clinical research enterprise (Burchard et al. 2003). A landmark empirical study by the Commonwealth Fund conducted in-­depth surveys of 272 hospitals in partnership with the American Hospital Association annual membership survey. Commonwealth investigators also conducted extensive site visits at six leading health systems, including Parkland Health and Hospital System, our county safety-­net teaching hospital here in Dallas. Study findings identified significant barriers to adequate data collection in operations processes and at the social-­behavioral level, that is, in administrative culture (Hasnain-­Wynia et al. 2004). For example, 30 percent of hospitals

87

88

Simon J. Craddock Lee

surveyed reported “drawbacks” to collecting race/ethnic identity data including discomfort on the part of the registrar or admitting clerk asking the patient for the information; problems associated with the accuracy of the data collected; a sense that patients might be insulted or offended, or resist answering questions about their race and ethnicity; the categories given often did not “fit” the patients; a fear that data may not be kept confidential; and the possibility that collecting data on race and ethnicity might be used to profile patients and discriminate in the provision of care. Similarly, studies of cancer patients have repeatedly found significant variations in the practice of collecting race, ethnicity, and birthplace information (for example, to track acculturation) for patients in community hospitals (Gomez et al. 2003; Polednak 2005) and clinical cancer research settings (Ford et al. 2002). A more recent study indicates that this variability persists despite increasing national attention to inclusivity policies in research and clinical care (Epstein 2007, 2008; Gomez, Satariano et al. 2009). Reporting Research Inclusivity

Building on my prior research at the National Cancer Institute, I have framed my research program in cancer disparities in terms of the organizational cultures of health science and health care (Lee 2009a, 2009b). But I have come to undertake concurrent patient-­focused anthropological research to understand the assumptions and social complexity inherent to the national assertion that “we need to increase minority accrual to clinical trials.” Working directly with cancer patients, however, brought this data collection and reporting challenge into focus. In this chapter, I use some of these field observations to illustrate the pragmatic challenges of operationalizing such identity data. In order to situate my research, I begin with two fieldwork vignettes. I met Bernard and Daniela as part of a pilot study to better understand a potential change in lung cancer therapy that my oncologist colleagues were debating in the journals as well as in their own clinical practice.2 Bernard was my age, late thirties and somewhat young to have advanced non-­small cell lung cancer. He was professionally involved in clinical care himself, with advanced training in his field, and as I have discussed elsewhere, many of us who work in healthcare settings find clinics surprisingly uncanny when we become patients ourselves. When I met him, Bernard had had surgery to resect the tumors where possible followed by mono-­therapy. I would learn later there was progressive disease. When he talked with me, he was due to receive more first-­line chemotherapy but was not in a therapeutic trial. (Field note, summer 2010)



Organizational Practice and Social Constraints 89

Daniela was nearly twenty years older but, unlike Bernard, she’d undergone targeted radiation for her metastases followed by two cycles of chemotherapy with only minor response to the regimen. She’d had further radiotherapy for spine metastases when I met her. She was also planning on more first-­line chemo, though a dose reduction had been recommended to alleviate the side effect of cytopenia. Daniela was also not in a therapeutic trial. (Field note, autumn 2010)

Our team planned a qualitative pilot study focused on patient attitudes to the standard chemotherapy regimen with an emergent paradigm shift to maintenance chemotherapy (Gerber et al. 2012). We suspected structural differences would exist for individuals who navigate the experience as a patient through university clinics and those who navigate their care through the county safety-­ net hospital although they may be managed by the same attending physicians. Our science was not focused on cultural group differences so much as the procedural experience of repeated oncology visits, the routine of drug administration, and the way that the tedium of cancer care might shape patient decision making from diagnosis through prognosis. Thus, although we identified patients like Bernard and Daniela through their attending physicians and used data from the electronic medical record (EMR), specific race and ethnicity were not inclusion or exclusion criteria (however, as an exploratory study, we did exclude no-­or low-­English proficiency patients). Thus, for the focus groups I conducted, the recruitment conversations did not include patient self-­report at all, only soliciting individual interest in the study. We could have added a short patient survey to confirm demographic factors (age, sex, education, race, ethnicity) at the time, but we were sensitive to patient burdens given the likelihood of medication-­induced fatigue, shortness of breath, and the fact that we were conducting these sessions between clinic appointments in lieu of asking participants to travel on top of their clinic schedule. Jane manages one of the disease-­oriented teams of physicians, nurses, and research coordinators that I work with and she runs a tight ship; she has an MBA and it shows. That winter, Jane helped me pull together the material needed to file a Continuing Review for the Institutional Review Board (IRB) that assures adequate human subject protections. At the time, we’d completed two focus groups in the university clinics and were in the midst of scheduling a third at our county hospital partner across the street. In submitting the documentation to extend IRB approval for another year, we were expected to report ethnic, racial, and vulnerable subject enrollment since study activation. This reporting

90

Simon J. Craddock Lee

requires two tables: the first depicts the number of participants to date by ethnic origin; the second, by racial origin. I reviewed the focus group transcripts and thought about themes, but because we would not begin coding until we completed all the sessions, we hadn’t yet needed to correlate any EMR data with specific participants in the transcripts. (Focus Group Analysis Plan, spring 2011)

When I explained to Jane that we had not collected patient self-­ report in the course of the focus groups, Jane offered to go pull the race/ethnicities as recorded in the EMR. Instead we opted to separate our participants by sex and indicate both racial origin and ethnic origin as “unknown or not reported.” At my suggestion, Jane noted in the continuing review that “race and ethnicity remain to be determined as this information was not directly collected from patients at the time the focus group was conducted.” At study completion, we submitted these data as indicated in the patient EMR, although we did not ourselves solicit it from patient self-­report either as part of study protocol or as part of the recruitment process. On reflection, this is both more straightforward and more convoluted. Jane was not particularly satisfied, though there was no scientific reason to have made self-­report part of the pilot study protocol: public or private payor, as a proxy for socioeconomic status (SES), was the more salient marker of difference for that particular research question (Yang et al. 2010; Yorio et al. 2009). Our EMR is a proprietary system based on Epic (Verona, WI) that was implemented across the university clinical functions in November 2008. The system went live at Parkland less than five months later. There is a common adjunct clinical trial management system called Velos (Fremont, CA) that we use to track patients who participate in clinical research (for example, cancer trials) at both institutions. Velos can autopopulate some fields by importing directly from the Epic EMR. Jane and her research study colleagues have an explicit patient self-­report protocol that they follow to solicit patient race and ethnicity identity information, according to a structured script, when they recruit patients for studies. Per federal Office of Management and Budget (OMB) guidelines, this script breaks out ethnicity (Hispanic, non-­Hispanic) separately from racial origin (African American/Black, Caucasian, Asian American, and so forth). But when the study coordinators input the data they’ve solicited from a patient into Velos, the software interface offers only the single drop-­down box for “race/ethnicity.” Clinical trial staff must know to scroll through all the combinations (White, non-­Hispanic; White, Hispanic; Black, non-­Hispanic; Black,



Organizational Practice and Social Constraints 91

Hispanic, and so forth) to find the appropriate combination. Thus, although the information is solicited from patients according to the U.S. Census survey design, the answers must be recombined by staff in order to be recorded in the clinical research database. Based on the fieldwork I have conducted since arriving at University of Texas Southwestern Medical Center in 2008, staff and clinicians are uncertain about any reciprocity that may occur between Velos and Epic. That is, if the fields are empty when drawn down from Epic but are then subsequently populated in Velos, we do not believe the Epic fields are automatically updated. Similarly, we don’t believe that updates in Epic migrate to Velos after the initial data draw by trial staff, and we don’t know whether partial updates are possible. Perhaps more importantly, staff and clinicians are unclear who is able to update or change existing field choices in Epic. For example, if the field is vacant, can anyone populate it at any time? Should updates override prior entries? Why and in what circumstances? How do we control for Type III error (updated answer but invalid source)? Various people over the course of my fieldwork have commented that physician notes often contain abbreviations for race and/or ethnicity in the opening comments of an encounter record (field notes, 2009). However, there is no standard notation; clinicians may note race but not ethnicity or treat them interchangeably. There is certainly no indication that this information was received by patient self-­report. Almost all of the physicians I have worked with, regardless of training level or seniority, do not ask patients to self-­identify but annotate their assumption based on (perhaps unconsidered) visual assessment. My own fieldwork further suggests that the race/ethnicity notation that characterizes the SOAP3 note tends to be an artifact of medical school and residency; the notation becomes less common with physician seniority, mediated by teaching responsibilities. Although race and ethnicity did not factor into our inclusion criteria, we know we see many more minority patients through the oncology service at Parkland than at the university clinics. Even with my tiny sample of focus group participants, the university clinic patients were all tagged as White (with one Asian exception) in the Epic record while the public hospital participants were evenly split between White and Black patients. At least, that’s what we are able to report based on what the patient records indicate—­although we have no idea whether these fields are accurate. I suspect at least one of the patient participants I worked with was Hispanic, although her last name maps to Anglo heritage (Fiscella and Fremont 2006; Sweeney et al. 2007). Her Epic data field

92

Simon J. Craddock Lee

only indicates White—­this could be White Hispanic but absent standardized patient self-­report, we cannot know. Attempts to Standardize Data Collection

In 2009, a draft policy for collecting race and ethnicity data across all university patient care units made an effort to articulate standard practice (table 6.1). I work through the draft policy here in order to illustrate what I believe are common, and largely understandable, challenges to collecting data in the context of service delivery. These efforts are undercut by real concerns about bias and discriminatory practice, as well as the simple fact that patients’ and staff ideas about identity often do not map to bureaucratic categories. My institution took on this effort in an open and transparent way and has continued to make advances in this area. Our process serves here as an instructional case in point. All too easily noted, the interpersonal anxiety that arises in discussions (and non-­ discussions) that touch on human group differences have direct effects on what people are willing to do even in the course of otherwise very professional behavior (Kim et al. 2008). For example, research indicates measurable differences in clinic communication where patients in race-­discordant interactions do less to prompt doctors for information and doctors in turn provide less information to these patients (Gordon et al. 2006). Nonverbal, even unintentional communication is also involved and both clinician and patient race seem to be contributing factors (Stepanikova et al. 2012).

Table 6.1   Initial Draft Policy Categories

American Indian or Alaska Native American Indian or Alaska Native–­Hispanic or Latino Asian Asian-­Hispanic or Latino Black or African American Black or African American–­Hispanic or Latino Declined Hispanic or Latino Native Hawaiian or Other Pacific Islander Native Hawaiian or Other Pacific Islander–­Hispanic or Latino Other/Unknown Other/Unknown-­Hispanic or Latino White White-­Hispanic or Latino



Organizational Practice and Social Constraints 93

The 2009 draft policy has several limitations. First, the options “to be used by all appropriate systems that register and/or schedule patients” were represented in a single column list of fixed options that I suspect was produced as an echo of how the data was laid out in the new electronic records system. In this list, the primary position is occupied by racial origin group, with a clausal option where ethnic group is a secondary classification, but Hispanic or Latino also exists as a primary category alone. Other/Unknown only occupies the primary position (racial) but also comes with a secondary classification (ethnic). The 2009 draft policy indicates that an attempt to collect these data must be made at appointment check-­in. If a patient refuses or shows as “declined” or “other/unknown” as an initial response, then the staff is instructed not to ask about race/ethnicity. However, although patient self-­report is the accepted gold standard, the next procedural point requests a “visual assessment” by the staff person who chooses “the most appropriate response based on visual verification.” The instructions further note that “this information is voluntary and will not in any way impact the quality of care patients receive while under our care.” Unremarked upon is the fact that the institution’s effort to collect this information is not voluntary, given federal requirements. The draft further explains that “staff can limit their list of responses based on patient appearance if in-­person collection; for example, we would not ask someone who appears African American if they are Caucasian or Asian and vice versa, etc.” The draft provides a script “to address patient concerns.” Here at UTSW we are a healthcare provider that also conducts research studies for new medicines, procedures, tests etc to improve the health of the patients we serve. In order to conduct research, we are required to request racial and ethnic information from all of the patients we care for. Your assistance in collecting this information plays an important part in our research efforts and the effectiveness of research is dependent on using the correct race and ethnic data to identify potential research participants. This information is voluntary and will not in any way impact the quality of care you receive while under our care. We will ask you to select from a list of choices so we may collect this information as easily as possible; please respond with your one best response to these 2 questions.

The oral questions ask, first, whether the patient is of “any Hispanic or Latino origins.” If so, only those choices from the list with Hispanic or Latino in the secondary position should be offered. If the patient answers in the negative,

94

Simon J. Craddock Lee

then staff should only offer the choices that lack Hispanic or Latino in the secondary position. This draft policy reflects both bureaucratic-­procedural and sociocultural obstacles to the consistent collection of race and ethnicity identity data in health care. Staff worry patients will be offended, consequently they are embarrassed or unwilling and thus ill-­prepared to solicit the data in a consistent way that follows best practices for patient self-­report. Further, the need to report these data to regulatory bodies creates the bureaucratic demand to populate the data fields even if the data entered are invalid. Thus, the draft encouraged visual assessment despite ample evidence in the literature that this is inconsistent, error-­prone, and subject to observer selection bias. The figure of “research studies” occupies an interesting position in this discourse. The draft policy positions “research” as the justification for and the beneficiary of this data collection. Moreover, in order to deny that the collection of these data figures into the quality of care that a patient receives at our academic medical center, the policy sidesteps the notion that race and ethnicity might play any role in the delivery of care, instead asserting such data are only salient to the abstract and distal research. As a researcher, I was intrigued that the draft also alludes to external government mandates without identifying their source. In the revised 2011 policy, the language has been changed. The document now emphasizes the value of using electronic health information in a “meaningful way that augments clinical quality improvement,” citing federal stimulus funding in the American Recovery and Reinvestment Act legislation (ARRA 2009). Thus, the collection of racial and ethnic data is framed as an issue of electronic records and is explicitly linked to quality of care rather than research. In making this shift, the institutional voice also moves toward taking ownership and responsibility for the need to collect these data for a purpose that is fundamentally grounded in their primary public mission. The impetus still derives from an external source, federal law, but it is couched in the more proximate terms of patient medical records rather than the abstract reference to “research.” Racial and ethnic terms are set forth in the minimum standard categories of OMB Directive 15 (1997) to establish at least logical consistency, regardless of construct validity. Thus, our 2011 policy specifically indicates that ethnicity data should be collected before race and rephrases the two questions to be asked of patients: “Do you consider yourself Hispanic or Latino? Which category best describes your race?” “Some other race” has been added to the list of fixed options and the “declined” option is followed by “unavailable/ unknown.” This enables personnel to record attempts to collect the data when



Organizational Practice and Social Constraints 95

the respondent’s response fell outside the given options. Importantly, the 2011 policy specifically prohibits any attempt to enter patient race or ethnicity information based on visual assessment. The script provided to address patient concerns has also been amended. No longer referring to research, it explains that the data is used “by our physicians to further understand the nature of diseases and opportunities that may exist in the future for the enhancement of treatment, therapies and possible cures. UT Southwestern healthcare providers are committed to developing programs that meet the needs of the community and all of our patients. Your assistance in collecting this information plays an important part in our efforts.” These policies received renewed scrutiny and revision following the stipulations of federal funding during the recession. Federal funds, especially the obligations tied to active participation in Medicare reimbursement as well as explicitly research and process-­improvement monies, are powerful incentive to standardization. However, the evolution I highlight here reflects a confluence of efforts, both internal and external, to make data relevant, consistent, and useful even while we continue to struggle with the excess of meaning that will probably always surround categorical thinking (Lee 2005; Lee 2006). I remain concerned about how these changes will be implemented because disparities in access to health care are related to disparities in health outcomes: lack of preventive care, later stage at diagnosis, delayed or incomplete treatment (Do et al. 2010; Freeman and Chu 2005; Gerend and Pai 2008). In Texas, nearly a third of our population is under-­or uninsured; Dallas County is the ninth most populous county in the United States, with an estimated 2.4 million residents, of whom 39 percent are Hispanic, 33 percent are White, and 23 percent are African American (U.S. Census). This cultural (and linguistic) diversity directly impacts the delivery of care even before a patient presents. Work in psychology and other fields has shown that constructs of identity like ethnicity and race are plastic and situation-­contingent (Gillborn 1995; Okamura 1981; Phinney 1996; Schnittker 2002). Symbolic interactionism earlier established dialogic models, demonstrating that subjectivity is both externally imposed and internally claimed (Dunn 1997; Mead 1967 [1934]). Identity claims can be made tactically within political strategies. An individual or group may frame claims on the basis of race identity or class position interchangeably, depending on which approach is most politically expedient. In the clinic setting, then, patient self-­report remains the gold standard, but nevertheless reflects the structural power imbalance inherent in a sick person seeking care from a formal organization. This means that however much a clinic or hospital seeks to reassure prospective patients and their families that these data

96

Simon J. Craddock Lee

are collected for benign, even benevolent reasons, asking about someone’s race or ethnicity remains a loaded question in the United States. Similarly, how people answer such questions depends on why they think you asked and how they think any given answer will serve. Patients are vulnerable by definition because providers have something they need and this imbalance colors any request for information. Front-­line personnel need explanation of the policy and training on the procedure for soliciting patient self-­report. For example, Patient Financial Services personnel who determine proof of insurance or assessments to receive public assistance may find it awkward to solicit race and ethnicity identity information. Administrative personnel can be as affected by the productivity pressures of a busy patient day as their counterparts who provide actual clinical care. But administrative personnel are much less likely to be recognized for the emotional labor involved in engaging patients (Ward and McMurray 2011) although courteous and efficient patient/family interactions clearly impact hospital satisfaction survey scores. Investigators have found that the number of Massachusetts hospitals using race, ethnicity, and language data internally nearly doubled following a 2007 state mandate (Jorgensen et al. 2010). But executives from seventeen facilities initially reported significant staff fears of upsetting patients whereas only nine facilities reported actual concerns from patients. The Massachusetts Hospital Association developed training materials to anticipate these concerns and reported uniform implementation. At least one facility suspected work-­arounds where staff members were modifying the process “to minimize their discomfort about asking potentially intrusive questions” and continued to monitor the data and planned for additional training with key players. Staff concerns subsided as the data collection became routine, although the change to standardize “Hispanic” as ethnicity rather than race remained a problem for many facilities serving for large Latino communities, whose members tended to see themselves as a racial rather than an ethnic group. As large health care providers in North Texas, we care for the rapidly growing Hispanic/Latino population; our hospitals and clinics need to develop specific question and answer scripts for our front-­line staff to feel comfortable discussing how the OMB categories are organized in order to walk patients through efforts to collect related data. Despite administrative anxiety, public attitudes support the need for health care providers to collect such identity data (nearly 88 percent), indeed a majority (greater than 75 percent) supported legislation to require identity data collection in health care (Baker et al. 2007). While 17 percent of respondents were uncomfortable reporting their own race/ethnicity, nearly half worried



Organizational Practice and Social Constraints 97

that the information could be used to discriminate against them. Importantly, more than a third of Hispanic patients were uncomfortable reporting English proficiency. Another study reveals that as compared to Whites, minority individuals reporting prior perceived experiences of discrimination were more likely to express significant worry that identity information could be used for discriminatory purposes in the provision of care (Kandula et al. 2009). Clear institutional statements and communication training are needed to call out specifically why these data are necessary and how these data will be used (see López, this volume). Provider policies should anticipate patient concerns with prepared statements that empower front-­line personnel with confidence that such data collection supports their mission to provide quality care for their patients (Weissman and Hasnain-­Wynia 2011). Validating Hospital Data

The collection of race/ethnicity identity data within our academic medical center is inconsistent, despite specific organizational policy addressing this practice. The only demographic data collected on a consistent basis is current age and gender (sex), according to my colleagues who manage the academic medical center’s data warehouse. We find that only a limited number of existing records (less than 30 percent at best) for all clinical conditions currently contain populated fields for race and ethnicity variables. In contrast, at Parkland, previous work by my partners indicate that while race and ethnicity identity fields are largely populated (greater than 90 percent), these data have never been validated; that is, it is unknown whether the recorded data reflect patient self-­report. In 2011, I was funded to conduct an audit of race and ethnicity identity data for oncology-­related services, first to determine baseline compliance and then to design and implement a follow-­up analysis to assess data validity in the current patient databases. We developed a script for telephone patient interviews to validate populated race and ethnicity fields. Prior to asking for self-­identification by OMB categories, we asked, “Please tell me how you describe your own race or ethnic background?” with an open answer field, followed by a question drawn from the CDC’s Behavioral Risk Factors Surveillance System (BRFSS) survey module on Reactions to Race, “How do other people usually classify you in this country?”4 We sought to acknowledge that the OMB categories rarely map to individual constructs. Incorporating these items into a hospital intake process could clarify both what we mean to ask and indicate institutional recognition of imperfect categories. However ironic, articulating limitations can alleviate frustration and animosity in both staff and presenting patient.

98

Simon J. Craddock Lee

We also included the Healthcare System Distrust scale to explore whether patient feelings toward provider organizations correlate with the absence or inaccuracy of EMR identity data (Shea et al. 2008). Psychologists have conducted studies indicating that the strength of implicit bias tends to exceed self-­report of bias (Sabin et al. 2009). Such findings raise our awareness of the potential impact of assumptions staff make about the people we care for and, in turn, patient-­public ideas about how we will treat them. My analyses are forthcoming; we hope to inform interventions with clinic staff designed to anticipate patient distrust and demonstrate how identity data requests serve patient care. Implications: Imperfect Data for an Imperfect World

I remain ambivalent about what is lost and what is reified in using OMB categories. In lieu of another theoretical argument, I have tried to show how the process of attempting to “capture these data” itself reveals the arbitrary and social artifice of the cultural practice of racializing. Nonetheless, to reach underserved communities, we have to work with these categorical data in order to establish more nuanced analyses of group differences in patient decision making or access to clinical trials. The use of race and ethnicity variables in federal and state reporting is a foundation for public health care funding (for example, Medicaid, Disproportionate Share Hospital “DSH” payments), as well as for targeting particular disparities in health care. Even after access to care has been demonstrated, Black patients obtain surgery for lung cancer less often than Whites (Lathan et al. 2006). They are more likely not to have surgery recommended and more likely to refuse surgery. Patient treatment preferences are informed by different cultural beliefs (Kim et al. 2008), but we also see racial differences in patient perceptions of and trust in physicians (Gordon et al. 2006). Without saying anything about the relationship between race constructs and causal mechanisms of disparities, OMB categories still offer a proxy by which to circumscribe groups we can then engage in more nuanced analyses of social disadvantage. Although limited to the cancer service lines, my audit will enable us to benchmark data collection and establish best practices for other service lines within our safety-­net system for dissemination to other provider organizations. Each of the database sources we audit corresponds to a different time point in the patient care and research participation continuum (figure 6.1). Our findings should help identify specific opportunities to intervene in the cancer patient data collection process to ensure consistent and accurate recording of race and ethnic identity data. And most importantly, race and ethnicity data can be used to provide immediate feedback to local public health partners seeking to redress disparities in service and outcomes (see López, this volume).

for reporting to State of TX

Current Epic Medical Record

Deceased

Race/ethnicity solicited from next of kin?

Enroll or Decline

Race/ethnicity solicited and recorded during research consent

Velos Research Patient Database

Prognosis

Invite to Clinical Trial

Figure 6.1   Conceptual Schematic of Identity Data Collection in a Cancer Service Line

Tumor Registry at Clinic

Race/ethnicity solicited and recorded

Race/ethnicity solicited and recorded

Archived Epic Medical Record

Cancer Diagnosis

Presentation at Clinic

Not Invited

No race/ethnicity data recorded?

Discharge or After Care

10 0

Simon J. Craddock Lee

Health services categories are intended to circumscribe. Analytic constructs exist to create instrumental specificity. Human lived experience offers a surplus of meaning beyond our efforts to delineate these distinctions. But as my accounts of IRB and patient care reporting illustrate, we must attend carefully to what these data do and do not represent as well as the specific purpose for which such reporting is intended. Disparities in cancer care, and health outcomes generally, require we address both fundamental causes and downstream challenges across the continuum of care (Franks and Fiscella 2008). Our ability to develop stronger measures of social disadvantage is increasing (Altschuler et al. 2004) although those models are driven by disparities originally identified using these necessarily imperfect categories of “ethnicity” and “race.” The IOM recommendations on data standardization and ARRA requirements for the intelligent use of electronic medical data provide strong impetus for health care providers to rationalize their data collection procedures—­creating not only standard policies but implementing the training, audit, and feedback processes we need to address the organizational culture and social constraints so well documented by the Commonwealth study. Notes 1. Other chapters address the conceptual limitations of the “race” construct, especially Kaufman, Garcia, López, and Saperstein, in this volume. Here, I recognize “racial” minorities not as homogenous biological groups but from a cancer disparities perspective focused on socially vulnerable, heterogeneous population groups (Kilbourne et al. 2006; LaVeist 1994). 2. Data collection and fieldwork, respectively, supported in part by a special interest award for cancer disparities research (American Cancer Society-­Institutional Research Grant 02–­196) and by a pilot award through the UT Southwestern Clinical and Translational Alliance for Research (NIH UL1-­RR024982). Individual patients and staff are intended to be representative; names have been changed to prevent identification, however, salient characteristics remain accurate. 3. Subjective-­Objective-­Assessment-­Plan: standard mnemonic for patient workup by physicians that informs structure of physician note. 4. Particular thanks to Camara Jones and Aliya Saperstein for continued conversations about variations on self-­report items and the BRFSS module that were started at the workshop hosted by the RWJF Center for Health Policy in Albuquerque (April 29–­ 30, 2011; see also Penner and Saperstein 2008). References Altschuler, Andrea, Carol P. Somkin, and Nancy E. Adler. 2004. “Local Services and Amenities, Neighborhood Social Capital, and Health.” Social Science and Medicine 59 (6): 1219–­29. ARRA. 2009. “American Recovery and Reinvestment Act of 2009.” In Public Law 111–­5, Washington, DC: U.S. Congress.



Organizational Practice and Social Constraints 101

Baker, David W., Romana Hasnain-­Wynia, Namratha R. Kandula, and Jason A. Thompson. 2007. “Attitudes toward Health Care Providers Collecting Information about Patients’ Race, Ethnicity, and Language.” Medical Care 45 (11): 1034–­42. Burchard, Esteban Gonzalez, Elad Ziv, Natasha Coyle, Scarlett Lin Gomez, Hua Tang, Andrew J. Karter, Joanna L. Mountain, Eliseo J. Perez-­Stable, Dean Sheppard, and Neil Risch. 2003. “The Importance of Race and Ethnic Background in Biomedical Research and Clinical Practice” New England Journal of Medicine 348 (12): 1170–­75. Do, Young, William Carpenter, Pamela Spain, Jack Clark, Robert Hamilton, Joseph Galanko, Anne Jackman, James Talcott, and Paul Godley. 2010. “Race, Healthcare Access, and Physician Trust among Prostate Cancer Patients.” Cancer Causes and Control 21 (1): 31–­40. Dunn, R. G. 1997. “Self, Identity, and Difference: Mead and the Poststructuralists.” Sociological Quarterly 38 (4): 687–­705. Emmons, Karen M., Mei Wong, Elaine Puleo, Neil Weinstein, Robert Fletcher, and Graham Colditz. 2004. “Tailored Computer-­based Cancer Risk Communication: Correcting Colorectal Cancer Risk Perception.” Journal of Health Communication 9 (2): 127–­41. Epstein, Steven. 2007. Inclusion: The Politics of Difference in Medical Research. Chicago Studies in Practices of Meaning. Chicago: University of Chicago Press. ———. 2008. “The Rise of ‘Recruitmentology’: Clinical Research, Racial Knowledge, and the Politics of Inclusion and Difference.” Social Studies of Science 38 (5): 801–­32. Fiscella, Kevin, and Allen M. Fremont. 2006. “Use of Geocoding and Surname Analysis to Estimate Race and Ethnicity.” Health Services Research 41 (4p1): 1482–­1500. Ford, M. E., D. D. Hill, D. Nerenz, M. Hornbrook, Jane Zapka, R. Meenan, S. Greene, and C. C. Johnson. 2002. “Categorizing Race and Ethnicity in the HMO Cancer Research Network.” Ethnicity and Disease 12 (1): 135–­40. Franks, Peter, and Kevin Fiscella. 2008. “Reducing Disparities Downstream: Prospects and Challenges.” Journal of General Internal Medicine 23 (5): 672–­77. Freeman, Harold P., and Kenneth C. Chu. 2005. “Determinants of Cancer Disparities: Barriers to Cancer Screening, Diagnosis, and Treatment.” Surgical Oncology Clinics of North America 14 (4): 655–­69. Gerber, David E., Heidi A. Hamann, Drew W. Rasco, Sharon Woodruff, and Simon J. Craddock Lee. 2012. “Patient Comprehension and Interest in Matintenance Chemotherapy for Lung Cancer.” Patient Education and Counseling. DOI: 10.1016/j.pec.2012.04.013. Gerend, Mary A., and Manacy Pai. 2008. “Social Determinants of Black-­White Disparities in Breast Cancer Mortality: A Review.” Cancer Epidemiology Biomarkers and Prevention 17 (11): 2913–­23. Gillborn, David. 1995. “Racism, Identity, and Modernity: Pluralism, Moral Antiracism, and Plastic Ethnicity.” International Studies in Sociology of Education 5 (1): 3–­23. Gomez, Scarlett Lin, Gem M. Le, Dee W. West, William A. Satariano, and Lilia O’Connor. 2003. “Hospital Policy and Practice Regarding the Collection of Data on Race, Ethnicity, and Birthplace.” American Journal of Public Health 93 (10): 1685–­88. Gomez, Scarlett Lin, William A. Satariano, Gem M. Le, P. Weeks, L. McClure, and Dee W. West. 2009. “Variability among Hospitals and Staff in Collection of Race, Ethnicity, Birthplace, and Socioeconomic Information in the Greater San Francisco Bay Area.” Journal of Registry Management 36 (4): 105–­10. Gordon, Howard S., Richard L. Street, Jr., Barbara F. Sharf, P. Adam Kelly, and Julianne Souchek. 2006. “Racial Differences in Trust and Lung Cancer Patients’ Perceptions of Physician Communication.” Journal of Clinical Oncology 24 (6): 904–­9.

102

Simon J. Craddock Lee

Hasnain-­Wynia, Romana, Debra Pierce, and Mary A. Pittman. 2004. Who, When, and How: The Current State of Race, Ethnicity, and Primary Language Data Collection in Hospitals. New York: Commonwealth Fund, Health Research and Educational Trust. Institute of Medicine (IOM). 1999. “Unequal Burden of Cancer: An Assessment of the NIH Research and Programs for Ethnic Minorities and the Medically Underserved.” Washington DC: National Academy Press. Jorgensen, Selena, Ruth Thorlby, Robin Weinick, and John Ayanian. 2010. “Responses of Massachusetts Hospitals to a State Mandate to Collect Race, Ethnicity, and Language Data from Patients: A Qualitative Study.” BMC Health Services Research 10 (1): 352. Kandula, Namratha, Romana Hasnain-­Wynia, Jason Thompson, E. Brown, and David Baker. 2009. “Association between Prior Experiences of Discrimination and Patients’ Attitudes towards Health Care Providers Collecting Information about Race and Ethnicity.” Journal of General Internal Medicine 24 (7): 789–­94. Kilbourne, Amy M., Galen Switzer, Kelly Hyman, Megan Crowley-­Matoka, and Michael J. Fine. 2006. “Advancing Health Disparities Research within the Health Care System: A Conceptual Framework.” American Journal of Public Health 96 (12): 2113–­21. Kim, A. W., M. J. Liptay, and R. S. Higgins. 2008. “Contemporary Review of the Inequities in the Management of Lung Cancer among the African-­American Population.” Journal of the National Medical Association 100 (6): 683–­88. Lathan, Christopher S., Bridget A. Neville, and Craig C. Earle. 2006. “The Effect of Race on Invasive Staging and Surgery in Non-­Small-­Cell Lung Cancer.” Journal of Clinical Oncology 24 (3): 413–­18. LaVeist, Thomas. 1994. “Beyond Dummy Variables and Sample Selection: What Health Services Researchers Ought to Know about Race as a Variable.” Health Services Research 29 (1): 1–­16. Lee, Simon J. Craddock. 2005. “The Risks of Race in Addressing Health Disparities.” Hastings Center Report 35 (4): Back. ———. 2006. “Rethinking Race and Ethnicity in Health Disparities.” Anthropology News 47 (3): 7–­8. ———. 2009a. “Notes from White Flint: Identity, Ambiguity, and Disparities in Cancer.” In Confronting Cancer: Metaphors, Inequality, and Advocacy, edited by Juliet McMullin and Diane Weiner, 165–­86. Santa Fe: SAR Press. ———. 2009b. “Science, Surveillance, and the Politics of Redress in Health Disparities Research.” Race/Ethnicity: Multidisciplinary Global Contexts 3 (1): 51–­74. Mead, George Herbert. 1967 (1934). Mind, Self, and Society: From the Standpoint of a Social Behaviorist. Works of George Herbert Mead, vol. 1, edited by Charles W. Morris. Chicago: University of Chicago Press. Okamura, Jonathan Y. 1981. “Situational Ethnicity.” Ethnic and Racial Studies 4 (4): 452–­65. Penner, Andrew, and Aliya Saperstein. 2008. “How Social Status Shapes Race.” Proceedings of the National Academy of Sciences 105: 19628–­30. Phinney, Jean S. 1996. “When We Talk about American Ethnic Groups, What Do We Mean?” American Psychologist (51): 918–­27. Polednak, A. P. 2005. “Collecting Information on Race, Hispanic Ethnicity, and Birthplace of Cancer Patients: Policies and Practices in Connecticut Hospitals.” Ethnicity and Disease 15 (1): 90–­96. Sabin, Janice A., Brian A. Nosek, Anthony G. Greenwald, and F. P. Rivara. 2009. “Physicians’ Implicit and Explicit Attitudes about Race by MD Race, Ethnicity, and Gender.” Journal of Healthcare for the Poor and Underserved 20 (3): 896–­913.



Organizational Practice and Social Constraints 103

Schnittker, J. 2002. “Acculturation in Context: The Self-­Esteem of Chinese Immigrants.” Social Psychology Quarterly 65 (1): 56–­76. Shea, Judy, Ellyn Micco, Lorraine Dean, Suzanne McMurphy, J. Schwartz, and Katrina Armstrong. 2008. “Development of a Revised Health Care System Distrust Scale.” Journal of General Internal Medicine 23 (6): 727–­32. Stepanikova, Irena, Qian Zhang, Darryl Wieland, G. Eleazer, and Thomas Stewart. 2012. “Non-­Verbal Communication between Primary Care Physicians and Older Patients: How Does Race Matter?” Journal of General Internal Medicine: 1–­6. Sweeney, Carol, Sandra L. Edwards, Kathy B. Baumgartner, Jennifer S. Herrick, Leslie E. Palmer, Maureen A. Murtaugh, Antoinette Stroup, and Martha L. Slattery. 2007. “Recruiting Hispanic Women for a Population-­ Based Study: Validity of Surname Search and Characteristics of Nonparticipants.” American Journal of Epidemiology 166 (10): 1210–­19. U.S. Census. “U.S. Census State and County Quickfacts: Dallas County, Tx.” Accessed February 21, 2012. http://Quickfacts.Census.Gov/Qfd/States/48/48113.Html. Ward, Jenna, and Robert McMurray. 2011. “The Unspoken Work of General Practitioner Receptionists: A Re-­examination of Emotion Management in Primary Care.” Social Science and Medicine 72 (10): 1583–­87. Weissman, Joel S., and Romana Hasnain-­Wynia. 2011. “Advancing Health Care Equity through Improved Data Collection.” New England Journal of Medicine 364 (24): 2276–­77. Yang, Relin, Michael C. Cheung, Margaret M. Byrne, Youjie Huang, Dao Nguyen, Brian E. Lally, and Leonidas G. Koniaris. 2010. “Do Racial or Socioeconomic Disparities Exist in Lung Cancer Treatment?” Cancer 116 (10): 2437–­47. Yorio, Jeffrey T., Jingsheng Yan, Yang Xie, and David E. Gerber. 2009. “Presence of Lung Cancer Treatment and Outcome Disparities within a Single Academic Medical Center.” Journal of Thoracic Oncology 4 (11): 1303–­4.

Chapter 7

Gabriel R. Sanchez and Vickie D. Ybarra

Lessons from Political Science Health Status and Improving How We Study Race

Over the past twenty-­plus years, hundreds of studies have verified the persistence of pervasive racial/ethnic inequities in access to health care, health outcomes, and health status in the United States (for a sampling, see AHRQ 2009; Berenson et al. 1996; Collins et al. 1999; Gornick 2000; James et al. 2007; McGuire and Miranda 2008; Satcher et al. 2005; Smedley et al. 2003; Waidmann and Rajan 2000). In fact, a recent edition of the premier academic journal for the study of race in the social sciences, the Du Bois Review, specifically focused on racial inequalities in health. The articles in this volume cover many critical topics related to the assessment of racial and ethnic health disparities, ranging from measuring multiracial populations (Woo et al. 2011), to how ethnic identity and acculturation influence health for Latinas (Viruell-­Fuentes et al. 2011), to racial differences in health issues associated with stress (Sternthal et al. 2011). Although the work of social scientists highlighted in this special issue constitutes a tremendous advance in this area of research, there remains room for continued knowledge production, and this chapter contributes to that goal. We specifically advocate for a more comprehensive approach to the measurement of race and ethnicity in this area of research in order to better understand the rich variation that exists both within and between racial/ethnic groups. This requires the implementation of multiple measures to capture race/ ethnicity beyond the dominant measure of self-­identification. For example, skin color has been employed in the U.S. health disparities literature for some time, with many scholars finding that individuals who are perceived by others to have darker skin as well as those who self-­report darker skin color tend to have poorer health (Gravlee et al. 2005; Klonoff and Landrine 2000). Furthermore,

104



Lessons from Political Science 105

experiences with discrimination have been incorporated in studies focused on determining the underlying causes of racial and ethnic health disparities (Borrell et al. 2007; Brondolo 2009; Finch et al. 2001; Gee et al. 2007; LaVeist 2000; Paradies 2006; Schulz et al. 2006; Williams 1996, 1999). Finally, group identity is a concept used often by political scientists to measure racial or ethnic differences in political behavior (Sanchez 2006; Sanchez and Masouka 2010) and has also been associated with health outcomes (for example, see Iwamoto et al. 2010; Viruell-­Fuentes et al. 2011). Although the extant literature has explored these three approaches to measuring race, they are very rarely incorporated in the same analysis. We attempt to fill this void by including measures for experiences with discrimination, skin color, and the perception that you need to “play down” your racial/ethnic identity to advance in society, in addition to self-­reported race/ethnicity, in our analysis, in order to better understand the impact of race/ethnicity on health. We are also highly interested in exploring internal variation within the Latino community. The U.S. Latino population is a diverse population that is comprised of individuals from more than twenty national origin groups, some of whom are recent immigrants and Spanish-­language dominant, and others who come from families that have been in the United States for more than five generations. This variation matters, as scholars have found that Latinos born in the United States and who are English speakers have higher rates of preventive service use, encounter fewer barriers in obtaining care, and are more likely to have health coverage (Lara et al. 2005). We therefore suggest that scholars move beyond research designs that treat Latinos (and other groups, for that matter) as a monolithic population. In line with this framework, we conduct split-­sample analyses for the Latino and Mexican-­specific populations. We anticipate that this split-­sample approach will reveal important variations in how our measures of race and acculturation influence health status among Latinos. The two primary foci of this chapter are therefore (1) to demonstrate the utility of incorporating multiple measures of race and/or ethnicity in research focused on the determinants of racial and ethnic health disparities, and (2) to demonstrate the gains from exploring internal variation within racial and ethnic groups for health disparities research. Our chapter addresses these two themes by studying the potential impact of political factors on health outcomes. We contend that the social determinants of population health and of racial/ethnic health inequities are themselves determined largely by political context and policy environment, and that these topics themselves constitute areas of scholarly inquiry ripe for political science perspectives. We engage in such analysis by investigating the relationship between trust in government, an

106

Gabriel R. Sanchez and Vickie D. Ybarra

important political variable connected to political behavior (Liu 2001; Verba et al. 1995), and health status.1 Measuring Race, Political Factors, and Health Status

This examination relies on data from the 2008 Collaborative Multi-­ Racial Post-­Election Survey (CMPS), a national telephone survey of registered voters conducted during the 2008 presidential election, with samples of self-­defined African Americans, Asian Americans, Latinos, and Whites who were surveyed between November 9, 2008, and January 5, 2009. The CMPS was the first multiracial and multilingual survey of registered voters across multiple states and regions in a presidential election, and was available in six languages.2 The final CMPS data set includes completed surveys of 4,563 respondents who self-­identified as Asian, Black, Latino, and White, permitting substantive comparisons among racial/ethnic groups. The targeted eighteen states contained two-­thirds of the 2008 registered voters nationwide, including 92 percent of all registered Latino voters, 87 percent of all registered Asian American voters, 66 percent of all registered African American voters, and 61 percent of all registered White voters (http://cmpstudy.com/index.html). This is critical for our initial analysis, which is focused on exploring the impact of our multidimensional approach to measuring race with the non-White sample. The CMPS provides individual level survey data that includes multiple measures of political context, a set of control variables, a measure of health status, and most importantly, multiple measures of race/ethnicity. The CMPS also provides an ideal setting to advance our interest in internal variation among the Latino population. Not only does the sample size of Latinos provide the ability to examine Latino-­specific models of health status, but we are also able isolate respondents of Mexican origin.3 Furthermore, the Latino sample of the CMPS has tremendous variation by nativity and language use, as 46 percent of the Latino respondents, for example, chose to conduct the survey in Spanish, and 57 percent of the respondents report being born in the United States. This is critical for our effort to explore internal variation within the tremendously diverse Latino population. The dependent variable used in this study is self-­reported health status (a novel variable in political studies) using the following question from the CMPS: “How would you rate your overall physical health—­excellent, very good, good, fair, or poor?”4 Although the operationalization of health status is defined by a single question in our analysis, there is justification for making generalizations based on this sole measure. For example, as a component of its Behavioral Risk Factor Surveillance System (BRFSS), the national Centers for Disease Control



Lessons from Political Science 107

and Prevention (CDC) conducts ongoing national surveys of adult health (CDC 2006). The self-­reported health status question included in the CMPS survey is similar to that used in the BRFSS, with both questions using a 1 to 5 Likert scale (with respondents rating their health status from excellent to poor). More importantly, studies have found that the overall BRFSS questionnaire produces reliable and valid results (CDC 2010). For example, Nelson et al. (2001) produced a comprehensive review of BRFSS reliability and validity studies and determined that most items on the main BRFSS survey are at least moderately, and many are highly, reliable and valid.5 We again believe that racial self-­identification may not provide a complete conceptualization of the racial and ethnic impact on health, so we explore racial identity in more detail through the CMPS. More specifically, we use measures of experiences with discrimination, perceptions of the need to “play down” one’s racial identity to succeed in the United States, and self-­rated skin color. We employ a five-­point measure of respondent’s self-­defined skin color, with options ranging from “very light” to “very dark” to assess the potential impact of skin color on health status. Skin color is an important measure of race and ethnicity, as this has been a measure of interest to those interested in racial and ethnic differences in health outcomes for some time. We approach the concept of skin color in line with Gravlee et al. (2005), who contend that skin color may be a key cultural, and not biological, variable of interest. Furthermore, our approach of using self-­reported skin color is consistent with the work of others (Gravlee et al. 2005; Klonoff and Landrine 2000), who have found both ascribed and self-­reported skin color to be correlated with health status. This is in contrast to those who have used reflectometers to assess skin color and have found this conceptualization of skin color to have no relationship with health status (Borrell et al. 2006; Gravlee et al. 2005; Krieger et al. 1998). We also include a measure of whether respondents had reported being discriminated against or treated unfairly due to their race or ethnicity over the past year. This measure is very similar to items identified in Paradies’s (2006) review of the extant work focused on self-­rated discrimination and health status. Although scholars have explored the potential relationship between African American experiences with discrimination and health for some time, there is much more limited work exploring this relationship among Latinos (but see Finch et al. 2001) and Asian Americans (but see Gee et al. 2007). We are therefore among the first to explore the relationship between discrimination and health across multiple groups. Finally, we incorporate a measure that assesses how much respondents believe that they have to “play down their racial identity to get ahead in

108

Gabriel R. Sanchez and Vickie D. Ybarra

American society.” This variable is intended to gauge the potential impact of racial/ethnic identity, which has been found to have significant health implications for minority populations. For example, Viruell-­Fuentes et al. (2011) find in their qualitative interviews of Latinas that the stress associated with “negotiating ethnic identities under stigmatizing environments” could be an explanation for the relationship between Latina ethnic status and poorer health outcomes. Our measure, which asks respondents directly whether they feel the need to play down their racial identity, provides some leverage on this question of whether the stigma associated with a racialized environment has health consequences. While not exhaustive, these additional measures add significantly to the self-­defined sample that we employ in the analysis, and as noted earlier, are all identified as relevant to health status in the extant literature. In addition, we include a four-­point acculturation scale that combines language preference and nativity with values ranging from foreign-­born and Spanish dominant to U.S.-­born and English dominant. We operationalize political factors through a cluster of variables: trust in government, voting, political participation (scale), mobilization and political knowledge. Both voting and mobilization are dichotomous variables, with voted in 2008 and contacted by a political party, candidate, or other interest group in 2008 as the positive value for these two variables. The political participation measure is a scale ranging from 0 to 8 based on the number of political activities that respondents engaged in during the 2008 election (contacted an official, donated money, attended a rally, signed a petition, and so forth). Our trust measure is based on responses to a question focused on how often respondents “trust the government in DC to do what is right,” with responses ranging from “never” to “just about always.” Finally, our political knowledge variable is based on the ability of respondents to identify the chief justice of the Supreme Court and the political party that had the most seats in the U.S. House of Representatives in 2008.6 The Impact of Race/Ethnicity on Health Status

We begin our discussion of Latino self-­defined health status with descriptive statistics to analyze the overall health status of Latinos relative to other racial and ethnic groups, as well as to consider variation in health status within the Latino population. We start with the common approach of using self-­reported race and ethnicity to explore potential differences across groups. When we compare the frequencies of our dependent variable across groups, we find that 86.4 percent of Whites and 84.4 percent of Asians reported health status in the good/ very good/excellent range, versus 81.7 percent of Blacks and only 78.4 percent



Lessons from Political Science 109

of Latinos. We therefore find that Latinos have lower levels of self-­rated health than other groups, which is consistent with previous research.7 However, additional exploration is needed to move beyond this limited treatment of race and ethnicity. When we explore the relationship between our multidimensional measures of race and ethnicity and health status, we see some interesting preliminary support for our argument that moving beyond dummy variables to capture race is essential. For example, we find that 78 percent of respondents of color who have experienced discrimination within the past year report having excellent, very good, or good health, compared to 84 percent who have not been discriminated against. There is a similar gap in health status due to skin color, as our health status measure is sensitive to the self-­reported skin color of Latino, Black, and Asian respondents. More specifically, the mean on the health measure for those respondents who indicated that they are “very dark” is 3.30 on the scale from 1 (poor health) to 5 (excellent health). Self-­reported health improves with lighter self-­reported skin color, however, increasing to 3.47 for those with “dark” skin, 3.63 for those with “light” skin, and nearly 4 for those with “very light” skin. These findings suggest that these alternative measures to race and ethnicity are important and enrich our understanding of racial and ethnic health disparities. Given our focus on exploring internal group variation, we also explore the impact of language use and nativity among Latinos. Here we see significant differences, with foreign-­born Latinos and those who are Spanish dominant having poorer health than their more acculturated counterparts. This provides initial evidence for our argument that the exploration of within-­group variation is critical, particularly for diverse populations such as Latinos. Given these trends, we created an acculturation scale that combines these two variables in our analysis to gauge the impact of acculturation on health outcomes. We now turn our attention to the potential impact of our multidimensional measures for race relative to other controls, including our measures for political context in the non-­White sample. Here we organize our measures according to theoretically related factors: political factors, socioeconomic/demographic factors, and racialization factors. Finally, in line with our desire for internal group analysis, we isolate the Latino and Mexican-­origin respondents in split-­sample models to examine whether political and other factors have a unique impact on health status for these more specific populations. Beginning with the full sample results depicted in table 7.1, we see that our argument for the expansion of approaches to capturing race and ethnicity is supported. More specifically, both skin color and discrimination experiences are negatively correlated with

110

Gabriel R. Sanchez and Vickie D. Ybarra

self-­reported health status. We find that non-­White respondents who report having darker skin color and report being discriminated against have poorer health than minority respondents who report having lighter skin color and who did not report having been discriminated against. Although playing down one’s racial identity is not statistically significant, the relevance of skin color and discrimination in our model that isolates self-­identified Latino, Asian, and African American respondents supports the more expansive approach to measuring race. It is important to note that these results hold even when controlling for several other relevant factors. For example, among the political factors specified in our model, both trust in government and political participation are positively correlated with health status. These results suggest that political factors are meaningful and, therefore, should be explored in more depth in the future. Finally, both educational attainment and gender are significant as well, indicating that minority respondents who are more highly educated and who are male have better health than female and less educated respondents. The final segment of our analysis focuses on the split-­sample models where we explore the impact of political factors within the Latino and Mexican-­origin samples. Our argument for within-­group modeling is strongly supported by the behavior of the remaining variables, as many factors are relevant for Latinos that were not significant in the full model. The Latino model indicates that socioeconomic status has a major influence on Latino health status, as both education and income are positively associated with health. Acculturation is significant in this model as well, with more acculturated Latinos having better health—­though this variable was not significant in the full model. Most important to our focus on race and ethnicity, among racial factors discrimination matters for Latinos as well, with respondents who report having experienced discrimination in the past year having worse health than those who have not. There is a similar trend within the Mexican-­origin model as well, highlighting the important role that discrimination plays in health among the Latino population. The consistent relevance of discrimination experiences across all models strongly implies that scholars interested in exploring racial and ethnic health disparities must account for this important dimension to race/ethnicity in future research. Skin color is not significant in the split-­sample models, however, providing further evidence that moving beyond a monolithic approach to assessing racial/ethnic health disparities leads to a different set of results and conclusions. If we were to have closed our analysis prior to conducting the Latino and Mexican-­origin specific models, we would have concluded that skin color influences health outcomes for all non-­White communities. Yet that would have been a flawed conclusion since skin color

Table 7.1   Determinants of Health Status among Full/Latino Sample (Ordered Logistic Regression) Full Minority

Explanatory Variable

Mexican Sample

Sample

Latino Sample

Coefficient

Coefficient

Coefficient

(standard error)

(standard error)

(standard error)

Political Factors

Efficacy (trust government)

.136** (.046)

-­.023 (.065)

.036 (.097)

Vote (2008)

-­.011 (.139)

.351** (.165)

.346* (.198)

Participation scale

.097*** (.016)

.036 (.025)

.046 (.034)

Mobilization (mobilized)

.093 (.072)

-­.205** (.107)

-­.364** (.142)

Political knowledge

.048 (.075)

-­.023 (.107)

.336** (.141)

Education (some college)

.501*** (.091)

.388*** (.115)

.193 (.154)

Education (college grad)

.854*** (.087)

.848*** (.147)

.822*** (.212)

Income

.022 (.018)

.047* (.028)

.085** (.039)

Acculturation scale

-­.015 (.030)

.088** (.041)

.046 (.056)

Gender (male)

.189** (.069)

.066 (.107)

-­.146 (.138)

Latino (full) Mexican (all Latino)

-­.029 (.126)

-­.093 (.101)

.

Asian

.084 (.179)

.

.

Black

-­.128 (.102)

.

.

Demographic/SES Factors

Racial/Immigration Context Factors

Skin color

-­.100** (.036)

-­.008 (.053)

-­.060 (.075)

Discrimination experiences

-­.175** (.070)

-­.204** (.101)

-­.293** (.136)

Play down racial identity

-­.098 (.078)

-­.129 (.127)

-­.084 (.180)

N = 2837

N = 1280

N = 642

*< .01**p < .05 ***p < .001

112

Gabriel R. Sanchez and Vickie D. Ybarra

does not appear to be correlated with health status for Latinos when other factors, including discrimination experiences, are included. Furthermore, while political factors are once again meaningful, there are some important distinctions. Specifically, voting and mobilization are significant in both the Latino and Mexican specific models, while trust in government, which was relevant in the full model, fails to reach statistical significance in the split-­sample models. Finally, for the most part, the same factors that were relevant in the full Latino model (education, income, and discrimination experiences) remained significant for Mexican Americans. However, we find that acculturation is not significant in the Mexican-­specific model. Although the exploration of this trend is beyond the scope of this chapter, this is something that scholars interested in immigration and health should explore further given the implications for the vast immigrant health paradox literature. In short, the immigrant health paradox suggests that foreign-­born Latinos exhibit better health than those born in the United States, particularly in regard to lifestyle and behavior patterns. Research within this literature finds that more acculturated Latinos, often measured by language use and nativity, have higher alcohol consumption, smoking, and body mass index, among other health risk factors (see Lara et al. 2005 for a review of this work). Our results suggest that acculturation may not be a strong predictor of self-­rated health status among specifically Mexican Americans. Implications for Future Health Disparities Research

Our paper is an illustration of the potential for greater knowledge generation through more comprehensive approaches to measuring race and ethnicity and the use of in-­group analysis, particularly for diverse groups such as the Latino population. By exploring the relationship between skin color, discrimination experiences, racial identity, and health status as a dependent variable in the same research design, we have essentially tested the primary argument of this volume—­that richer and more nuanced approaches to measuring race will lead to better science and policy foundations. We advance this theory by finding that experiences with discrimination and skin color impact self-­rated health status for non-­White Americans, even after controlling for a host of other factors. Furthermore, our split-­sample approach to exploring factors that predict differences in health status among Latinos reaffirms that minority groups are not monolithic communities. Within the context of the measurement of race, although skin color is significantly correlated with health status when we explore all populations of color, it is not significant in the Latino and Mexican-­origin specific models. On the other hand, acculturation proved to be a meaningful predictor of self-­defined health status for Latinos while it lacked



Lessons from Political Science 113

statistical significance for the more general minority population. Thus, if we had not moved forward to a research design intended to uncover in-­group differences in health outcomes, we would be left with the conclusion that acculturation is not relevant to health status—­a misguided interpretation of our data after we look at our overall findings. We therefore suggest that scholars look to explore internal variation among other communities to assess what explains the racial and ethnic disparities apparent from research focused on exploring differences across groups. Many chapters in this volume provide some guidance on how to advance our study of racial and ethnic health disparities with existing data sets (see Garcia, this volume, and Saperstein, this volume). This should include identifying data sets with large enough sample sizes to explore within-­group differences. Finally, our efforts here have provided preliminary evidence that political factors such as trust in government and political participation have a meaningful impact on self-­defined health status. Given that there are striking similarities between discussions of racial and ethnic disparities within the political and health systems, it is not surprising to see that there are significant correlations between these two phenomena in the United States. We believe that this is an area where political scientists are poised to make major contributions to the social determinants of health literature, and we invite others to expand on these initial findings. For example, by providing a model of the utility of in-­ group analyses, scholars might better explore internal variation within groups to better understand the complex effects of race and political status on health status. Moreover, by teasing out the impact of racialization on health status and other outcome variables, we can better explore the utility of using multiple measures of racial and ethnic identification in social science research. Notes 1. Due to space limitations, we are not able to provide a theoretical justification for the measurement of the remaining political factors we employ in our analysis: mobilization, political knowledge, and political participation (both voting and nonvoting). However, each of these factors captures a well-­established political science concept that has also been robustly shown to be affected by race and ethnicity. 2. The CMPS survey was available in English, Spanish, Mandarin, Cantonese, Korean, and Vietnamese and respondents were offered the opportunity to interview in their language of choice. 3. The most significant limitation of the CMPS data is that it is restricted to registered voters, and consequently not inclusive of noncitizens. We compared frequencies in health status between the CMPS and BRFSS, and we find that CMPS respondents report slightly better health than those from the 2007 BRFSS given the differences in sampling: BRFSS includes noncitizens. However, we contend that the nature of the sample effectively controls for the role of citizenship status, as any differences

114









Gabriel R. Sanchez and Vickie D. Ybarra

in health status due to racial factors should be considered robust given that we are only examining the health of Americans who have at a minimum gone through the important step of becoming a citizen if they were not born in the United States. 4. We have reordered the categories in the dependent variable for this study so that higher values reflect better health: 5 = excellent, 4 = very good, 3 = good, 2 = fair, and 1 = poor. 5. Self-­reported health status, as measured by the CDC BRFSS, has been especially well-­studied in relation to mortality. In their review of twenty-­seven studies including international and U.S. contexts, Idler and Benyamini report “impressively consistent findings” in self-­rated health status as an independent predictor of mortality (1997). Self-­reported health status has also been found to be associated with a variety of health behaviors and health status indicators including physician-­rated health status, smoking behavior, alcohol use, healthy eating, physical activity, healthy days, diabetes-­related complications, and cardiovascular disease (Mossey and Shapiro 1982; Rubin and Peyrot 1999; Tsai et al. 2010a; Tsai et al. 2010b; Zullig and Hendryx 2010). 6. The rest of our control variables are straightforward, but please see the following link for the toplines to the CMPS data set: http://cmpstudy.com/assets/CMPS-toplines .pdf. 7. The CMPS Asian weighted sample reports 15.6 percent in fair/poor health, while the 2004–­2006 BRFSS non-­disability sample reports only 8.1 percent in fair/poor health. We are exploring potential issues with the Asian sample that may be leading to unusually high rates of self-­defined poorer health status in our data.

References Agency for Healthcare Research and Quality (AHRQ). 2009. National Healthcare Disparities Report: 2008. AHRQ Publication No. 09–­002. Rockville, MD: U.S. Department of Health and Human Services. Berenson, Gerald S., Wendy A. Wattigney, and Larry S. Webber. 1996. “Epidemiology of Hypertension from Childhood to Young Adulthood in Black, White, and Hispanic Population Samples.” Public Health Reports 3 (suppl. 2): 3–­6. Borrell, Luisa, N. Catarina I. Kiefe, David R. Williams, Ana V. Diez-­Roux, and Penny Gordon-­Larsen. 2006. “Self-­Reported Health, Perceived Racial Discrimination, and Skin Color in African Americans in the CARDIA Study.” Social Science and Medicine 63: 1415–­27. Borrell, Luisa N., David R. Jacobs Jr., David R. Williams, Mark J. Pletcher, Thomas K. Houston, and Catarina I. Kiefe. 2007. “Self-­Reported Racial Discrimination and Substance Use in the Coronary Artery Risk Development in Adults Study.” American Journal of Epidemiology 166 (9): 1068–­79. Brondolo, E., Brady ver Halen, Melissa Pencile, Danielle Beatty, and Richard Contrada. 2009. “Coping with Racism: A Selective Review of the Literature and a Theoretical and Methodological Critique.” Journal of Behavioral Medicine 32: 64–­88. Centers for Disease Control and Prevention (CDC). 2006. BRFSS Operational and User’s Guide: Version 3.0. Accessed July 8, 2012. ftp://ftp.cdc.gov/pub/data/brfss/userguide .pdf. ———. 2010. BRFSS Data Quality, Validity, and Reliability: BRFSS Data Quality and National Estimates. Accessed July 8, 2012. http://www.cdc.gov/brfss/pubs/quality.htm. Collins, K., A. Hall, and C. Neuhaus. 1999. U.S. Minority Health: A Chartbook. New York: Commonwealth Fund.



Lessons from Political Science 115

Finch, Brian Karl, Robert A. Hummer, Bohdan Kolody, and William A. Vega. 2001. “The Role of Discrimination and Acculturative Stress on the Physical Health of Mexican-­ Origin Adults.” Hispanic Journal of Behavioral Sciences 23 (November): 399–­429. Gee, Gilbert C., Michael S. Spencer, Juan Chen, and David Takeuchi. 2007. “A Nationwide Study of Discrimination and Chronic Health Conditions among Asian Americans.” American Journal of Public Health 97: 1275–­82. Gornick, Marian. 2000. “Disparities in Medicare Services: Potential Causes, Plausible Explanations, and Recommendations.” Health Care Financing Review 21 (Summer): 23–­44. Gravlee, Clarence C., William W. Dressler, and H. Russell Bernard. 2005. “Skin Color, Social Classification, and Blood Pressure in Southeastern Puerto Rico.” American Journal of Public Health (December): 2191–­97. Idler, Ellen L., and Yael Benyamini. 1997. “Self-­Rated Health and Mortality: A Review of Twenty-­ seven Community Studies.” Journal of Health and Social Behavior 38 (March): 21–­37. Iwamoto, D., Liu, and William Ming. 2010. “The Impact of Racial Identity, Ethnic Identity, Asian Values, and Race-­Related Stress on Asian Americans and Asian International College Students’ Psycholgical Well-­Being.” Journal of Counseling Psychology 57 (1): 79–­91. James, Cara, Megan Thomas, and Marsha Lillie-­Blanton. 2007. “Key Facts: Race, Ethnicity, and Medical Care.” Henry J. Kaiser Family Foundation. Accessed July 8, 2012. http://www.kff.org/minorityhealth/upload/6069–02.pdf. Klonoff, Elizabeth, and Hope Landrine. 2000. “Is Skin Color a Marker for Racial Discrimination? Explaining the Skin Color–­Hypertension Relationship.” Journal of Behavioral Medicine 23: 329–­38. Krieger, N., S. Sidney, and E. Coakley. 1998. “Racial Discrimination and Skin Color in the CARDIA Study: Implications for Public Health Research.” American Journal of Public Health 88: 1308–­13. Lara, Marielena, Christina Gamboa, M. Iya Kahramanian, Leo S. Morales, and David Hayes Bautista. 2005. “Acculturation and Latino Health in the United States: A Review of the Literature and Its Sociopolitical Context.” Annual Review of Public Health 26: 367–­97. LaVeist, Thomas A. 2000. “On the Study of Race, Racism, and Health: A Shift from Description to Explanation.” International Journal of Health Services 30: 217–­19. Liu, A. 2001. “Political Participation and Dissatisfaction with Democracy: A Comparative Study of New and Stable Democracies.” In Research Methods Working Paper Series. Los Angeles: Research Methods Institute. McGuire, Thomas G., and Jeanne Miranda. 2008. “New Evidence Regarding Racial and Ethnic Disparities in Mental Health: Policy Implications.” Health Affairs 27 (March/ April): 393–­403. Mossey, Jana M., and Evelyn Shapiro. 1982. “Self-­Rated Health: A Predictor of Mortality among the Elderly.” American Journal of Public Health 72 (August): 800–­808. Nelson, David E., D. Holtzman, J. Bolen, C. A. Stanwyck, and K. A. Mack. 2001. “Reliability and Validity of Measures from the Behavioral Risk Factor Surveillance System (BRFSS).” Social and Preventive Medicine 46 (suppl. 1): S03–­S42. Paradies, Y. 2006. “A Systematic Review of Emperical Research on Self-­Reported Racism and Health.” International Journal of Epidemiology 35: 888–­901. Rubin, Richard R., and Mark Peyrot. 1999. “Quality of Life and Diabetes.” Diabetes/ Metabolism Research and Reviews 15 (May-­June): 205–­18.

116

Gabriel R. Sanchez and Vickie D. Ybarra

Sanchez, Gabriel R. 2006. “The Role of Group Consciousness in Latino Public Opinion.” Political Research Quarterly 59 (3): 435–­46. Sanchez, Gabriel R., and Natalie Masouka. 2010. “Brown Utility Heuristic? The Presence and Contributing Factors of Latino Linked Fate.” Hispanic Journal of Behavioral Sciences 32 (4): 519–­31. Satcher, David, George E. Fryer Jr., Jessica McCann, Adewale Troutman, Steven H. Woolf, and George Rust. 2005. “What If We Were Equal? A Comparison of the Black-­White Mortality Gap in 1960 and 2000.” Health Affairs 24 (March/April): 459–­64. Schulz, Amy J., Clarence C. Gravlee, David R. Williams, Barbara A. Israel, Graciela Mentz, and Zachary Rowe. 2006. “Discrimination, Symptoms of Depression, and Self-­ Rated Health among African American Women in Detroit: Results from a Longitudinal Analysis.” American Journal of Public Health 96 (7): 1265–­70. Smedley, Brian D., Adrienne Y. Stith, and Alan R. Nelson, eds. 2003. Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington, DC: National Academies Press. Sternthal, Michelle J., Natalie Slop, and David R. Williams. 2011. “Racial Disparities in Health: How Much Does Stress Really Matter?” Du Bois Review: Social Science Research on Race 8 (1): 95–­113. Tsai, James, Earl S. Ford, Chaoyang Li, Guixiang Zhao, and Lina S. Balluz. 2010a. “Physical Activities and Optimal Self-­Rated Health of Adults with and without Diabetes.” BMC Public Health 10: 365–­73. Tsai, James, Earl S. Ford, Chaoyang Li, Guixiang Zhao, William S. Pearson, and Lina S. Balluz. 2010b. “Multiple Healthy Behaviors and Optimal Self-­Rated Health: Findings from the 2007 Behavioral Risk Factor Surveillance System Survey.” Preventive Medicine 51: 268–­74. Verba, S., K. Schlozman, and H. Brady. 1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press. Viruell-­Fuentes, Edna A., Jeffrey D. Morenoff, David R. Williams, and James S. House. 2011. “Language of Interview, Self-­Rated Health, and the Other Latino Health Puzzles.” American Journal of Public Health 101 (7): 1306–­13. Waidmann, Timothy A., and Shruti Rajan. 2000. “Race and Ethnic Disparities in Health Care Access and Utilization: An Examination of State Variation.” Medical Care Research and Review 57: 55–­84. Williams, David R. 1996. “Racism and Health: A Research Agenda.” Ethnicity and Disease 6 (Winter-­Spring): 1–­8. ———. 1999. “Race, Socioeconomic Status, and Health: The Added Effects of Racism and Discrimination.” Annals New York Academy of Sciences: 173–­88. Woo, Meghan, S. Bryn Austin, David R. Williams, and Gary G. Bennett. 2011. “Reconceptualizing the Measurement of Multiracial Status for Health Research in the United States.” Du Bois Review: Social Science Research on Race 8 (1): 25–­36. Zullig, Keith J., and Michael Hendryx. 2010. “A Comparative Analysis of Health-­related Quality of Life for Residents of U.S. Counties with and without Coal Mining.” Public Health Reports 125 (July-­August): 548–­55.

Derek Kenji Iwamoto, Mai M. Kindaichi, Chapter 8

and Matthew Miller

Advancing Asian American Mental Health Research by Enhancing Racial Identity Measures

Asian Americans are often racialized by society through the minimization of Asian ethnic group variations and the exaggeration of similarities across these ethnic groups (Liu and Iwamoto 2006). As a result, diverse people with a particular phenotype are categorized racially as “Asian” regardless of various ethnic differences (Iwamoto and Liu 2010). Terminology for race, ethnicity, ethnic identity, and culture are often conflated or used interchangeably to lump people of various Asian ethnicities and cultural groups into one homogenous group: the racial category “Asian American” is comprised of at least thirty different ethnic groups and potentially many more cultural groups (U.S. Census Bureau 2011). In this chapter, the term Asian American will be used to reflect the “racialization” of Asian Americans whereupon ethnic variations within this group are often minimized and similarities are exaggerated (for example, “all Asians look alike”). As this is one of the fastest growing racial groups in the United States (U.S. Census 2010), there is a pressing need to understand how this racialization process affects Asian Americans’ physical and mental health. Racial identity development—­the process of how “members of racially oppressed groups respond to and internalize race-­related stress and discrimination into their overall identity or self-­consciousness” (Alvarez and Helms 2001, 218)—­is one dynamic way to conceptualize race and helps explain within-­ group differences among Asian Americans. Racial identity development differs from ethnic identity development, in which ethnic identity represents an individual’s subscription to and maintenance of ethnic group cultural practices (Cokley 2005). In contrast, racial identity captures an individual’s level of awareness of racism and connection with their own racial group and orientation

117

118

D. K. Iwamoto, M. M. Kindaichi, and M. Miller

toward the dominant racial group (Whites). Furthermore, whereas many quantitative analyses operationalize racial identity merely and/or exclusively via the subject’s racial self-­identification, racial identity development provides a much wider lens through which to examine how racism and racialization may impact physical and mental health outcomes among people of Asian descent in the United States. This study advances this literature by refining the People of Color Racial Identity Attitudes Scale (PRIAS), a psychological research instrument designed to capture an individual’s racial identity attitudes that was pioneered by psychologist Janet Helms (1995a). With some exceptions that we will discuss, this scale has been primarily based on studies of African Americans; this chapter presents data from two new studies with Asian American samples, thereby refining the model and advancing the literature in important ways. Racial Identity Development

Racial identity development is typically conceptualized as occurring via a number of distinct statuses. According to Helms, racial identity statuses are reflective of the “the dynamic cognitive, emotional, and behavioral processes that govern a person’s interpretation of racial information in her or his interpersonal environments” (1995a, 184). The term “status” was adopted in contrast to “stage,” which implies a progressing, unidirectional sequence of adaptive racial identity development; “status,” on the other hand, communicates that an individual can vary in their experience of race and awareness of racism. Helms (1995a) identified multiple statuses that characterize how individuals perceive and interact with their environment, including conformity, dissonance, immersion/emersion, and internalization. Individuals operating from conformity tend to value “Whiteness” at the expense of their own racial group and seem to show a lack of awareness of the social-­political histories of people of color. Individuals operating from dissonance have become aware of the existence of racial differences, and they struggle to process anxiety-­provoking racial information in a meaningful way on account of their ambivalence between their identification with their own racial group and the dominant group. An individual who endorses immersion/emersion is fully cognizant of the negative impact of racism and, as a result, detaches from White standards of merit in favor of cultural values associated with one’s minority racial group. The individual also harbors some resentment and resistant attitudes toward Whites and White culture. Finally, individuals with internalization are able to respond objectively to racial messages and show positive commitment to their own social-­racial group. Beyond the demographic variable of self-­identified race, racial identity development represents a psychological lens through which individuals’



Asian American Mental Health Research 119

process racial messages cognitively, affectively, and behaviorally. As a result, racial identity development provides a more nuanced perspective through which physical and mental health outcomes among people of Asian descent can be understood and researched. Emerging research among Asian Americans has demonstrated that racial identity development is associated strongly with perceptions of racism (Alvarez and Helms 2001; Chen et al. 2006) and psychological adjustment and well-­ being (Alvarez and Helms 2001; Iwamoto and Liu 2010). Chen and colleagues (2006) examined the association between racial identity, racism-­related stress (Liang et al. 2004), and color-­blind racial attitudes (for example, beliefs that racism should not and does not exist; Neville et al. 2000) among Asian American young adults. The results suggest that individuals who endorsed internalization tended to report less racism-­related stress and minimized the effects of institutional racism and racial privilege (Chen et al. 2006). Individuals who endorsed conformity attitudes, specifically those who aligned with White cultural standards, also endorsed minimization of racism and low race-­related stress. On the other hand, individuals who endorsed dissonance tended to report more stress and higher levels of perceived racism. Lastly, individuals who reported heightened awareness of the negative effects of racism and affiliated with the sociocultural values of their minority group (for example, immersion/emersion) also reported high racism-­related stress and low minimization of present-­day racism. These findings suggest that individuals who endorse statuses of dissonance and immersion share a similar sensitization to the personal impact of racism such that they experience more pointed levels of stress that may have physical and mental health significance. While Chen and colleagues (2006) investigated the link between racial identity and perceptions of racism, Iwamoto and Liu (2010) examined the relationships between racial identity, ethnic identity, endorsement of Asian values, and psychological well-­being (including an individual’s openness to growth and personal expressiveness, self-­actualization, and the feeling of purpose and meaning in life) among diverse university students of Asian descent. They found that higher internalization was positively related to psychological well-­being, whereas racial identity attitudes related to racial conflict (for example, dissonance and immersion/emersion attitudes) were negatively related to psychological well-­being. They also reported that respondents who indicated high amounts of race-­related stress reported higher psychological well-­being if they had high endorsement of Asian cultural values along with lower conformity and dissonance attitudes. In effect, this study implies that people of Asian descent who identify more strongly with White cultural norms and

120

D. K. Iwamoto, M. M. Kindaichi, and M. Miller

expectations may not recognize or may internalize race-­related incidents in such a manner as to influence their psychological states negatively. These findings suggest that racial identity development represents a crucial component of identity and mental health for Asian Americans, which necessitates greater research to examine how racial identity may impact other health outcomes and health disparities experienced by this population. Although the current research on racial identity experiences is promising, there is a paucity of research in this area despite recent calls to investigate the measurement of racial identity and its influence on Asian Americans’ mental health (Ponterotto and Park 2007). In fact, much of the extant racial identity assessments have been based on the racial identity development of African Americans, and the few racial identity assessments designed specifically for Asian Americans have significant measurement problems (Alvarez, Juang, and Liang 2006; Chen et al. 2006; Iwamoto and Liu 2010). To date, only one published factor analytic study (Perry, Vance, and Helms 2009)1 has examined the measurement of racial identity development among Asian Americans. Perry et al.’s study provided empirical support for the four-­factor model of racial identity: conformity, dissonance, immersion/emersion, and internalization. However, many of the items from the original PRIAS did not effectively assess racial identity, and the original surveys still include a large number of items (50), which detracts from the feasibility of this important measure in epidemiological studies. This current investigation extends the Asian American racial identity literature by examining the factor structure of the PRIAS among two samples of Asian Americans: a young adult and a community sample. Specifically, the current study advances the racial identity literature by (1) improving the measurement of the PRIAS through factor analytic procedures; (2) streamlining and reducing the number of items in order to enhance its feasibility in epidemiological studies; (3) providing empirical support of the predictive validity of the revised measure; and (4) providing construct validity evidence for the theoretical factor structure among more diverse samples of Asian Americans (a young adult, university sample, and community sample). Refining the PRIAS in Two New Studies

Two hundred and eighty Asian American young adults participated in the first study conducted in March 2008. The sample included 64.3 percent women, and the average age of the participants was 21.17 years old (SD = 3.9). In terms of generation, 68 percent were second generation (first born in the United States), 26.4 percent were first generation, and 5.6 percent were third generation or higher. The majority of the participants identified as Chinese American, Korean



Asian American Mental Health Research 121

American, or Vietnamese American. The sample was recruited from a large public university located in Southern California. More information about the sample characteristics and recruitment can be founded in Iwamoto and Liu (2010). The PRIAS (Helms 1995a) is based on the Minority Identity Development Model developed by Atkinson, Morton, and Sue (1989), which includes the racial identity statuses (1) conformity, (2) dissonance, (3) immersion/emersion, (4) introspection, and (5) internalization. The PRIAS is a fifty-­item measure that uses responses of participants to items on a five-­point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). The PRIAS consists of four subscales that make up the entire measure and reflect the racial identity factors (conformity, dissonance, immersion/emersion, and internalization), which slightly differ from the Minority Identity Development Model. In the PRIAS, the internalization subscale included components of introspection and internalization schemas. The Scales of Psychological Well-­ Being (SPWB) variable is based on Ryff’s conceptualization (1989) of eudaimonic (that is, positive) psychological well-­being and consists of six dimensions: positive relationships with others, autonomy, environmental mastery, personal growth, purpose of life, and self-­ acceptance. The present study used the shortened version of the SPWB, which includes three items per subscale, and all of the subscales were summated to create a composite score of psychological well-­being. Because we were interested in examining the underlying structure of racial identity and streamlining the PRIAS by reducing the number of items, we conducted a series of exploratory factor analyses (Worthington and Whittaker 2006), examining three-­, four-­, and five-­factor models that were performed. We used restrictive inclusion criteria given that we were interested in retaining the strongest factor loading, which is important for replicating results. Accordingly, items with factor loadings lower than .40 (for example, factor loading above .40 is desirable) and items that measured or were highly correlated with multiple factors (for example, cross-­loaded) were removed. Based on the aforementioned criteria, the four-­factor model was retained, which was consistent with the theoretical factors described by Helms (1995a, b). Using exploratory factor analysis procedures and evaluating items that cross-­loaded or had a loading of less than .40, a number of items were removed since they did not measure racial identity well. Thus the original PRIAS items were reduced from fifty to thirty-­two, which resulted in the empirically derived PRIAS-­32. A confirmatory factor analysis (CFA) was conducted to examine the model fit of the PRIAS-­32 (for example, the thirty-­two-­item, four-­factor PRIAS). Based on the results, some of the factor loadings in the PRIAS-­32 were relatively low or below .50, thus a model with only the four highest per-­factor loadings—­that

122

D. K. Iwamoto, M. M. Kindaichi, and M. Miller

is, the four best items that measured each factor most optimally—­was performed, with the exception of the dissonance factor, in which only three of the four items met the factor loading of .50 or higher. Inspection of the statistical tests of model fit suggested modifying the original PRIAS model to one that includes four items per factor and three items for the dissonance factor to create a more parsimonious alternative named the PRIAS-­15 (table 8.1). Finally, since we were interested in further streamlining the PRIAS, we also tested the model fit of the PRIAS-­12, which consisted of three items per factor (Kline 2005). Based on the CFA, the results suggested that the three-­item-­per-­factor model assessed each racial identity factor well (table 8.1). Table 8.1   PRIAS-­1 5 and PRIAS-­1 2 Items Derived from Confirmatory Factor Analysis on the Original PRIAS Factor Item Conformity

4. Embarrassed to be Asian. 6. Whites are more attractive. 7. Should act more like Whites. 8. Engage only in White activities. Immersion/Emersion

10. Unable to involve self in White culture so only involve self with Asian culture. 14. Whites are untrustworthy. 45. In social situation, I prefer to be around Asians more than Whites. 49. I get angry when I see Asians act like Whites. Internalization

22. All races have limitations and strengths. 25. Asians and White should learn from each other. 26. I enjoy some White customs. 28. All races have good and bad people. Dissonance

31. I don’t know which race I belong to. 33. I now question what I believe in. 40. I am unsure about myself. Source: Helms 1995b. Note: Items are abbreviated items from Helms (1995b). Bold: Indicates items just from the PRIAS­12. The PRIAS-­15 includes all of the items listed above.



Asian American Mental Health Research 123

We next performed three regression analyses to investigate the extent to which the original PRIAS, PRIAS-­15, and PRIAS-­12 factors are associated with psychological well-­being. Specifically, we were interested in investigating whether the main effects and R-­square would significantly differ between the original fifty-­item PRIAS (Helms 1995a) and the PRIAS-­15 and PRIAS-­12. The first regression using the original PRIAS factors revealed that dissonance and immersion attitudes were related negatively with psychological well-­being, while internalization attitudes were related positively (see table 8.2). In the second regression model, using the PRIAS-­15, the results differed from the previous model: specifically dissonance attitudes were most strongly negatively associated with well-­being. Additionally, internalization attitudes were positively related to well-­being, and less strongly, immersion yielded a negative relationship (see table 8.2). In contrast, the results of third regression model, using the PRIAS-­12, indicated that all of the PRIAS factors were associated significantly with well-­being. Contrary to the first two analyses, conformity attitudes were related negatively to psychological well-­being. Dissonance attitudes remained the strongest negative factor associated with well-­being; additionally, internalization attitudes related positively to well-­being, while immersion attitudes yielded a negative relationship. Table 8.2 includes the statistics for all three regression models. Collectively, all three versions of the PRIAS factor structures explained a large percentage of the variance (range from 28 percent to 30 percent) in psychological well-­being. In addition, the more streamlined versions of the original PRIAS, namely the PRIAS-­12 and PRIAS-­15, operated very similarly in terms of the significant associations between the racial identity factors and well-­being, as well as explaining the amount of variance attributed to each racial identity factor. Finally, the conformity factor in the PRIAS-­12 emerged as a significant predictor of well-­being. Table 8.2   Multiple Regression Using the PRIAS, PRIAS-­1 5, and PRIAS-­1 2 Measures as Predictors and Psychological Well-­B eing as the Outcome Three Versions of the PRIAS PRIAS Variables

Conformity

β

.00

PRIAS-­15

β

PRIAS-­12

β

-­.10

-­.13*

Dissonance

-­.16**

-­.35**

-­.35**

Immersion-­Emersion

-­.16**

-­.15***

-­.18**

Internalization

.46***

.26***

.27**

R²

30%

29%

26%

*p < .05, **p < .01, ***p < .001

124

D. K. Iwamoto, M. M. Kindaichi, and M. Miller

Given the strong empirical support for the PRIAS-­12 and PRIAS-­15, the second study examined the extent to which the factor structure derived from the college sample would operate similarly for the non-­college, community sample of Asian Americans. One hundred and ninety-­eight Asian Americans were recruited in May 2008 via a web-­based survey that was posted on two popular Asian American blogs and listservs (http://blog.angryasianman.com; http://www.asian-nation.org/). The webmasters for the two blogs were provided details of the study and agreed to post the link on their sites. The average age of the community sample was twenty-nine years (SD = 6.4), where 54 percent were men and 59 percent of the participants were second-generation (that is, the first generation born in the United States). The majority of the participants identified as Chinese American, Korean American, Filipino, and Vietnamese Americans. In order to attract participants, they were provided the option to enter a raffle to win one of six fifty-dollar gift certificates. In order to assess the validity of the more streamlined and revised PRIAS measures, we conducted a series of confirmatory factor analyses (CFAs) using the community sample of Asian Americans. The results suggest generally that the revised PRIAS (for example, PRIAS-­32, PRIAS-­15, and PRIAS-­12) measured racial identity well in the community sample. In fact, statistical markers used in CFA suggested that the revised PRIAS measures appeared to measure racial identity more accurately for the community sample compared to the young adult college sample.2 In sum, the revised PRIAS operated similarly for the college students and the community sample, and the findings from these two studies provide strong support for the construct validity of the PRIAS-­15 and PRIAS-­12. Asian Americans’ Racial Identity Development

Racial identity development is a salient component in understanding Asian Americans’ mental health. Research attention to the measurement of racial identity among Asian Americans, however, has been sparse. This multi-­study investigation extends theory and the racial identity literature by providing measurement and construct validity evidence for the utility and factor structure of the PRIAS for Asian American young adults and community samples. Our analysis provides strong empirical support for the PRIAS-­15 and PRIAS­12. The PRIAS-­15, in particular, may be valuable in future research given that this refined version has been shown to have lower measurement error; thus, the main effects of the racial identity factors on psychological well-­being are estimated more accurately. Additionally, reducing the length of the original PRIAS minimizes participant burden. Therefore, the PRIAS-­15 and PRIAS-­12



Asian American Mental Health Research 125

may have more practical utility (for example, by reducing participant burden) and, as a result, can be better integrated in epidemiological studies. Through the refinement of the PRIAS to fifteen-­and twelve-­item measures, the subscales and the overall assessment were improved with respect to Asian American populations. Additionally, the present study offered support for the predictive validity of the PRIAS-­15 and PRIAS-­12 among a university sample. Consistent with the fifty-­item PRIAS, the dissonance racial identity factor was negatively related to self-­reported psychological well-­being using the PRIAS-­15 and PRIAS-­12. Within the racial identity development model, individuals who strongly endorse dissonance are likely to wrestle with their allegiance to White cultural standards and norms; simultaneously, their awareness of the continued existence of racism, specifically racism experienced subjectively and/or by other Asian Americans, may sharpen and become more acute. Consequently, individuals who strongly endorse dissonance may experience internal conflicts that may compromise their overall sense of psychological well-­being (Iwamoto and Liu 2010). In contrast, and also consistent with previous research, the racial identity factor internalization was related positively to psychological well-­being as measured by the original PRIAS, the PRIAS-­15, and the PRIAS-­12. Individuals who endorse internalization attitudes are conceptualized to have a positive commitment to their racial group and have the capacity to appreciate and respond adaptively to members of their own and the dominant racial group (Helms and Cook 1999). The ability to retain an affirmative connection to other Asian Americans while understanding the racial perspectives of Asian Americans and White Americans may allow individuals to navigate potentially stressful racism-­related encounters effectively. This explanation implies cognitive, behavioral, and potential affective coping strategies, similar to universal-­diverse worldview orientation (Miville et al. 1999), in which interpersonal similarities and differences are appreciated at individual and group levels. Asian Americans who endorse internalization may impart a perspective similar to universal-­diverse world-­view orientation when confronted with a range of stresses, and thus maintain positive psychological well-­being. Coupled with the findings that the relationships between dissonance and internalization with psychological well-­being were replicated using both the PRIAS-­15 and PRIAS-­12, the present study reinforces the predictive validity of the revised and shortened PRIAS measures. The relationships that emerged between the conformity and immersion/ emersion and psychological well-­being using the PRIAS-­15 and PRIAS-­12 were equivalent in the present study. More specifically, endorsement of conformity was negatively related to psychological well-­being using the PRIAS-­12, though a nonsignificant relationship emerged using the original PRIAS and the PRIAS-­15.

126

D. K. Iwamoto, M. M. Kindaichi, and M. Miller

Immersion/emersion attitudes yielded significant negative relationships with psychological well-­being using the three forms of the PRIAS. It is possible that questioning adherence and subscription to dominant assumptions about people of Asian descent may be uniquely complex in social climates in which some racist stereotypes about Asian people (for example, the model minority myth) position the racial group in relative privilege compared to other racial groups (or where race-­related stress stimuli may be highly subtle; Sue et al. 2007). Future research using the PRIAS-­15 and PRIAS-­12 may offer opportunities for a more comprehensive examination of race-­related attitudinal variables and psychological well-­being. The present study provides empirical support for the use of the PRIAS­15 and PRIAS-­12 within university and community samples, thus encouraging inclusion of race-­related attitudinal variables in epidemiological research. The literature on health disparities in illness prevalence rates and help-­seeking among people of Asian descent has tended to integrate race, ethnicity, and/or proxy variables for acculturation, such as language fluency (for example, Meyer et al. 2009), whereas explorations of racial identity development have been less developed. In the spirit of acknowledging the diversity of race-­related experiences of people of Asian descent, this vein of research would inform health disparities disciplines that address the sociocultural and racial contexts through which both illness and treatment variance emerge. Drawing from a national sample of Asian Americans, Gee and colleagues (2007) reported that higher levels of discriminatory experiences were related to increased likelihoods of cardiovascular conditions, respiratory health issues, and chronic pain. Thus, it would appear essential to examine how individuals’ sensitization and adaptive responses to racial discrimination, as conceptualized within the racial identity development model, influence the likelihood of developing health complications. Although these two studies contribute to the literature by providing strong support for a more streamlined and empirically validated measure of racial identity for an Asian American college and community sample, there are several limitations worth noting. First, the participants were drawn from convenience samples; this implies that measurement studies using nationally representative samples would provide a stronger basis for generalizability. Relatedly, all of the participants in both samples were relatively highly educated, and the community sample was primarily drawn from two popular Asian American advocacy blogs. In effect, the community sample may be more conscious about racial issues and have stronger connections with the Asian American community compared to “average” Asian Americans. Similarly, it is possible that individuals who align more strongly with specific racial identity statuses (for example, dissonance or immersion/emersion) may have been overrepresented



Asian American Mental Health Research 127

in our samples. Lastly, given these studies were cross-­sectional, no causality can be inferred, and caution is warranted in interpreting the results. Implications for Health Disparities Research

Racial identity statuses encapsulate cognitive, affective, and behavioral dimensions regarding individuals’ styles of navigating race and racism with respect to their identification with White and non-­White norms. Additionally, racial identity development may serve to explain individuals’ awareness and sensitivity to various race-­related stimuli. In turn, such acuity may precede emotional and behavioral coping strategies, the onset of health outcomes, and help-­seeking attitudes. Thus, examining the complexities by which Asian Americans understand their racial identity development may inform epidemiological, public health, and treatment initiatives that address physical and mental health disparities. By providing strong empirical support and practical utility of the PRIAS, continued research concerning the multidimensional relationships among race, racism, and health outcomes for Asian Americans can be encouraged. Research regarding the incidence of health and risk behaviors that incorporates considerations of racial identity can serve to elucidate important precursors to detrimental and resilience behaviors and health outcomes, and thus deepen professionals’ prevention and intervention efforts. Several recent studies across ethnic and age cohorts among Asian Americans reported a positive relationship between perceived discrimination and chronic health conditions (Gee et al. 2007), mental health concerns (such as suicidal ideation, intermittent explosive symptoms, and depressive symptoms; see Hahm et al. 2010), and depressive symptomatology (Jang et al. 2010). Additionally, Jang et al. reported that having a sense of mastery mediated the relationship between perceived discrimination and depressive symptoms among a community sample of adult Korean Americans. It can be implied that clarifying Asian Americans’ racial identity development can inform paths through which they identify, process, and internalize discriminatory events, as well as how their cognitive and affective styles may serve to protect against negative health outcomes. Additionally, racial identity assessment may help us better understand substance use behaviors among Asian Americans (Iwamoto, Corbin, and Fromme 2010). Assessment of racial identity may also inform incidence and prevalence rates of physical and mental health concerns with respect to the intersections of various social locations and experiences of oppression and discrimination. Such avenues of research may serve to address the relative lack of epidemiological data on health outcomes and discrimination among Asian Americans. For example, using data from the National Latino and Asian American Survey

128

D. K. Iwamoto, M. M. Kindaichi, and M. Miller

(NLAAS), Cochran et al. found that gay/bisexual men were more likely than heterosexual men to report a recent suicide attempt and lesbian/bisexual women were more likely than heterosexual women to report depressive disorders within their lives and within the past year (2007). Although the researchers implied that individuals encountering multiple sources of oppression and discrimination may experience increased vulnerability for psychological and psychiatric morbidity, the dataset excluded measures of how racial identity may have factored into the respondents’ psychological distress. Including measures of racial identity may inform studies related to physical and mental health service use among Asian Americans as well. Recent studies (for example, Abe-­Kim et al. 2007; Meyer et al. 2009) have reported that Asian Americans underuse mental health services compared to the general population. At the same time, U.S.-­born Asian Americans reportedly use mental health services more frequently than immigrant Asian Americans (Abe-­Kim et al. 2007; Meyer et al. 2009). Abe-­Kim and colleagues reported that although U.S.-­born respondents tended to use mental health services more frequently than immigrant respondents, generational differences emerged such that second-­generation (that is, the children of immigrants) individuals were more similar to immigrant Asian Americans in their service use than third-­generation individuals. Further, within-­group difference based upon nativity has been reported to relate to strengthening use of mental health services, such that U.S.-­born Asian Americans who were connected to mental health services through primary care providers were more likely to seek help while such a connection was not significant to mental health service use among immigrant Asian Americans (Meyer et al. 2009). Although considerations of generation status and language proficiency are ethnic and cultural within-­group variables, racial identity development may influence help-­seeking behaviors and attitudes among people of Asian descent. For example, individuals’ recognition of discriminatory experiences (by race, by language ability or the assumption of such) may vary as a function of racial identity, which consequently may influence help-­seeking attitudes and treatment adherence. Thus, disaggregating the potential effects of Asian Americans’ racial identity attitudes may contribute to the explanation of disparities in mental health help-­seeking and perceived helpfulness of care. Acknowledgments

This study was supported by a supplemental grant from the National Institute of Drug Abuse (R01-­DA018730–­04A1S1). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Institutes of Health.



Asian American Mental Health Research 129

Some data used in this working paper has been previously published in Iwamoto and Liu (2010); however, the findings in the present paper are original. Notes 1. Factor analysis is a statistical procedure used to uncover the underlying factor structure from a larger number of variables that is often used in psychometric scale development (Garson 2012). 2. The first CFA on the PRIAS-­32 resulted in poor model fit, χ² (844) = 428, p < .001, CFI = .73, TLI = .69, and RMSEA =.07 (CI = .07–­.08). Model fit for the PRIAS-­15 was better than the 32-­item PRIAS, though it still did not fit the data well, χ² (181) = 84, p < .001, CFI = .87, TFI = .82, and RMSEA =.07 (CI = .06–­.09). The model fit using the PRIAS-­12 provided an excellent fit with the data, χ² (301) = 98, p < .001, CFI = .95, TFI = .93, and RMSEA =.05 (CI = .022–­.074). References Abe-­Kim, J., D. T. Takeuchi, S. Hong, N. Zane, S. Sue, M. S. Spencer, H. Appel, E. Nicdao, and M. Algeria. 2007. “Use of Mental Health-­Related Services among Immigrant and U.S.-­Born Asian Americans: Results from the National Latino and Asian American Study.” American Journal of Public Health 97: 91–­98. Atkinson, D. R., G. Morton, and D. W. Sue. 1989. “A Minority Identity Development Model.” In Counseling American Minorities, 3rd ed., edited by D. R. Atkinson, G. Morten, and D. Sue, 35–­47. Dubuque, IA: William C. Brown. Alvarez, A. N., and J. E. Helms. 2001. “Racial Identity and Reflected Appraisals as Influences on Asian Americans’ Racial Adjustment.” Cultural Diversity and Ethnic Minority Psychology 7: 217–­31. Alvarez, A. N., L. Huang, and C.T.H. Liang. 2006. “Asian Americans and Racism: When Bad Things Happen to ‘Model Minorities.’” Cultural Diversity and Ethnic Minority Psychology 12: 477–­92. Chen, G. A., P. LePhuoc, M. R. Guzman, S. S. Rude, and B. G. Dodd. 2006. “Exploring Asian American Racial Identity.” Cultural Diversity and Ethnic Minority Psychology 12: 461–­76. Cochran, S. D., V. M. Mays, M. Alegria, A. N. Ortega, and D. Takeuchi. 2007. “Mental Health and Substance Use Disorders among Latino and Asian American Lesbian, Gay, and Bisexual Adults.” Journal of Consulting and Clinical Psychology 75: 785–­94. Cokley, K. O. 2005. “Racial(ized) Identity, Ethnic Identity, and Afrocentric Values: Conceptual and Methodological Challenges in Understanding African American Identity.” Journal of Counseling Psychology 52: 517–­26. Garson, D. 2012. Accessed March 8, 2012. http://faculty.chass.ncsu.edu/garson/PA765/ factor.htm. Gee, G. C., M. S. Spencer, J. Chen, and D. T. Takeuchi. 2007. “A Nationwide Study of Discrimination and Chronic Health Conditions among Asian Americans.” American Journal of Public Health 97: 1275–­82. Hahm, H. C., A. Ozonoff, J. Gaumond, and S. Sue. 2010. “Perceived Discrimination and Health Outcomes: A Gender Comparison among Asian-­Americans Nationwide.” Women’s Health Issues 20: 350–­58. Helms, J. E. 1995a. “An Update on Helms’ White and People of Color Racial Identity Models.” In Handbook of Multicultural Counseling, edited by J. G. Ponterotto, J. M. Casas, L. A. Suzuki, and G. M. Alexander, 181–­98. Thousand Oaks, CA: Sage.

130

D. K. Iwamoto, M. M. Kindaichi, and M. Miller

———. 1995b. “The People of Color (POC) Racial Identity Attitude Scale.” Unpublished manuscript. University of Maryland, College Park. Helms, J. E., and D. A. Cook. 1999. Using Race and Culture in Counseling and Psychotherapy: Theory and Practice. Needham Heights, MA: Allyn and Bacon. Iwamoto, D. K., W. Corbin, and K. Fromme. 2010. “Trajectory Classes of Heavy Episodic Drinking among Asian American College Students.” Addiction 105: 1912–­20. Iwamoto, D. K., and W. M. Liu. 2010. “The Impact of Racial Identity, Ethnic Identity, Asian Values and Race-­Related Stress on Asian Americans’ Well-­Being.” Journal of Counseling Psychology 57: 79–­91. Jang, Y., D. A. Chiriboga, G. Kim, and S. Rhew. 2010. “Perceived Discrimination, Sense of Control, and Depressive Symptoms among Korean American Older Adults.” Asian American Journal of Psychology 1: 129–­35. Kline, R. 2005. Principles and Practice of Structural Equation Modeling. New York: Guilford Press. Liang, C., L. C. Li, and B.S.K. Kim. 2004. “The Asian American Racism-­Related Stress Inventory: Development, Factor Analysis, Reliability, and Validity.” Journal of Counseling Psychology 51: 103–­14. Liu, W. M., and D. K. Iwamoto. 2006. “Asian American Men’s Gender Role Conflict, Distress, Self-­Esteem, and Asian Values.” Psychology of Men and Masculinity 7: 153–­64. Meyer, O. L., N. Zane, Y. I. Cho, and D. T. Takeuchi. 2009. “Use of Specialty Mental Health Service by Asian Americans with Psychiatric Disorders.” Journal of Consulting and Clinical Psychology 77: 1000–­1005. Miville, M. L., C. J. Gelso, R. Pannu, W. Liu, P. Touradji, P. Holloway, and J. Fuertes. 1999. “Appreciating Similarities and Valuing Differences: The Miville-­Guzman Universality-­ Diversity Scale.” Journal of Counseling Psychology 46: 291–­307. Neville, H. A., Lilly, R. L., Duran, G., Lee, R. M., Browne, L. 2000. “Construction and Initial Validation of the Color-­Blind Racial Attitudes Scale (COBRAS).” Journal of Counseling Psychology 47: 59–­70. Perry, J., K. S. Vance, and J. E. Helms. 2009. “Using the People of Color Racial Identity Attitude Scale among Asian American College Students: An Exploratory Factor Analysis.” American Journal of Orthopsychiatry 79: 252–­60. Ponterotto, J. G., and J. Park-­Taylor. 2007. “Racial and Ethnic Identity Theory, Measurement, and Research in Counseling Psychology: Present Status and Future Directions.” Journal of Counseling Psychology 54: 282–­94. 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: 1069–­81. Quintana, S. M., and S. E. Maxwell. 1999. “Implications of Recent Developments in Structural Equation Modeling for Counseling Psychology.” Counseling Psychologist 27: 485–­527. Sue, D. W., J. Bucceri, A .I. Lin, K. L. Nadal, and G. C. Torino. 2007. “Racial Microaggressions and the Asian American Experience.” Cultural Diversity and Ethnic Minority Psychology 13: 72–­81. U.S. Census. 2010. http://2010.census.gov/news/releases/operations/cb12-cn22.html. U.S. Census Bureau. 2011. “Asian/Pacific American Heritage Month: May 2011.” http:// www.census.gov/newsroom/releases/archives/facts_for_features_special_editions/ cb11-ff06.html. Worthington, R., and T. A. Whittaker. 2006. “Scale Development Research: A Content Analysis and Recommendations for Best Practices.” Counseling Psychologist 34: 806–­38.

Chapter 9

Aliya Saperstein

Representing the Multidimensionality of Race in Survey Research

When designing and conducting research, taking the claim that race is socially constructed seriously is easier said than done. Some methods, such as historical/ archival research or in-­depth interviewing and ethnography, are better suited to reveal the specific moments, mechanisms, and processes through which racialized categories and boundaries shape life chances and come to have meaning in people’s lives. However, that does not—­and should not—­let those of us who employ demographic methods or quantitative survey research off the hook. The critique of standard practices in survey research and official data collection is twofold: first, most practitioners continue to use a single set of mutually exclusive, exhaustive categories that have been cobbled together piecemeal through institutional inertia and political compromise (see Nobles 2000 and Skerry 2000 for discussions); second, the drawbacks of these categories are compounded by using overly simplistic methods and providing interpretations that rely on and reinforce an essentializing view of race (for similar critiques, see Helms and Mereish this volume; Martin and Yeung 2003; Zuberi 2000). Of course, it is easy to point to the problems with most survey research about race, but it is much harder—­and thus less common—­to provide concrete, practical solutions. Widespread adoption of better methods for studying race, ethnicity, and social inequality will not be achieved until survey directors, researchers, and applied practitioners can be shown that taking a different approach has substantive consequences for theory, policy, and our empirical understanding of how the world works. This presents a catch-­ 22: demonstrating concrete consequences (rather than arguing they exist in the abstract) requires collecting racial and ethnic data in nonstandard ways, yet because

133

134

Aliya Saperstein

of the costs associated with large-­scale data collection efforts, few are willing to pursue these alternative methods without knowing the payoff in advance. To meet this challenge, I recommend a two-­pronged approach: scholars must encourage large, national surveys and data collection agencies to rethink how they measure the complex and continually shifting constructs of race and ethnicity, while simultaneously exploiting the potential of existing data to demonstrate that the multidimensionality of race exists empirically, is not limited to particular subgroups in the population, and matters for studies of inequality. The latter option is feasible because many large-­scale data sets already include multiple measures of race, in one form or another, though this fact is not widely known and thus the data largely has gone unexamined. To promote research along these lines, this chapter proposes suggestions for future directions in data collection and reviews the current possibilities for incorporating multidimensional notions of race and ethnicity into survey research on social inequality. It does so with a particular focus on health disparities. I address both the strengths and weaknesses of existing data, and conclude by raising additional methodological issues that should be considered, alongside improved variable selection and construction, in order to fully incorporate a constructivist theory of race into survey research. A Multidimensional Wish List for Data Collection

Theoretical debates over how to conceptualize and define race have been waged among intellectuals and political elites for centuries (see Banton 1998; Gossett 1997; Smedley and Smedley 2005 for reviews). Similar conversations related explicitly to methodology have emerged much more recently, as scholars across disciplines became concerned with how to translate a particular conception of race into a research design, a set of survey questions, or a list of categories (for example, Kaplan and Bennett 2003; Sollors 2002). Making the shift from abstract ideas to concrete measures is an age-­old challenge in science (Blalock 1968), but the intense scrutiny of operationalization decisions in studies of race, due in part to their political implications, continues to raise the stakes of doing research on racial inequality. While the necessity of collecting racial data has been subject to periodic challenge in the United States (see, for example, Connerly 2001), even European governments that previously banned the collection of racial and ethnic data are currently reconsidering their policies because failing to capture this information also means being unable to monitor progress on racial and ethnic discrimination (Simon 2012). Much like disagreements over the definition of race, theories also abound as to why racial distinctions are among the most enduring in American society



Representing the Multidimensionality of Race 135

despite changing populations, increasing diversity, and decades of attempts to address racialized inequality. Some stress the importance of perceived physical differences that have long been equated with inherited “essences” and behaviors. Others focus on family socialization, culture and tradition, or individual identity development. Some direct attention to the accumulated disadvantages of past discrimination, while others stress the continued existence of racist norms and stereotypical expectations that shape the perceptions of individuals as a result of their ascribed membership in a particular racial group (see, for example, Brubaker, Loveman, and Stamatov 2004; Dressler, Oths, and Gravlee 2005; Omi and Winant 1994). These varied explanations of racial inequality, and the underlying conceptions of race they reference, each point to a different mechanism or a different aspect of race that explains unequal opportunities and outcomes. Trying to adjudicate between them raises questions like the following: Is it specific physical features such as skin tone that matter most, or the overall racial categorization? Is the effect of race on life chances best captured by how individuals identify or how other people perceive them? Will the strongest effects be found in ancestry and family history or through current processes? Or do several factors combine to create the overall experience of race in the United States? None of these questions can be answered using the standard methods of racial data collection. Most surveys and official statistics in the United States rely on a single measure of self-­identification, which effectively conflates the many different ways that one’s race could shape or constrain life outcomes—­including access to health care and physical and mental well-­being (Mays et al. 2003). Thus, only by shifting away from a single-­measure approach to employing multiple measures of race in data collection and survey research can large-­ scale quantitative studies move closer to incorporating a theoretically driven, constructivist approach to race and racial inequality. The specific processes through which racial categories are made meaningful in people’s lives cannot be captured directly in a survey but, by adding measures to the standard query for self-­identification, a multidimensional conception of race can be adapted to survey settings. In subsequent analysis, measures of the different dimensions of race (appearance, ancestry, and so forth) can be used in combination to help provide the point from which to start a more fruitful search for the mechanisms that drive racial disparities in health. Ideally, surveys would employ as many measures as possible, and always more than one. For example, the differential effects of external racial classification versus expressed racial identity cannot be weighed without measures of both aspects of race in the same data set. Exploiting this contrast has already

136

Aliya Saperstein

deepened several recent studies of racialized inequality in Brazil (Telles and Lim 1998), the United States (Saperstein 2006), and Eastern Europe (Ahmed, Feliciano, and Emigh 2007). Similarly, researchers cannot distinguish the effects of skin tone or other physical features from those of racial categorization without including the various measures simultaneously (see Stepanova and Strube 2012)—­but nationally representative data with these different dimensions of race are surprisingly difficult to obtain. In the United States, skin tone data has generally been collected only in national studies of non-­Whites; the few exceptions are group-­specific studies, such as the New Immigrant Survey (http://nis.princeton.edu), or the National Longitudinal Study of Adolescent Health (http://www.cpc.unc.edu/projects/addhealth), which only sample particular cohorts of Americans.1 Recent work by Roth (2010) suggests that another important distinction should be made between self-identification, other-classification, and “reflected appraisals.” The term “reflected appraisal” comes from sociological work in the symbolic interactionist tradition on how people create a sense of self and communicate it to others; it references how people think they are perceived by others, which might not match either the person’s preferred identity or how they are actually perceived in social interactions. For example, a recent immigrant from the Dominican Republic might identify as White, think she is perceived in the United States as Hispanic, but be categorized by most native-born Americans as Black. A focus on reflected appraisals, and how they differ from other measures of race and racial experience, could be particularly important for studies of perceived discrimination— a mechanism that has gained increased attention in studies of racial health disparities (Paradies 2006). Whether the multiple measures of race in surveys should also include direct biological or genetic information is a subject of fierce debate. Recent studies claim that clusters of human populations can be identified through gene mapping and that these resemble groupings based on self-­identification. Critics counter that, even if these studies’ methods are sound (which is debatable), the results do not demonstrate that “ancestry informative markers” have explicit consequences for health or well-­being (see Graves, this volume; Marks 2006). Of course, a thorough test of the competing hypotheses that genes matter more or less than perceived race or self-­identification, or any of the other dimensions of race, requires that all be included in the same study (for example, see Gravlee, Non, and Mulligan 2009). This multidimensional wish list would not be complete without acknowledging that each of these dimensions of race can and do vary over time (as well as across countries or regions). Skin tone may change with the seasons



Representing the Multidimensionality of Race 137

and types of employment. Racial and ethnic identities are added, dropped, and edited across the life course. Even external racial categorizations can shift depending on the social context (for example, one neighborhood versus another) and other available cues, including information on the individual’s social status (Freeman et al. 2011). This suggests that studies of racial inequality are best conducted longitudinally, with repeated measures of race and ethnicity in each wave. Researchers should also pursue linking several data sources that include different measures of race to one another. The Utah Population Database is exemplary in this respect because it includes longitudinally linked data for 6.5 million people from a range of sources including the 1880 census and family genealogical histories, annually updated vital statistics, cancer registries, and inpatient hospital claims (http://www.huntsmancancer.org/research/shared-resources/ utah-population-database/overview). The National Longitudinal Mortality Study represents a similar effort on a national scale that links data from the 1980 census and several contemporary Current Population Surveys to Medicare claims data and the SEER cancer registry. The database currently includes 3 million Americans with 250,000 matched death records (http://surveillance .cancer.gov/disparities/nlms/). Of course, access to both of these data sets is highly restricted due to privacy concerns, and the cost of such access can be prohibitive for most research purposes; nevertheless, they illustrate the potential of this approach. Finally, in addition to incorporating multiple measures of race and ethnicity, researchers should give more serious thought to the specific pathways that might lead from racial differences (however they might be measured) to various health disparities, and develop survey instruments and studies that examine the intervening mechanisms directly. One especially difficult puzzle to sort out in the absence of both multiple measures of race and a longitudinal framework is how various dimensions of race relate to both perceived discrimination and other social disparities. For example, having experienced discrimination and seeking to avoid such experiences are both related to an individual’s self-­identification (Basler 2008; Golash-­Boza 2006). Similarly, one’s appearance should predict the likelihood of experiencing racial discrimination, but longitudinal data suggests that how one is perceived racially reflects one’s social position, which is partly a result of disadvantage or discrimination in the past (Penner and Saperstein 2008). The various dimensions of race described here are also closely related to one another—­as when self-­identification is shaped and constrained by external perceptions. Despite these challenges, the task remains to untangle how different aspects of race are linked to experiences of

138

Aliya Saperstein

discrimination and unequal access across the life course and how these factors combine to shape health and life chances. The Untapped Potential of Existing Health Data

Making widespread changes to the way race is operationalized in survey research is not going to be easy, and there are upsides and downsides associated with any particular tactic or starting point. The upside to pursuing research using existing data sets is that it allows researchers to explore relationships and test theories of racial inequality without the expenditure of time and effort that comes with creating, validating, and deploying new measures and methods of data collection. It also has the potential to persuade others of the importance of the general enterprise, and may help to inform innovations in data collection or target future interventions. (For example, if push comes to shove, which measures of race or ethnicity should be advocated for most strongly?) The downside to using existing surveys comes when the results do not turn out as expected, and it is not clear whether to blame the inadequacies of the theories, the data, or some combination of the two. (Of course, in that case, the solution is simple: collect more data to confirm the results and test new theories!) Interestingly, several of the most promising existing surveys that include multiple measures of race also were designed to study health. These include the National Longitudinal Study of Adolescent Health (Add Health; which has followed a sample of adolescents in grades 7–­12 in the United States during the 1994–­95 school year through four waves of in-­home interviews as of 2008) and the Behavioral Risk Factor Surveillance System (BRFSS; an annual telephone survey of American adults, established in 1984 by the Centers for Disease Control). Each has strengths and weaknesses: as with most race measures in surveys, the questions, categories, and answer options change over time, making analysis of trends a challenge. Though none of the data sources described below is ideal, they represent important starting points for exploring the multidimensionality of race and its relationship to health disparities and inequality in general. Add Health is well known across the social sciences and has already been employed by a number of scholars interested in the experiences of multiracial Americans and the consistency of racial self-­identification (see, for example, Doyle and Kao 2007; Harris and Sim 2002). The survey includes six measures of race and ethnicity: self-­identified race (with multiple responses), self-­identified single “best” race, self-­identified Hispanic origin, self-­identified ethnic background, interviewer-­classified race, and interviewer-­classified skin tone. Self-­ identifications are available in the 1994–­ 95 wave 1 in-­ school and in-­ home



Representing the Multidimensionality of Race 139

samples, as well as for wave 3 (2001–­2). Interviewer classification is available for the wave 1, wave 3 and wave 4 (2007–­8) in-­home samples. Some oddities of the survey design include removing the “other” category for self-­identification between waves 1 and 3. Also, although the respondents were allowed to give multiple race responses, the interviewer classifications are limited to a single category. Despite these drawbacks, Add Health has the most comprehensive set of race measures of any current national longitudinal survey. One published study compared the wave 3 self-­and interviewer-­classified race measures and finds that among self-­identified American Indians, being “misclassified” as another race is associated with worse-­than-­average mental health outcomes (Campbell and Troyer 2007), but many important questions regarding inequality-­producing processes remain to be explored using these data. Thanks to the efforts of the Measures of Racism Working Group at the Centers for Disease Control, multiple measures of race and questions tapping respondents’ experience with racial discrimination have been included in the public-­use BRFSS since 2004, along with the standard questions on self-­identification. As a telephone survey, the BRFSS cannot include an explicit measure of racial classification, but instead asks respondents, “How do other people in this country typically classify you?” This represents a racial “reflected appraisal” as discussed above. The primary drawback to the BRFSS is that the Measures of Racism module is optional and has only been adopted by a handful of states. Thus, results are not nationally representative (though each sample is representative within its own state), and one must pool data across years to acquire enough cases for analysis of the smaller populations that can be defined using multiple measures of race. At least one published study has contrasted the BRFSS reflected appraisal measure with the respondent’s self-­identification to examine differences in self-­ reported health. Using data from the 2004 survey, Jones et al. (2008) find that, among self-­ identified American Indians, Hispanics, and people who report multiple races, those who are perceived as White (or think they are) have better self-­reported health on average than people who are not perceived as White. Similarly, Stepanikova (2010) and Campbell and Troyer (2011) use the reflected appraisal comparison to examine differences in mental health outcomes, though the two studies report different results regarding whether perceived race “status gains” (for example, being seen as White among self-­identified non-­Whites) and “status losses” (being seen as Black among self-­identified non-­Blacks) are associated with worse mental health—­in part, perhaps, because they pool different years (and thus different states) in their analyses. Still other surveys remain underused because the fact that they include multiple measures of race is buried in their voluminous variable lists and

140

Aliya Saperstein

mentioned only briefly, if at all, in their user’s guides. These include the National Survey of Family Growth (NSFG), the General Social Survey (GSS), and the 1979 cohort of the National Longitudinal Survey of Youth (NLSY). Each of these surveys also includes measures of health, though, thus far, only the NSFG data has been used to examine racial health disparities in a multidimensional framework (Saperstein 2009). The study shows that variation in access to preventative health care among women was more closely related to perceived race rather than self-­identification; however, net of all other measured factors, including self-­identification, being seen as White was not necessarily associated with reporting more frequent screenings (Saperstein 2009). The NSFG has measures of self-­identified race and ethnic origin/ancestry, as well as interviewer-­classified race in each of its first four cycles (1973, 1976, 1982, and 1988) though, at that time, the in-­home interviews were limited to women of childbearing age (http://www.cdc.gov/nchs/nsfg.htm). The GSS is a nationally representative household-­based sample of American adults that included the interviewer’s classification along with self-­ identified race, Hispanic origin, and ethnic background in its 1996 and 2000 surveys (http://www3.norc.org/gss+website/). The sample size of the GSS hinders multivariate analysis of the multiple measures of race in combination, but results using self-identified and interviewer-classified race have been fruitfully contrasted in an analysis of income inequality (Saperstein 2006). The NLSY is an ongoing nationally representative study of 12,686 young men and women who were fourteen to twenty-two years old when they were first surveyed in 1979 (http://www.bls.gov/nls/nlsy79.htm). It includes self-identified “origin or descent” in 1979 and self-identified race and Hispanic origin in 2002, as well as interviewer classifications in each survey year between 1979 and 1998. Many of these older surveys present a challenge for researchers because the measures of race were not explicitly designed to measure multiracial status or the multidimensionality of race. For example, in the NSFG, the additional measure of race (classification by the interviewer), was included primarily for internal use, to ensure there was no missing data on race that would complicate the calculation of post-­stratification weights.2 The GSS briefly included multiple measures of race to determine whether changing from primarily interviewer classification pre-­1996 to self-­identification post-­2000 would affect time trends. From the survey directors’ perspective, the finding that interviewer classification and self-­identification were concordant 95 percent of the time justified the switch and the “extra” measure was dropped (Smith 1997, 2001).3 Similarly, the NLSY dropped its interviewer classification after 1998 for perceived redundancy and because it was moving primarily to telephone interviews. In each



Representing the Multidimensionality of Race 141

case, the interviewer’s classification is coded in very broad “White,” “Black,” and “Other” terms, and American Indians and Asians were not oversampled, effectively limiting multivariate analysis to exploring stratification along the Black-­ White (and possibly Latino) divide. These drawbacks, along with the ones noted above, make it especially important to advocate for improved multidimensional measurement of race in future surveys. Reorienting Research Designs

Regardless of how scholars move forward, whether by capitalizing on the multiple measures of race in existing data sets or pushing for measures of multidimensionality in future surveys, there are other valuable methodological lessons to be gleaned from a constructivist perspective on race. It is not only the way we try to capture the effects of race in our survey questions that requires revisiting. Outdated and atheoretical assumptions about what race is and how it operates to generate and maintain inequality permeate the research process. This section offers additional recommendations and cautions regarding both research design and interpretation of results. Perhaps most important, in terms of its ability to shape entire research agendas, is the often cited but rarely followed maxim against essentializing race and racial populations. Although the consensus view in the social sciences asserts that race is not based solely in biological or genetic differences, and also acknowledges that the category boundaries shift over time and across countries, a nonessentialist conception of race has more implications than is commonly realized. First, the social construction of race is not only a macro-­ level process that involves the state or social movements; it also occurs for individuals across their life course (Saperstein and Penner 2010). That is, both theoretically and, for a nontrivial proportion of the population, empirically, one’s “Whiteness” or “Blackness” or “Asian-­ness” is not given. Yet recognizing that racial perceptions can be fluid does not make the effects of race any less real. Indeed, it makes racialized inequality all the more insidious. It means we are all players in a sort of game for which the rules are continually changing. It is also important to note that racial disparities are likely not limited to “either/or” scenarios in which one dimension or aspect of race always trumps the other. While it is tempting to focus on research questions, such as “Which matters more: self-­identification or ascription?” this approach ignores the possibility of unique interactions among various measures. Much as proponents of intersectionality stress the interplay among axes of difference, such as race, class, and gender (see Hancock 2007 for further discussion), survey researchers may find that the effects of different dimensions of race are better described

142

Aliya Saperstein

as “both/and” or that different dimensions matter differently for different outcomes. For example, using data from the GSS and the NSFG to compare racial self-­identification and interviewer classification, I find that while self-­identifying as non-­White is linked to lower average income, being perceived as White is not always related to being better off in terms of either socioeconomic status or health disparities. Indeed, whether “Whiteness” is an advantage, a disadvantage, or a nonfactor varies both by outcome and by the particular combination of perceived race and self-­identification in question (Saperstein 2006, 2009, 2012). Another standard practice that takes on new meaning in a critical, constructivist framework is the interpretation of the effects of race and the existence of racial discrimination. Most commonly, race is estimated using one or more binary (“dummy coded”) indicators for the categories in question and interpreted net of the effects of other variables in a regression model. While in biomedicine the residual effects of race are often attributed to genetic factors, in the social sciences they are often read as reflecting discrimination. This “residual method” of capturing racial difference due to discrimination has been widely critiqued on methodological grounds (Blank, Dabady, and Citro 2004). Less commonly recognized is that this approach also fits poorly with theories about the institutional nature of racial discrimination. For example, if we know that racial inequalities are reinforced in the education system, is it substantively meaningful to separate the effect of “education” from the effect of “race” when studying health disparities? While many argue that finding a statistically significant effect of race after controlling for a host of potential risk factors or other demographic characteristics is evidence of discrimination or an unacceptable disparity, the converse is not also true: one cannot conclude that a nonsignificant effect of race, net of the other characteristics, indicates the absence of discrimination or disparity. In either case, one or more of the controls could (also) be capturing the effects of discriminatory processes (cf. Stewart 2008). A final research convention that does not match our theories about race and racial inequality is the widespread practice of comparing outcomes for multiple racial populations within the same regression model. If race structures life outcomes across multiple domains, from cradle to grave, that is, from residential segregation and education to the labor market and the health clinic, then it does not make sense to constrain the effects of control variables intended to tap into these experiences to be equal across the racial categories of interest. Instead, it should be an empirical question whether the effects of education, age, gender, and so forth related to a particular outcome also vary by race. Yet incorporating interaction effects, or estimating separate models for each racial or ethnic category, which allow estimates of the other controls to



Representing the Multidimensionality of Race 143

vary, is rare—­or at least rarely reported—­in most studies, including in sociology (Martin and Yeung 2003). Improving the way race is conceptualized and measured in surveys is an important step toward advancing research on health disparities. But it is only one of many that will be required to bridge the gap between our methods and our theories about racial inequality. Perhaps the most important, overarching lesson survey researchers can learn from a constructivist perspective is that much of the persistently divisive power of race rests in its malleability and its multiplicity. That is what we need to figure out how to capture in our work. Notes 1. There are several smaller-­scale studies that include skin tone data for all respondents, including the Coronary Artery Risk Development in Young Adults study (http://www.cardia.dopm.uab.edu) and the Multi-City Study of Urban Inequality (http://www.sociology.emory.edu/MCSUI). The first wave of the nationally representative Portraits of American Life Study also includes skin tone, but the measure is only available in the restricted data set (see http://www.palsresearch.org/pals/ researchers/restricteduse.asp). 2. In the NSFG’s recommended RACE variable, self-­ identification is preferred but interviewer classification is substituted when self-­identification is missing. Interestingly, in the few cases where respondents identified as monoracial White but were perceived by the interviewer as Black, the interviewer’s classification is given precedence in the calculated variable. 3. Of course this ignores that for respondents who were either perceived or identified as “other,” the two measures matched only 50 percent of the time. Further, when the two measures were used in two otherwise similar analyses of racial differences in family income, they produced significantly different results (Saperstein 2006). References Ahmed, Patricia, Cynthia Feliciano, and Rebecca Jean Emigh. 2007. “Internal and External Ethnic Assessments in Eastern Europe.” Social Forces 86 (1): 231–­55. Banton, Michael. 1998. Racial Theories. Cambridge: Cambridge University Press. Basler, Carleen. 2008. “White Dreams and Red Votes: Mexican Americans and the Lure of Inclusion in the Republican Party.” Ethnic and Racial Studies 30 (1): 123–­66. Blalock, Hubert M. 1968. “The Measurement Problem: A Gap between the Languages of Theory and Research.” In Methodology in Social Research, edited by Hubert M. Blalock Jr. and Ann B. Blalock, 5–­27. New York: McGraw-­Hill. Blank, Rebecca M., Marilyn Dabady, and Constance F. Citro, eds. 2004. Measuring Racial Discrimination. Washington DC: National Academies Press. Brubaker, Rogers, Mara Loveman, and Peter Stamatov. 2004. “Ethnicity as Cognition.” Theory and Society 33 (1): 31–­64. Campbell, Mary E., and Lisa Troyer. 2007. “The Implications of Racial Misclassification by Observers.” American Sociological Review 72: 750–­65. ———. 2011. “Further Data on Misclassification: A Reply to Cheng and Powell.” American Sociological Review 76 (2): 356–­64.

144

Aliya Saperstein

Connerly, Ward. 2001. “Don’t Box Me In: An End to Racial Check Offs.” National Review 53 (7): 24–­26. Doyle Jamie M., and Grace Kao. 2007. “Are Racial Identities of Multiracials Stable? Changing Self-­Identification among Single and Multiple Race Individuals.” Social Psychology Quarterly 70: 405–­23. Dressler, William W., Kathryn S. Oths, and Clarence C. Gravlee. 2005. “Race and Ethnicity in Public Health Research: Models to Explain Health Disparities.” Annual Review of Anthropology 34 (1): 231–­52. Freeman, Jonathan B., Andrew M. Penner, Aliya Saperstein, Matthias Scheutz, and Nalini Ambady. 2011. “Looking the Part: Social Status Cues Shape Race Perception.” PLoS One 6 (9): e25107. Golash-­ Boza, Tanya. 2006. “Dropping the Hyphen? Becoming Latino(a)-­ American through Racialized Assimilation.” Social Forces 85 (1): 27–­56. Gossett, Thomas F. 1997 (1963). Race: The History of an Idea in America. New York: Oxford University Press. Gravlee, Clarence C., A. L. Non, and C. J. Mulligan. 2009. “Genetic Ancestry, Social Classification, and Racial Inequalities in Blood Pressure in Southeastern Puerto Rico.” PLoS ONE 4 (9): e6821. Hancock, Ange-­Marie. 2007. “When Multiplication Doesn’t Equal Quick Addition: Examining Intersectionality as a Research Paradigm.” Perspectives on Politics 5 (1): 63–­79. Harris, David R., and Jeremiah Joseph Sim. 2002. “Who Is Multiracial? Assessing the Complexity of Lived Race.” American Sociological Review 67: 614–­27. Jones, C. P., B. I. Truman, L. D. Elam-­Evans, C. A. Jones, C. Y. Jones, R. Jiles, et al. 2008. “Using ‘Socially Assigned Race’ to Probe White Advantages in Health Status.” Ethnicity and Disease 18: 496–­504. Kaplan, Judith B., and Trude Bennett. 2003. “Use of Race and Ethnicity in Biomedical Publication.” Journal of the American Medical Association 289 (20): 2709–­16. Marks, Jonathan. 2006. “Genes, Race, and Population: Avoiding a Collision of Categories.” American Journal of Public Health 96 (11): 1965–­70. Martin, John Levi, and King-­To Yeung. 2003. “The Use of the Conceptual Category of Race in American Sociology, 1937–­99.” Sociological Forum 18: 521–­43. Mays, Vickie M., Ninez A. Ponce, Donna L. Washington, and Susan D. Cochran. 2003. “Classification of Race and Ethnicity: Implications for Public Health.” Annual Review of Public Health 24: 83–­110. Nobles, Melissa. 2000. Shades of Citizenship: Race and the Census in Modern Politics. Stanford, CA: Stanford University Press. Omi, Michael, and Howard Winant. 1994. Racial Formation in the United States: From the 1960s to the 1990s. New York: Routledge. Paradies, Yin. 2006. “A Systematic Review of Empirical Research on Self-­Reported Racism and Health.” International Journal of Epidemiology 35: 888–­901. Penner, Andrew M., and Aliya Saperstein. 2008. “How Social Status Shapes Race.” Proceedings of the National Academy of Sciences 105 (50): 19628–­30. Roth, Wendy. 2010. “Racial Mismatch: The Divergence between Form and Function in Data for Monitoring Racial Discrimination of Hispanics.” Social Science Quarterly 91 (5): 1288–­1311. Saperstein, Aliya. 2006. “Double-­ Checking the Race Box: Examining Inconsistency between Survey Measures of Observed and Self-­Reported Race.” Social Forces 85 (1): 57–­74.



Representing the Multidimensionality of Race 145

———. 2009. “Different Measures, Different Mechanisms: A New Perspective on Racial Disparities in Health Care.” Research in the Sociology of Health Care 27: 21–­45. ———. 2012. “Capturing Complexity in the United States: Which Aspects of Race Matter and When?” Ethnic and Racial Studies 35 (8): 1484–­1502. Saperstein, Aliya, and Andrew Penner. 2010. “The Race of a Criminal Record: How Incarceration Colors Racial Perception.” Social Problems 57 (1): 92–­113. Simon, Patrick. 2012. “Collecting Ethnic Statistics in Europe: A Review.” Ethnic and Racial Studies 35 (8): 1366–­91. Skerry, Peter. 2000. Counting on the Census? Race, Group Identity, and the Evasion of Politics. Washington DC: Brookings. Smedley, Audrey, and Brian D. Smedley. 2005. “Race as Biology Is Fiction, Racism as a Social Problem Is Real—­Anthropological and Historical Perspectives on the Social Construction of Race.” American Psychologist 60 (1): 16–­26. Smith, Tom W. 1997. “Measuring Race by Observation and Self-­Identification.” GSS Methodological Report No. 89. Chicago: National Opinion Research Center. ———. 2001. “Aspects of Measuring Race: Race by Observation vs. Self-­Reporting and Multiple Mentions of Race and Ethnicity.” GSS Methodological Report No. 93. Chicago: National Opinion Research Center. Sollors, Werner. 2002. “What Race Are You?” In The New Race Question: How the Census Counts Multiracial Individuals, edited by Joel Perlmann and Mary C. Waters, 263–­68. New York: Russell Sage. Stepanikova, Irena. 2010. “Applying a Status Perspective to Racial/Ethnic Misclassification: Implications for Health.” In Advances in Group Processes, vol. 27, edited by S. R. Thye and E. Lawler, 159–­83. Bingley, UK: Emerald Group Publishing. Stepanova, Elena V., and Michael J. Strube. 2012. “The Role of Skin Color and Facial Physiognomy in Racial Categorization: Moderation by Implicit Racial Attitudes.” Journal of Experimental Social Psychology 48: 867–­78. Stewart, Quincy Thomas. 2008. “Chasing the Race Effect: An Analysis of Traditional Quantitative Research on Race in Sociology.” In Racism in Post-­Race America: New Theories, New Directions, edited by C. Gallagher. Chapel Hill, NC: Social Forces. Telles, Edward E., and Nelson Lim. 1998. “Does It Matter Who Answers the Race Question? Racial Classification and Income Inequality in Brazil.” Demography 35: 465–­74. Zuberi, Tukufu. 2000. “Deracializing Social Statistics: Problems in the Quantification of Race.” Annals of the American Academy of Political and Social Science 568: 172–­85.

Chapter 10

Janet E. Helms and Ethan H. Mereish

How Racial-­Group Comparisons Create Misinformation in Depression Research Using Racial Identity Theory to Conceptualize Health Disparities

A multitude of studies in various physical and mental health domains attests to disparities in health outcomes of African Americans, Latina/Latino Americans, Asian/Pacific Island Americans, Native or Indigenous Americans (ALANAs), and immigrant groups of Color relative to their White American counterparts (U.S. Department of Health and Human Services 2001).1 According to the World Health Organization (WHO 2001), depression is the leading cause of disability in the general population; it damages the quality of life of sufferers, and failure to treat depression places an economic burden on the depressed person as well the communities in which the person lives. Moreover, depression co-­occurs with many physical and mental health disorders, which results in slower recovery and increased rates of mortality relative to people who are not depressed (Leentjens 2010). As compared to White Americans, ALANAs and immigrant groups of Color have similar rates of mood disorders, but they are less likely to receive appropriate treatment for them (Miranda, Williams, and Escobar 2002). Also, ALANAs have higher rates of disorders typically associated with depression. These include physical ailments, such as chronic pain and cardiovascular and respiratory health problems among Asian Americans (Gee et al. 2007) and high levels of cigarette smoking (Landrine and Klonoff 2000), alcohol consumption (Taylor and Jackson 1990), high blood pressure, and HIV/AIDS among African Americans (Crane et al. 2010; Krieger and Sidney 1996). Yet no satisfactory explanations exist to account for the between-­group disparities. Efforts to explain such disparities have generally involved comparing one or more ALANA samples to White Americans or predominantly White

146



Racial-Group Comparisons and Depression 147

aggregated samples on some health-­or mental-­health-­related symptom or outcome measure and then inferring conceptual or systemic reasons for observed between-­group differences. Depending on the researcher’s predilections, discovered between-­group differences may be attributed to systemic factors (for example, social class differences, institutional racism), racial socialization life experiences (for example, person-­level experiences of racism), or deficits in the person or the presumed racial/cultural environment in which the research participant is assumed to have been socialized. That is, in the relevant physical or mental health research, racial or ethnic cultural groupings are used as if they are conceptual constructs or variables rather than the nominal, atheoretical categories that they actually are (Helms, Jernigan, and Mascher 2005). Moreover, different explanatory attributions are made for the same ostensible racial groups without any measurement or manipulations that support such explanations. In psychology, variables or manipulations used to infer reasons for particular results on outcomes or criteria are called “independent” or “predictor” variables. Use of independent/predictor variables presumes an underlying hypothetical construct, which the variable represents. In depression research, the construct of depression is operationally defined in terms of observable symptoms that are often measured by self-­report scales, such as the Center for Epidemiologic Studies Depression scale (CES-­D; Radloff 1977). Responses to such measures can be observed and tested by others who then decide whether the construct has been adequately represented by scores on the focal depression measure. However, in such research, racial groups consistently have been used as if they were independent variables, although no theoretical constructs underlie them. Conceptual approaches and measures for studying race-­related variables exist, but they have not been integrated into the health disparities literature (Helms 1995; Yip, Seaton, and Sellers 2006). In particular, Helms (1984; 1995) proposed models of racial identity development by which she argued that, rather than merely dichotomizing people into ostensibly mutually exclusive racial categories, understanding and measuring the multiple ways in which people respond to such categorization would provide more sensitive race-­related independent/predictor variables in health and mental health research and treatment. The purposes of this chapter are to (1) provide a rationale for why racial groups ought to be replaced as independent/predictor variables in mental health research, (2) discuss some research limitations arising from treating race as categories, and (3) suggest how racial identity theory might be used to better investigate health disparities with a specific focus on depression.

148

Janet E. Helms and Ethan H. Mereish

Replacing Race with Racial Conceptual Constructs

In health research, racial groups often are used as if they are diagnostic categories in a manner similar to, say, DSM diagnoses (Neighbors et al. 1989). That is, persons assigned to categories are presumed to have certain psychological or behavioral characteristics. For example, Neighbors et al. noted the conflict between two competing camps of clinicians and researchers wherein one group contends that Blacks do not suffer from depression, whereas the other contends that clinicians misdiagnose Blacks relative to Whites, perhaps because of preexisting biases or cultural insensitivity. Perhaps such diagnostic inconsistencies occur because health disparities are virtually always discussed as if all members of each racial/ethnic group express symptoms similarly rather than acknowledging that it is individuals who express symptoms regardless of whether they have been assigned to or acknowledge membership in designated racial groups. Consequently, improvement of the health status of ALANAs requires that researchers and practitioners focus on conceptualizating and measuring race-­related constructs at the level of individuals rather than presuming health conditions from between-­group differences. Because health researchers consistently investigate racial groups as if they are real entities, the following three problems abound in the relevant literature: (i) Use of racial categories (for example, African American, White American) makes comparisons of individuals within racial groups untenable because each person within a racial group has the same “score” on the independent variable (that is, racial group). Consequently, understanding the different ways in which racial experiences and health symptoms or outcomes are manifested within groups is impossible. (ii) Because racial categories have no conceptual meaning (Helms, Jernigan, and Mascher 2005), it is easy to imbue them with both systemic and person attributes that they do not have, thereby conferring more meaning to racial categories than they merit. (iii) The absence of a priori rather than post hoc theoretical or explanatory constructs accompanying racial categories means that there is no empirical support for explanations of observed differences and, therefore, no way of using research findings to affect symptoms at the individual level. Moreover, Helms and Cook (1999) advised that it is fallacious to attribute genetic or physiological characteristics to researcher-­imposed racial categories.



Racial-Group Comparisons and Depression 149

Citing Allen and Adams (1992), they noted that racial groups lack biological or physiological meaning because (1) no biologic or genetic criteria are used to differentiate one group from another, (2) there is no evidence of conformance to such criteria within groups, and/or (3) no a priori procedures for accounting for within-­group overlap exist. Instead they contend that racial categories are social constructions with no scientific veridicality, an argument that has been around for some time (Brown 2001; Yee 1983; Yee et al. 1993; Zuckerman 1990). Furthermore, geneticists and population researchers acknowledge a lack of correspondence between social constructions of racial categories and actual gene variations among humans (Graves, this volume). Nevertheless, because racism and discrimination potentially have insidious effects on health outcomes (Thompson and Neville 1999), it is important to investigate such effects directly rather than inferring them from racial categories (Bamshad et al. 2004; Karlsen and Nazroo 2002). Socially constructed racial groups are born of and reflect racial stereotypes or societal customs rather than scientific constructs because they do not conform to the standards by which such constructs are operationally defined in the social and behavioral sciences. In psychology, for example, independent/ predictor variables are supposed to be defined by measurable or observable behaviors as broadly defined (such as symptoms, self-­reports, and so forth), whose presence gives rise to consensually agreed-­on labels that summarize or give meaning to those behaviors rather than the converse. Helms, Jernigan, and Mascher (2005) illustrate why racial categories are not scientific constructs with respect to psychological research. Here, a more pragmatic example might be useful: Suppose you invite a new neighbor, “Chris” (whom you have never met in person), to a get-­acquainted party and she or he appears at your door dressed in a bag that completely covers his or her body from head to toe except for small eye slits. Also, suppose that you need to know what race Chris is for some unknown reason (maybe you want to prove that you are an equal opportunity party giver). Of course, you would not ask such a question because it would be impolite. So, what attitudes, behaviors, or emotions (that is, psychological characteristics) would you use to determine to what racial category Chris should be assigned? The correct answer to the conundrum is that there are no such criteria. Any characteristic that one would suppose only occurs in one socially ascribed racial group can be found in others (Zuckerman 1990). The example illustrates that racial groups are at best nominal categories with no defining characteristics. Nevertheless, they are used as proxies for many concepts that people believe are racial in nature. Table 10.1 summarizes

150

Janet E. Helms and Ethan H. Mereish

Table 10.1   Some Examples of Risk Factors Used as Proxies for Race in Depression Research Risk Factors

Effect

Description

Indicators

Institutional barriers

Distal

Social or societal practices, policies, or laws intended to maintain a social hierarchy; effects may not be obvious to the person

Socioeconomic status (usually poverty); census tract

Systemic-­ environmental

Distal

Nature of the environmental contexts available to the person that are presumed to be racial

Neighborhood or school segregation by race or class

Cultural manifestations

Intermediate

Socialization experiences presumed to typify a group and presumed to shape health beliefs and practices

Manifestation of somatic rather than affective symptoms

Demographic

Proximal

Sociologic attributes of a person or group that are treated as if they are racial constructs

Age, gender, race

Biologic-­medical

Proximal

Medical conditions or genetic characteristics that allegedly affect some racial groups rather than others

Health status, phenotypes

some of the concepts that racial categories have been used to represent or symbolize in their role as independent or predictor variables. The first column lists types of risk factors that either have been studied or hypothesized to be possible causes of racial groups’ differential manifestations of depression specifically. The types of risk factors are not mutually exclusive as some researchers use more than one, perhaps in an attempt to cover the full domain of “race” (Sriwattanakomen et al. 2010). A risk is defined as “a proxy” if the researcher/ practitioner does not actually measure it but instead assumes that it pertains to the experiences of members of a certain socioracial group. In the table, proxy risk factors or independent variables are described as having potential effects on the group members that range from distal to proximal (Helms and Cook 1999; Sue and Zane 1987). Distal indicators refers to factors that, when measured, perhaps provide summary information about the attributes of research participants’ ascribed group(s), but may have virtually no quantifiable effects on individuals’ immediate experiences. As shown in the table, distal risk factors (for example, institutional barriers) may include



Racial-Group Comparisons and Depression 151

societal practices, policies, and conditions. When the focus of distal risk factors is allegedly racial, Helms and Cook use the label “socioracial” to suggest that distal attributes are imposed on the individual by society. Proximal refers to direct, immediate, person-­level life experiences. When the focus of proximal experiences is racial-­group ascriptions and individuals have internalized reactions (for example, attitudes, emotions, behaviors) to the experiences (for example, racial stereotyping), Helms and Cook label the reactions “psychoracial” to suggest that the racial experiences affect the person’s internal core. Also, Helms and Cook recommend using “sociocultural” to refer to distal or intermediate cultural socialization processes (for example, family child-­rearing practices) that are inherent to an ethnic cultural group and “psychocultural” to pertain to internalized cultural learning (for example, languages, values, health beliefs). Thus, they suggest that construct confusion is reduced by treating constructs derived from cultural factors and racial groups as separate entities given that cultural factors are measurable, whereas racial groups are not. Of the five types of risk factors (institutional barriers, systemic environmental factors, cultural manifestations, demographics, and biologic/medical origins), only biologic/medical and demographic are potentially obtained from person-­level or individuals’ behaviors or symptoms. Yet neither has anything to do with why a person is assigned to one racial group rather than another; nor can any of the five be changed by the symptomatic person, although researchers frequently and practitioners sometimes routinely change them to meet their own needs. Therefore, the types of risk factors outlined in table 10.1 do not merit the uncontested attention that they have received in the health disparities literature because they have no potential for explaining how people’s health or mental health is affected by their racial group status. Research Implications

Common approaches to studying or inferring racial risk factors from racial categories, as summarized in table 10.1, often have nothing to do with either racial socialization (for example, racism) or related life experiences or ethnic culture. Consider that some researchers control for gender and age when studying the applicability of the CES-­D across socioracial groups (Mossakowski 2008); yet it is difficult to devise an argument for why gender or age reflect any mutually exclusive racial characteristic, given that each person within every socioracial group has some of each. That is, age and gender (which are demographic factors) are not the result of being treated in society as if one belongs to a particular racial group, and, therefore, they cannot account for racial-­group differences conceptually. Similarly, the variables in table 10.1 cannot actually explain

152

Janet E. Helms and Ethan H. Mereish

racial-­group health and mental health effects, but they could be quite useful for explaining something more germane to them, such as societal barriers to equality of treatment. In their critique of the use of racial groups as constructs in psychological research, Helms et al. (2005) noted the following: (1) who is contained within a particular ascribed racial category changes according to the whims of the researcher such that supposedly mutually exclusive minority-­status groups are often combined so that the researcher has sufficient numbers of participants to conduct desired statistical analyses; (2) no consensually agreed on inclusion or exclusion criteria or procedures are used to assign persons to one group rather than another; and (3) virtually the only methodology for assigning people to groups is self or researcher report, with the researchers’ perceptions being most decisive. Thus, when racial groups are used as if they are diagnoses or explanatory constructs, none of the conditions that one would logically expect of meaningful diagnoses are met. Conceptions of Race in Depression Research

Research participants’ self-­reports of depression symptoms are the primary methodology for assessing depression between racial groups. Three primary strategies have been used to study the properties of depression measures and, by implication, levels of depression with respect to different racial groups. They are (1) between racial-­group comparisons of item responses or scale scores, (2) within-­group analyses of factor structures of item responses, and (3) studies of measurement invariance (that is, equivalence of item responses) across groups. Racial-­Group Comparisons. In the typical racial-­ group comparison study of the CES-­D scale, the depression scale scores of two or more ALANA or immigrant groups are compared to those of Whites explicitly by means of univariate statistics (for example, t-­tests, analysis of variance) or implicitly by means of multiple regression analyses in which table 10.1 demographic risk factors often are controlled. Racial categories in the comparison studies are nominal variables, which are dummy coded (for example, Black = 1, White = 0) for analytic purposes. Thus, each individual within a supposedly mutually exclusive category has the same score, and only group, as opposed to individual, comparisons and interpretations are possible. Factor Analyses. Sophisticated statistical techniques (for example, exploratory factor analysis) are often used to analyze the depression



Racial-Group Comparisons and Depression 153

item responses of a single racial group for the purpose of determining whether specific “cultural” dimensions or factors underlie the group’s item responses (Love and Love 2006). Identified dimensions are then qualitatively compared to results from similar analyses of other groups to provide a rationale for racial group differences in manifestations of depression in this case. Measurement Invariance. In this approach, two or more racial groups’ item responses to depression measures are compared statistically rather than conceptually. In one approach, called “differential item functioning,” racial groups and their interaction with level of depression are used to predict the groups’ differential response to each depression item. If the racial group interaction terms are not significant, then the measure is assumed to yield comparable scores across groups. In a second approach, called “confirmatory factor analysis,” dimensions of depression are specified a priori and the researcher attempts to discover whether the dimensions are the same for each group. With either approach, to the extent that the racial groups’ depression item responses are ostensibly the same, then researchers exclude race as meaningful. In each of the three research approaches, neither race-­related behavioral nor psychological inclusion criteria define the groups; thus, any label might be substituted for the racial categories (for example, apples, oranges) without changing the results of the analyses. In other words, quantification or assignment of a number to a categorical variable does not change its status from label to measurement. Measurement requires that the behavior of individuals within groups or conditions with respect to the independent/predictor variable can vary from person to person. Related to this point, researchers in epidemiological, biomedical, and other health fields perpetuate the illusion that race has been measured when they control for it or use it as a covariate in statistical analyses involving authentic conceptual independent/predictor variables and outcomes such as depression. In such instances, the researcher is essentially describing the groups’ averaged response rather than literally removing racial effects, which further misrepresents the relationships between race and health disparities as well as the estimates of health disparity rates (Kaufman, this volume). Several theoretical perspectives contend that it is not race per se but rather differential experiences of racism or racial discrimination that contribute to ALANAs’ poorer health status relative to their White counterparts or,

154

Janet E. Helms and Ethan H. Mereish

conversely, that the relatively better health status of Whites results from their advantaged status with respect to racism and ethnocultural violence (Clark et al. 1999; Helms, Nicolas, and Green 2010; Thompson and Neville 1999). Jones (1972) delineated three types of racism: (1) institutional—­laws, social policies, and customs whose function is to maintain the sociopolitical and economic status of ALANAs, who as groups are situated at the lower end of the sociopolitical/economic continuum relative to Whites as a group; (2) cultural—­ denigration of a socioracial group’s presumed cultural processes (for example, language usage, family patterns) or products (for example, foods) for the purpose of belittling the focal group members; and (3) individual or person-­level—­ racial slurs, stereotyping focused on the person because he or she is perceived as being a member of the devalued group. It might be noted that these types of racism vary along the distal-­proximal continuum summarized in table 10.1, with institutional being the most far removed from the person’s known, everyday experience, though it is perhaps the most potent influence systemically. Racial Identity Theory

Racial identity theories have been proposed as conceptual models to assist researchers in deconstructing racial categories. Health researchers typically have used self-­reported racial labels or government designations as their a priori definitions of racial groups or classifications, which they confusingly refer to as “racial identity” or “ethnic identity.” For example, in their examination of genetic properties across racial groups, Bamshad et al. use the Office of Management and Budget’s (OMB) classification system even though they acknowledged that it was “particularly crude and contentious” (2004, 604). They cite the OMB classification as involving “five categories for race based on physical features and/or country of origin: African-­American, ‘White,’ American-­Indian or Alaska Native, Asian, and Native Hawaiian or Pacific Islander” (2004, 604, emphasis added). Yet they also point out that the geographical boundaries that have been used historically to identify genetic markers for classifying populations into different racial groups for African Americans are situated on the African continent (Bamshad et al. 2004, 603). The relevant ethnic groups are Biaka pygmy, Mbuti pygmy, Mandenka, Yoruba, San, and Bantu (Kenya). Of course, the irony is that Africa does not produce Americans. Furthermore, most researchers do not observe the “physical features” of their research participants because they do not interact with research participants directly. Thus, the OMB classification system, although widely used, is of dubious quality (for example, see Lee, this volume). Addressing the issue of the lack of correspondence between the racial label and the people who



Racial-Group Comparisons and Depression 155

are supposed to be described by the label, and with respect to African Americans specifically, Bamshad et al. note that “comparisons in which two individuals from different racial groups were more similar [genetically] would have been higher if admixed populations such as African-­Americans had also been sampled” (2004, 604; see also Graves, this volume). Thus, another paradoxical aspect of the OMB classification system is that even though it uses popular U.S. American racial designations as labels for its racial categories, it excludes all of the ALANA groups who, for the most part, either are native to the United States (such as indigenous African Americans) or have adopted it. According to Helms and associates (Helms 1984, 1995, 2007; Helms and Cook 1999; Parham and Helms 1981), racial identity is not merely a matter of classifying people on the basis of some preordained societal classification system but rather refers to psychological (that is, psychoracial) attributes that people develop as a result of how they are treated because they are so classified. She contends that, as a result of experiencing various degrees of negative racial socialization (for example, racism), ALANAs potentially develop various schemas to help them understand and ideally overcome internalized racism, whereas White people experience different levels of undeserved positive racial socialization (for example, illegitimate privilege) that they may overcome by means of White racial identity schemas. Thus, the ultimate task for developing a healthy racial identity for People of Color or ALANAs is to recognize and overcome self-­demeaning racial socialization, whereas the ultimate task for developing a healthy racial identity for White people is to recognize and overcome self-­aggrandizing racial socialization. We hypothesize that each schema is related to different information processing styles, affective reactions, and behavioral styles. Table 10.2 summarizes the racial identity schemas for each group.2 People of Color/ALANA Schemas

The People of Color Racial Identity Attitude Scale is the measure used to assess the schemas summarized in table 10.2 (Helms 2010). Although a full review of the relevant studies is beyond the scope of this chapter, researchers have studied profiles or factor structures of item responses of Asian American (Perry, Vance, and Helms 2009) and Native American college students (Bryant and Baker 2003; Watson 2009). Of the few studies that have examined relations among racial identity schemas and mental health psychological outcomes, most have focused on global measures of psychological distress or positive adjustment rather than depression per se (Iwamoto, Kindaichi, and Miller, this volume; Iwamoto and Liu 2010). Nevertheless, depending somewhat on the population under study,

Table 10.2   Racial Identity Schemas for People of Color and White People Racial Identity Schema

People of Color/ALANA Racial Identity

Conformity

Idealization of White racial norms, valuing of White institutions, culture, and people and denigration of one’s socially ascribed racial group

Dissonance

Confusion and anxiety associated with anomie resulting from having no clear affiliation group

Immersion

Psychological withdrawal into one’s socially ascribed racial group and denigration of all things considered to be alien

Emersion

Gaining sustenance and positive racial socialization from members of the person’s ascribed racial group(s)

Internalization/ Integrative Awareness

The capacity to recognize racism, transcend its personal implications, and resist participation in it as it affects one’s own group as well as similarly disadvantaged racial and/or ethnic groups

Racial Identity Schema

White Racial Identity

Internalizing Racism Schemas

Contact

Innocence, ignorance, or neutrality about the impact of race on oneself and other people

Disintegration

Confusion resulting from conscious acknowledgment of the benefits that accrue from being White that can be lost if one violates basic principles of White racial socialization

Reintegration

Idealization of proper forms of Whiteness (for example, White people who believe as the perceiving person does) and denigration of everything perceived to be other than White

Abandoning-­Racist Identity Schemas

Pseudo-­Independence

Positive but not aggrandizing beliefs about Whiteness and oneself as a White person accompanied by a desire to help people of color become more like Whites

Immersion-­Emersion

Realistic self-­appraisal of the personal benefits of Whiteness accompanied by a self-­initiated exploration of the racial dynamics (for example, racism) in the person’s social contexts

Autonomy

Egalitarian or humanitarian attitudes and actions toward others regardless of race, which are based on an ability to recognize the various forms of racism and analogous types of oppression as they are manifested in society



Racial-Group Comparisons and Depression 157

some researchers have found that high levels of dissonance and/or immersion combined with low levels of internalization/integrative awareness are associated with poor health outcomes (Franklin-­Jackson and Carter 2007; Iwamoto and Liu 2010; Perry, Vance, and Helms 2009; Watson 2009). Carter and Reynolds (2011) found that the combination of high levels of conformity (that is, dissociation from one’s ascribed racial group) and low levels of internalization (that is, self-­ actualizing commitment to one’s ascribed racial group) were associated with high levels of depression, anger, confusion, tension, and fatigue in a sample of Black multiethnic adults. Items analogous to these symptoms appear in the twenty-­ item version of the CES-­D, which suggests that depression symptoms might be explained by race-­related (for example, racial identity) theoretical constructs. White Racial Identity Schemas

According to Helms (2008), positive White identity development potentially involves two sets of racial identity schemas, three whose focus is on different forms of internalizing racism and three whose focus is evolving a nonracist identity. The internalizing racism schemas result from White racial socialization in which White privilege is automatic and presumed to be ordained albeit invisibly to the White person. The abandoning-­racist identity schemas involve challenging White racial socialization norms and replacing them with principles that are more personally and interpersonally fair. The White Racial Identity Attitude Scale (WRIAS) was developed to assess the White racial identity schemas described in table 10.2 (Helms and Carter 1993). Most studies involving the measure have focused on psychometric properties of scales or relations between scale scores and self-­reported racism as outcomes, but have virtually ignored the possible implications of the schemas for the mental health of White people (for reviews, see Helms 1999, 2005, 2007). However, using the NEO Five-­Factor Inventory, Silvestri and Richardson (2001) did find relations between the WRIAS schemas and openness to new experiences and agreeableness (that is, altruism rather than self-­interest). Low pseudo-­independence and high autonomy were associated with openness and high reintegration was associated with low agreeableness. Their findings have no direct implications for assessing depression as manifested in White samples, but suggest that associations among mental health symptoms and WRIAS schemas merit further investigation. Conclusions and Recommendations

Racial identity as a potential conceptual replacement for racial-­group comparisons represents only a small segment of the theoretical constructs and measures available in the race-­culture assessment literature. Nevertheless, an explicit

158

Janet E. Helms and Ethan H. Mereish

focus on such concepts as independent/predictor variables provides direction for more focused health disparities research and interventions. For instance, because individuals internalize their racial socialization experiences, those experiences themselves suggest a rationale for understanding health disparities beyond merely comparing racial/ethnic groups. Comparing entire racial groups can reveal how much arbitrarily defined groups differ from each other with respect to specified outcomes, but not why they differ. Thus, they are merely descriptions of what is. An implication of our work that perhaps deserves more emphasis is that health (physical and mental) researchers interested in studying the health status of ALANAs and related immigrant groups should not rely so heavily on large-­sample statistical analysis. If it is possible that ALANA and White samples differ in their manifestations of health risk factors and outcomes because of differences in their historical racial and cultural socialization experiences, then small within-­group studies provide a better opportunity to discover what factors are important within these discrete ALANA samples. Moreover, even though as separate groups ALANAs are numerically small samples in most settings, aggregating across racial/ethnic groups to obtain large enough samples for sophisticated between-­group analyses may obscure important conceptual dynamics within groups that influence their health outcomes and therefore their health disparities (Mereish, Liu, and Helms 2012). Furthermore, if researchers must limit their investigations of health disparities to comparisons of racial groups, then it is important that they actually examine the types of risk factors summarized in table 10.1 rather than merely inferring them from racial group disparities. Helms et al. (2005) recommend examination of socioracial factors presumed to be racial constructs, such as institutional or systemic barriers, via the use of hierarchical regression analyses in which the typical order of entry of “control” and conceptual variables is reversed. They note that if proxy variables are entered in the first step of the analysis, then racial groups entered in the second step should disappear as significant or meaningful variables if the selected risk factors explain racial-­group disparities in the outcome variable. When racial groups effectively disappear in one’s analyses, then it is not “race” but institutional barriers, environmental factors, and so forth that should be the focus of health disparities research and interventions. Nevertheless, speaking to psychologists, Helms et al. note that most of those risk factors cannot be changed by psychologists, and so they recommend greater focus on theory-­driven concepts that are consistent with their scientific discipline. Yet inclusion of racial theoretical constructs in their research



Racial-Group Comparisons and Depression 159

and practice as a means of understanding health disparities is imperative for all health researchers, especially epidemiologists and public health researchers, who can best afford large-­scale studies. With their expertise in population-­ based and nationally representative research studies, health researchers provide important information and statistics regarding health disparities and valuable datasets for other researchers to study. Thus, including more rigorous, robust, and relevant racial constructs in health research designs is critical for examining and illuminating health disparities. More racially sensitive research on health disparities will permit more and better empirically informed prevention programs and policies. Notes 1. According to the Boston College web site (http://www.bc.edu/offices/ahana-adm/), “AHANA [pronounced ‘a-Ha-na’] is a term that refers to persons of African-American, Hispanic, Asian, and Native American descent, which was coined at Boston College in 1979 by two students, Alfred Feliciano and Valerie Lewis.” The students developed the term to replace the appellation “minority,” which is commonly used to refer to these groups, because the latter connotes “less than” and they wanted a label that “celebrated social and cultural differences.” Boston College has registered the term AHANA as a trademark. However, Helms and Cook (1999) broadened the term to ALANA (pronounced a-La-na) to incorporate the Americans of Color whose linguistic heritage is Portuguese and other Romance languages (for example, Brazilians). 2. The term People of Color (POC) includes ALANAs and immigrants of Color and is sometimes used interchangeably. References Allen, Bem P., and J. Q. Adams. 1992. “The Concept of ‘Race’: Let’s Go Back to the Beginning.” Journal of Social Behavior and Personality 7: 163–­68. Bamshad, Michael, Stephen Wooding, Benjamin, A. Salisbury, and J. Claiborne Stephens. 2004. “Deconstructing the Relationship between Genetics and Race.” Nature Reviews: Genetics 5: 598–­609. Brown, Tony N. 2001. “Measuring Self-­Perceived Racial and Ethnic Discrimination in Social Surveys.” Sociological Spectrum 21: 377–­92. Bryant, Alfred, and Stanley Baker. 2003. “The Feasibility of Constructing Profiles of Native Americans from the People of Color Racial Identity Attitude Scale: A Brief Report.” Measurement and Evaluation in Counseling and Development 36: 2–­8. Carter, Robert, and Amy Reynolds. 2011. “Race-­Related Stress, Racial Identity Status Attitudes, and Emotional Reactions of Black Americans.” Cultural Diversity and Ethnic Minority Psychology 17: 156–­62. Clark, Rodney, Norman Anderson, Vernessa Clark, and David Williams. 1999. “Racism as a Stressor for African Americans: A Biopsychosocial Model.” American Psychologist 54: 805–­16. Crane, P. K., L. E. Gibbons, J. H. Willig, M. J. Mugavero, S. T. Lawrence, J. E. Schumacher, M. S. Saag, M. M. Kitahata, and H. M. Crane. 2010. “Measuring Depression and

160

Janet E. Helms and Ethan H. Mereish

Depressive Symptoms in HIV-­Infected Patients as Part of Routine Clinical Care Using the Nine-­Item Patient Health Questionnaire (PHQ-­9).” AIDS Care 22: 874–­85. Franklin-­Jackson, Deidre, and Robert Carter. 2007. “The Relationship between Race-­ Related Stress, Racial Identity, and Mental Health for Black Americans.” Journal of Black Psychology 33: 5–­26. Gee, Gilbert, Michael Spencer, Juan Chen, and David Takeuchi. 2007. “A Nationwide Study of Discrimination and Chronic Health Conditions among Asian Americans.” American Journal of Public Health 97: 1275–­82. Helms, Janet. 1984. “Toward a Theoretical Explanation of the Effects of Race on Counseling: A Black and White Model.” Counseling Psychologist 12: 153–­65. ———. 1995. “An Update of Helms’s White and People of Color Racial Identity Models.” In Handbook of Multicultural Counseling, edited by Joseph Ponterotto, J. Manuel Casas, Lisa Suzuki, and Charlene Alexander, 181–­98. Thousand Oaks, CA: Sage. ———. 1999. “Another Meta-­Analysis of the White Racial Identity Scale’s Cronbach Alphas.” Measurement and Evaluation in Counseling and Development 32: 122–­37. ———. 2005. “Challenging Some Misuses of Reliability in Studying the White Racial Identity Attitude Scale.” In Handbook of Race and Cultural Psychology, vol. 1, edited by Robert Carter. New York: John Wiley. ———. 2007. “Some Better Practices for Measuring Racial and Ethnic Identity Constructs.” Journal of Counseling Psychology 54: 235–­46. ———. 2008. A Race Is a Nice Thing to Have: A Guide to Being a White Person or Understanding the White Persons in Your Life. 2nd ed. Hanover, MA: Microtraining Associates, Inc. ———. 2010. “The People of Color Racial Identity Attitude Scale (PRIAS).” Newton, MA: Huentity. Helms, Janet, and Robert Carter. 1993. “Development of the White Racial Identity Inventory.” In Black and White Racial Identity: Theory, Research, and Practice, edited by Janet Helms, 67–­80. Westport, CT: Praeger. Helms, Janet, and Donelda Cook. 1999. Using Race and Culture in Counseling and Psychotherapy. Boston: Allyn and Bacon. Helms, Janet, Maryam Jernigan, and Jackquelyn Mascher. 2005. “The Meaning of Race in Psychology and How to Change It.” American Psychologist 60: 27–­36. Helms, Janet, Guerda Nicolas, and Carlton Green. 2010. “Racism and Ethnoviolence as Trauma: Enhancing Professional Training.” Traumatology 4: 53–­62. Iwamoto, Derek, and William Liu. 2010. “The Impact of Racial Identity, Ethnic Identity, Asian Values, and Race-­Related Stress on Asian Americans and Asian International College Students’ Psychological Well-­Being.” Journal of Counseling Psychology 57: 79–­91. Jones, James. 1972. Prejudice and Racism. Reading, MA: Addison Wesley. Karlsen, Saffron, and James Nazroo. 2002. “Relation between Racial Discrimination, Social Class, and Health among Ethnic Minority Groups.” American Journal of Public Health 92: 624–­31. Krieger, Nancy, and Stephen Sidney. 1996. “Racial Discrimination and Blood Pressure: The CARDIA Study of Young Black and White Adults.” American Journal of Public Health 86: 1370–­78. Landrine, H, and E. A. Klonoff. 2000. “Racial Discrimination and Cigarette Smoking among Blacks: Findings from Two Studies.” Ethnicity and Disease 10: 195–­202. Leentjens, A. F. 2010. “Recognizing Depression in Physical Illness: Clinical Alertness, Case Finding, or Screening?” Journal of Psychosomatic Research 68: 507–­9.



Racial-Group Comparisons and Depression 161

Love, Ashley, and Robert Love. 2006. “Measurement Suitability of the Center for Epidemiological Studies—­Depression Scale among Older Urban Black Men.” International Journal of Men’s Health 5: 173–­89. Mereish, Ethan, Marcia Liu, and Janet Helms. 2012. “Effects of Discrimination on Chinese, Filipino, and Vietnamese Americans’ Physical Health: Mediating Effects of Psychological Distress.” Asian American Journal of Psychology. DOI: 10.1037/a0025876. Miranda, Jeanne, L. Williams, and Javier Escobar. 2002. “Ethnic Minorities.” Mental Health Services Research 4: 231–­327. Mossakowski, Krysia. 2008. “Dissecting the Influence of Race, Ethnicity, and Socioeconomic Status on Mental Health Adulthood.” Research on Aging 30: 649–­71. Neighbors, Harold, James Jackson, Linn Campbell, and Donald Williams. 1989. “The Influence of Racial Factors on Psychiatric Diagnosis: A Review and Suggestions for Research.” Community Mental Health Journal 4: 301–­11. Parham, Thomas, and Janet Helms. 1981. “The Influence of Black Students’ Racial Identity Attitudes on Preferences for Counselor’s Race.” Journal of Counseling Psychology 28: 250–­57. Perry, Justin, Kristen Vance, and Janet Helms. 2009. “Using the People of Color Racial Identity Attitude Scale among Asian American College Students: An Exploratory Factor Analysis.” American Journal of Orthopsychiatry 2: 252–­60. Radloff, Lenore. 1977. “The CES-­D Scale: A Self-­Report Depression Scale for Research in the General Population.” Applied Psychological Measurement 1: 385–­401. Silvestri, Timothy, and Tina Richardson. 2001. “White Racial Identity Statuses and NEO Personality Constructs: An Exploratory Analysis.” Journal of Counseling and Development 79: 68–­79. Somervell, Philip, Janette Beals, J. David Kinzie, James Boehnlein, Paul Leung, and Spero Manson. 1993. “Use of the CES-­D in an American Indian Village.” Culture, Medicine, and Psychiatry 16: 503–­17. Sriwattanakomen, Roy, Jesse McPherron, Jamie Chatman, Jennifer Morse, Lynn Martire, and Charles Reynolds. 2010. “A Comparison of the Frequencies of Risk Factors for Depression in Older Black and White Participants in a Study of Indicated Prevention.” International Psychogeriatrics 22: 1240–­47. Sue, Stanley, and Nolan Zane. 1987. “The Role of Culture and Cultural Techniques in Psychotherapy: A Critique and Reformulation.” American Psychologist 42: 37–­45. Taylor, Jerome, and Beryl Jackson. 1990. “Factors Affecting Alcohol Consumption in Black Women, Part 2.” International Journal of Addictions 25: 1415–­27. Thompson, Chalmer, and Helen Neville. 1999. “Racism, Mental Health, and Mental Health Practice.” Counseling Psychologist: Special Issue: Racism and Psychological Health 27: 155–­223. U.S. Department of Health and Human Services. 2001. Mental Health: Culture, Race, and Ethnicity—­A Supplement to Mental Health: A Report of the Surgeon General. Rockville, MD: U.S. Department of Health and Human Services, Substance Abuse, and Mental Health Services Administration, Center for Mental Health Services. Watson, Joshua. 2009. “Native American Racial Identity Development and College Adjustment at Two-­Year Institutions.” .Journal of College Counseling 12: 125–­36. Williams, David, and Selina A. Mohammed. 2009. “Discrimination and Racial Disparities in Health: Evidence and Needed Research.” Journal of Behavioral Medicine 32: 20. World Health Organization (WHO). 2001. The World Health Report 2001—­Mental Health: New Understanding, New Hope. Geneva: World Health Organization.

162

Janet E. Helms and Ethan H. Mereish

Yee, Albert. 1983. “Ethnicity and Race: Psychological Perspectives.” Educational Psychologist 18 (1): 14–­24. Yee, Albert, Halford Fairchild, Fredric Weizmann, and Gail Wyatt. 1993. “Addressing Psychology’s Problem with Race.” American Psychologist 48: 1132–­40. Yip, Tiffany, Eleanor Seaton, and Robert Sellers. 2006. “African American Identity across the Lifespan: Identity Status, Identity Content, and Depressive Symptoms.” Child Development 77: 1504–­17. Zuckerman, Marvin. 1990. “Some Dubious Premises in Research and Theory on Racial Differences: Scientific, Social, and Ethical Issues.” American Psychologist 45: 1297–­1303.

Chapter 11

Arline T. Geronimus

Jedi Public Health Leveraging Contingencies of Social Identity to Grasp and Eliminate Racial Health Inequality

Eliminating racial health inequality remains seemingly impervious to efforts and intentions. Of significance in addressing this dilemma is the concept of race and how this concept is and can be linked to health. Historically, public health has conceptualized race either as static, essentialist characteristics (genetic, behavioral, cultural, or social attributes and predispositions) or as entrenched conditions (poverty and social disadvantages related to the legacies of slavery and systematized racial segregation). However, increasingly, public health researchers are approaching race dynamically. Some are looking at how ongoing and new social processes maintain race as a lived experience with health impacts, and how dominant structural and cultural processes—­ and the social, physical, and policy environments they create—­work through a complex interplay of psychosocial, physiological, and molecular mechanisms to produce population variation in morbidity and mortality (Geronimus and Thompson 2004; Geronimus et al. 2010; Schulz et al. 2005). In addition, race is being considered beyond the Black-­White dichotomy to encompass a set of social relationships that emanate from pervasive ideologies that advantage dominant groups at the expense of others and that occur at all socioeconomic levels and in the minor ethnic, religious, or nativity divisions within racial groups (Geronimus 2000; Geronimus and Thompson 2004; James 1993; Pearson and Geronimus 2011; Viruell-­Fuentes 2007). In the United States, recent decades have witnessed growing income inequality, large waves of immigration, newly emergent or intensified xenophobia, and tensions around whether our vision for a postracial society should be race-­blind or multicultural. In this context, by acknowledging that marginalization of any

163

164

Arline T. Geronimus

identified social group may have population health repercussions, by broadening the theories of how such marginalization is enacted to impact health in a growing set of groups, and by viewing marginalization and its consequences as dynamic and relational, the field can move beyond the impasse occasioned by static and binary conceptions of how race and health are entwined (see also Garcia, this volume; Geronimus 2000; Geronimus and Thompson 2004; James 1993, 1994; Pearson 2008; Pearson and Geronimus 2011; Viruell-­Fuentes 2007). Despite these promising theoretical developments, public health research and practice continue to operate from the traditional assumption that a person’s race is fixed. This assumption is most widely recognized in the form that genetic predispositions are the starting point for understanding racial disparities. This viewpoint has been critiqued by social epidemiologists and population geneticists (for example, Cooper et al. 2003; Graves, this volume; Lewontin 1972). Now, most agree that the notion that everyone with the same phenotypic characteristics used to assign race—­most notably skin color—­would have the same health outcomes invariant to social and physical environments, access to resources, or the nature and timing of critical exposures is untenable. Now, those interested in the role genes play in population health are increasingly emphasizing the environment side of gene-­environment interactions. Others are moving into areas such as epigenetics or human stress genomics, wherein the regulation of gene expression is viewed as dynamic at the molecular level (Kuzawa and Sweet 2010),1 or are focusing on telomere length in a subset of leukocytes called peripheral blood mononuclear cells (PBMC), a measure of biological aging that appears to be sensitive to stressful life conditions (Epel et al. 2004; Geronimus et al. 2010).2 The Role of Stress Physiology

While embracing race as a social construction and assuming population differences in health along racial lines reflect social patterning, objective and subjective experiences that are socially patterned on the population level ultimately work via physiological processes and mechanisms to influence morbidity and mortality. Increasingly, public health researchers posit that prolonged psychosocial or physical challenges to metabolic homeostasis in marginalized groups can increase disease susceptibility, promote the early onset of chronic conditions (Geronimus and Thompson 2004; Geronimus et al. 2007; James 1994; Steptoe et al. 2006), and accelerate aging via a process of “weathering”—­the cumulative biological impact of chronic exposure to and coping with subjective and objective stressors (Geronimus 1992, 2001; Geronimus et al. 2006; Geronimus et al. 2010; McEwen 1998; Sapolsky et al. 2000). Everyday challenges



Jedi Public Health 165

shaped by social disadvantage may trigger chronic activation of stress processes to the health detriment of disadvantaged racial, ethnic, socioeconomic, gender, residential, or geographical ancestry groups. On a biological level, persistent high-­effort coping with acute and chronic stressors may have profound health effects. Stressors may be objective (for example, temperature extremes) or subjective (for example, financial anxiety) and, notably, need not be perceived as stressful to exert a physiological impact. Stress-­ activated biological (allostatic) systems enable people to respond to changing physical states and to cope with ambient stressors such as noise and crowding, imminent danger, hunger, extreme temperature shifts, or infection. As McEwen (1998) notes, the body’s response to a stress-­inducing challenge is twofold: turning on an allostatic response that introduces a complex cascade of stress hormones into the body and then shutting off this response when the threat has receded. When allostatic systems are not completely deactivated, the body experiences overexposure to stress hormones. Long periods of overexposure result in “allostatic load,” which can cause wear and tear on the cardiovascular, metabolic, and immune systems. This wear and tear increases susceptibility to infectious disease, early onset of chronic diseases such as hypertension, diabetes, morbid obesity, and metabolic syndrome, which, in turn, can lead to functional limitations or early death. The structural positioning of a racial or ethnic group influences its exposure and vulnerability to stressors. Stressors uniquely faced by Black Americans may accumulate and interact with one another to increase allostatic load (Geronimus and Thompson 2004). U.S. studies have found that, compared to Whites, Blacks have higher levels of cortisol and sympathetic nerve activation, as well as higher levels of oxidative stress, suggesting more frequent or intense episodes of physiological stress activation. Researchers also found higher allostatic load scores, a summary biological measure of stress-­mediated wear and tear on the body, among U.S. Black compared to White adults—­with the disparity widening from youth through middle adulthood (Geronimus et al. 2006). Among forty-­five-­to sixty-­year-­olds, Mexican immigrants have been found to be less likely than U.S.-­born Mexican Americans or non-­Hispanic Whites to have a high allostatic load if they recently arrived in the United States, but more likely than either group to have a high allostatic load if they had resided in the United States for twenty years or more (Kaestner et al. 2009).3 Researchers studying telomere length estimated that, at ages forty-­nine to fifty-­five, Black women study participants were 7.5 biological years older than their White counterparts (Geronimus et al. 2010). Indicators of perceived stress and poverty accounted for an important share of the estimated Black-­White difference in

166

Arline T. Geronimus

telomere length, but data limitations precluded evaluation of objective stressors. These lines of evidence suggest that racial or ethnic health disparities may result, in part, from repeated activation of biological stress processes in members of marginalized groups who cope over long periods with stressors inherent to their structural positioning. In addition to continued study of the physiology of stress—­its impacts on molecular dynamics, biological systems, and ultimate links to poor health—­ investigation of the nature and potential mitigation of structurally inherent stressors that activate the stress process is needed. For Blacks, researchers have focused on interpersonal racism, crime victimization, discrimination in housing and employment, material hardship, concern for physical safety, overburdened or disrupted social support networks, toxic or decaying environments, ambient stressors in residential or work environments (such as noise, pollution, crowding), and restricted access to healthy food or physical activity (see Burton and Whitfield 2003; Geronimus 2000; Schulz et al. 2005; Shapiro 2004; Williams et al. 2003). Investigators have also considered how persistently difficult conditions contribute to an increased tendency to engage in unhealthy behaviors, to feel hostility or anger, to suffer depression, or to engage in persistent high-­effort coping—­all risk factors for stress-­related diseases (Dallman et al. 2003; James 1994; Northridge et al. 1998). Yet the idea that an individual’s racial assignment or identity is a fixed trait, albeit a socially constructed one, remains restrictive. If we stipulate that the social construct of race is relational and dynamic, then it is important to look to active social and psychological processes generated by racialized ideologies and their related structures and systems to understand how racial health inequality is maintained (Geronimus 2000). If stress physiology is a key link, we need to look at the ways environments, interactions, and shared intersubjectivity of race-­conscious ideologies may be stressors. Concrete examples include overt acts of interpersonal racism; legacies of race-­based oppression, such as accumulated disparities in wealth; race conscious, inequitable policies; glass ceilings; and public disinvestment in the polluted, decaying structures and health deserts that often characterize segregated residential areas. In addition to these, how else can we imagine the relational quality of race as a dynamic force that imparts socially patterned health advantages and disadvantages? In leaping from structural conditions and physical environments to molecular mechanisms, we may have overlooked an important set of social-­psychological processes that activate physiological mechanisms precisely because the extent to which race affects biology is not constant within or across the experiences of individuals, but rather contingent upon the situational salience of the aspect



Jedi Public Health 167

of an individual’s social identity that is racial in specific settings, interactions, or relationships.4 This position assumes, then, that individuals experience their race as health-­promoting or health-­harming depending on situational factors and cues that vary in their daily rounds. For example, former NBA player John Amaechi (2011) describes this college experience: In the Big Ten, I ended up a bit of a star. And I got very used to walking around campus and having people respond to me in this really effusive way . . . I started to really enjoy it. I would walk around campus and people . . . would yell . . . , “Meach!” “Yeah . . . This is cool.” . . . We had a good game the night before, walking to class, “Meach . . . Good game!” “Mmm, hmm . . . you know it.” And then one day I was walking just off campus, and . . . I was really feeling good about my life, and everything I was achieving. . . . A car goes by, window winds down, jams the brakes on, I’ve got my hand right here [raises hand in a wave] . . . All of a sudden, out the window, this kid goes, “Nigger!” . . . it affected everything. Some random kid, in the back of a crappy car that I can’t even describe . . . someone I’ve never seen since, I don’t think, someone I had certainly not met before . . . yells that word out of the window, and my world kind of crushed. There I was as an athlete at the time . . . But he crushed me flat. I thought I was this rounded three-­dimensional person involved in all kinds of stuff, and a word from a stranger made me doubt myself.

In the course of this example, Amaechi’s skin color phenotype does not change, yet his Blackness does. Its salience to and impact on him becomes more central and harmful because of its sudden social psychological relevance. What made him Black in a health-­harmful way was what another person said and what they both understood it to mean. Identity group membership awareness—­and the experience or valence of this awareness—­is not fixed. The degree to which an individual is self-­conscious of their racial identity, and whether that self-­ consciousness is comforting or stressful, a source of self-­doubt or ethnic pride, is situational. It varies with social context, historical moment, expectations for performance, and situation-­specific cues. Steele’s Concept of Contingencies of Social Identity

With the above vignette in mind, we turn to Claude Steele’s concept of “contingencies of social identity.” Steele applies this concept most directly to the actions of “stereotype threat” on performance: the increased potential for

168

Arline T. Geronimus

underperformance in situations where an individual experiences cues to the negative stereotypes attached to his or her social identity (Aronson and Steele 2005). Those cues have demonstrable psychosocial impacts in the moment that can result in deteriorated performance. As in Amaechi’s case, such cues are fluid and an individual’s sense of identity safety can be swiftly undermined, with potential long-­term consequences for well-­being. How a person experiences his or her social identity at any given moment is additionally influenced by the stakes inherent in the situation for the individual. So, for example, the tragic irony, as Steele observes in the case of stereotype threat, is that the largest negative impacts on academic performance occur among individuals performing in an area where they are highly talented, trained, and personally invested, but in which members of their identified racial, ethnic, or gender group are thought to be incompetent. These high-­stakes instances increase the potential of negative stereotyping to engender self-­doubt, self-­consciousness, or social paralysis, despite one’s talent or preparation and in the face of pressure to overcome or disprove the stereotype. Steele observes that such stereotype threat is sufficiently powerful to “single out an identity and make it the center of a person’s functioning, powerful enough to make it more important, for the duration of the threat, at least, than any of the person’s other identities” (2010, 72). As Amaechi describes the impact of that one unexpected racist epithet: “I had felt myself to be this person that was rounded, that had everything that was going on, that was involved and connected to people in a real way . . . And some random stranger . . . made me think, is this what they really see? Is this what people see? Is it true that you can do all this stuff, that you can work this hard and try to achieve all these things, and still all people will see when they look at you, is that?” Amaechi’s example involves his response to a charged, direct interpersonal verbal attack. Yet individuals whose social identities place them in marginalized groups relative to others in the same setting can also be negatively impacted by shared understandings of marginalizing social ideologies and environmental cues that aren’t targeted specifically at them. Amaechi provides an example, referring to an e-­mail he had received from a college sophomore who was a starter on the varsity basketball team: He said, “Kobe’s my favorite player” . . . And now that’s all changed, because one day . . . on TV . . . he could see Kobe’s lip curl with contempt, as [Kobe] said what he thought . . . is the worst thing you can say about another man. He said that F-­word, right there, big as life. That word, this kid said in his email, that he heard every single day. Never



Jedi Public Health 169

directed at him because he’s a Black kid, basketball superstar . . . and therefore it’s impossible for him to be gay. But . . . the word was always around him. He made it sound . . . like it swarmed him like a mosquito, sucking the life out of him one drop at a time, infecting him with despair.

This, too, is a stark example. Cues to stereotyped identity are usually more subtle than hate speech: they can be unspoken and still have measureable effects. Thus, for example, researchers found that Asian American girls perform better on tests of mathematical achievement when they are cued to remember they are Asian, and do worse on the same math tests when they are cued to remember they are girls (Ambady et al. 2001). This impact has been observed in Asian girls in the United States as young as five years old. We may posit a similar link between contingencies of social identity and health that hinges on recognized and largely unquestioned social ideologies, hierarchies, attitudes, and expectations that are structured by identity group. Through a broad range of social classifications related to a person’s or a group’s phenotypic characteristics, national origin, or religious affiliation, each of us becomes a gendered, racialized actor in a pervasive racial paradigm, shaping our lives and health in countless and consequential ways. In considering how Steele’s documented instances of underperformance are linked to neurological and physiological processes in real time, it is logical to consider the possibility that threatening contingencies of social identity are important triggers of chronic activation of physiological stress processes. Steele draws on the model Schmader and Johns (2003) developed of the racing mind to interpret how stereotype threat results in deteriorated performance: “First, the threat of confirming the stereotype makes us vigilant to all things relevant to the threat, and to what our chances of avoiding it are. Second, it raises self-­doubt and then rumination over how warranted the doubts are. Third, these concerns lead us to constantly monitor how we are doing (something that can cause choking in athletes, for example) . . . that’s a lot of mental activity and while it’s going on there isn’t much mind left over for other things” (Steele 2010, 124). Poorer performance itself contributes to the cumulative disadvantage already faced by marginalized group members—­ manifesting in restricted or foreclosed options for advancing education or employment, with related health implications. But it also appears that the repeated experience of stereotype threat expends cognitive resources and activates physiological stress processes that can cause wear and tear on important body systems over time. Even the finite lab-­induced encounters with stereotype threat that Steele and

170

Arline T. Geronimus

others used in their experiments were enough to affect pulse rate and stability, raise blood pressure, increase rumination, selectively recruit neural networks, and reduce working memory, leading to the observed deteriorated performance in their experimental groups (Blascovich et al. 2001; Krendl et al. 2008). Continuing Inquiry

Considering race in its situational dimensions opens up new opportunities for inquiry and, ultimately, for intervention. It offers researchers a possible perch from which to better understand persistent puzzles in social epidemiology. For example, racial health disparities are apparent at all income levels and regardless of access to health care (Smedley et al. 2003). Historically, the “common sense” interpretation has been to assume that some essential feature of being Black (whether genetic, behavioral, or cultural) must explain the residual “race” effect after controlling for income or health care use. This interpretation has been well-­critiqued with reference to the extent to which race and socioeconomic variables are subject to considerable measurement error and, thus, to residual confounding (Kaufman et al. 1997), as well as on the grounds that interpreting a residual as due to factors that were not measured in the study (such as genes, behavior, or culture) is scientifically inappropriate (Kaufman et al. 2007; Graves, this volume). Examining how the contingencies of social identity may act as a mechanism through which lived experience in a racialized society is translated into physiological stress process activation allows researchers to identify and gauge psychosocial processes that differ between Blacks and Whites of similar socioeconomic status as alternative explanations to essentialist interpretations of residual race effects. For example, Amaechi was the son of British physicians, graduated from a Big Ten college, and became an NBA player, presumably with a top one percent salary. His high socioeconomic position certainly must have protected his health in many respects—­through avoidance of material hardship and access to knowledge, health care, food security, and health-­promoting physical environments, for example. Yet it did not inoculate him from feeling crushed when a stranger reduced him through racist hate speech. In that drive-­ by interaction, the racist ideology of Black inferiority—­ with all its power to undo him—­was embedded in a single word. Turning to physiological impacts of contingencies of social identity for other socially marginalized groups, studies document that poor Mexican immigrants often have better health than other poor people in the United States (Markides and Eschbach 2005), and also that this health advantage disappears



Jedi Public Health 171

in the next generation or even for the immigrants themselves the longer they reside in the United States (Collins and Shay 1994; Kaestner et al. 2009). Perhaps, as new immigrants, Mexicans in America are impervious to racialized contingencies of social identity (Pearson 2008), either because they have not internalized the U.S. racial hierarchy or because their social interactions are primarily with other Mexicans (Viruell-­Fuentes 2007). As they reside in the United States longer, or, as the children of immigrants, they are raised here, they become more integrated with Whites, more aware of racial hierarchies and ideologies, and more vulnerable to contingencies of social identity pursuant to being of Mexican descent in the United States. Some evidence across the minor ethnic divisions within racial groups may also be interpreted from this perspective. One study found the self-­reported health of Jewish Americans, who are highly educated and economically well-­off as a group, is comparable to that of other Whites until socioeconomic resources are controlled, where upon their reported health more closely approximates that of Black Americans (Pearson and Geronimus 2011). This disparity might be explained by the impacts of contingencies of social identity for the stigmatized White group—­in this case, Jewish Americans. In another intriguing example, Arabic-­named women in California suffered higher rates of poor birth outcome in the six months after September 11, 2001, than they had in the same six months the previous year, while rates for other ethnic groups did not change (Lauderdale 2006). One might hypothesize that the contingencies of having an Arabic social identity in the United States underwent a major transformation in the wake of 9/11, with concomitant stress process activation showing rapid health affects through its negative impact on pregnancy. The above examples provide suggestive, indirect evidence, and there is much more research to be done. Whether temporary activation of the physiological stress process in response to specific, sometimes laboratory-­ manipulated cues really progresses into the early onset of chronic disease remains to be tested (although animal studies also suggest the biological plausibility of this hypothesis, see Blanchard et al. 2001; Jayo et al. 1993; Sapolsky 1998; Sekl and Meany 2004). To test hypotheses stemming from a focus on contingencies of social identity, interdisciplinary, collaborative efforts are needed, including between population health researchers and social psychologists. Together they could devise ways of monitoring the extent to which members of marginalized groups experience stress process activation secondary to changing contingencies of social identity and then identify the situations or cues that trigger these processes.

172

Arline T. Geronimus

Translational Research

To determine best practices, translational research is also needed. Translational research refers to the transformation of interdisciplinary research findings into practical applications to improve health. Increasingly, translational research related to racial health disparities investigates biochemical pathways, molecular dynamics, and potential pharmaceutical treatments. However, findings from the social and behavioral sciences suggest alternate translational routes. In the current case, several lessons may be drawn from the research on the effect of stereotype threat on performance. First, the stress activation process that accompanies deteriorated performance appears to be triggered by contingency signaling cues in a specific setting. Members of groups at risk of devaluation based on their racialized social identity are attuned to such cues. For example, Purdie-­Vaughns et al. (2008) found low minority representation in workplace settings coupled with an explicit organizational creed of “color blindness” (as opposed to “valuing diversity”) apparently led African American professionals to become distrustful of the setting and perceive threatening identity contingencies within it. Such findings point to the ubiquity of encounters with threatening contingencies of social identity for members of underrepresented groups but also to potential foci for ameliorative interventions to preserve identity safety in potentially threatening institutional environments, including schools, workplaces, doctors’ offices, health care facilities, and neighborhoods. As Steele notes, If there is nothing in these settings that you have to deal with because you are a woman, or Black, or older, or have a Spanish accent then these characteristics will not become important social identities for you in that setting. They’ll be characteristics you have. You might cherish them for a variety of reasons. But in that setting they won’t much affect how you see things, whom you identify with, how you react emotionally to events in the setting, whom you relate to easily, and so on. They won’t become central to whom you are there. (2010, 73)

Unlike studies that connect structural background factors to molecular dynamics and health status, a focus on contingencies of social identity allows consideration of guiding principles for straightforward, low-­tech, but potentially fundamental primary prevention strategies. The finding that situational cues, even cues that are incidental or spontaneous representations of entrenched racialized ideology, can throw a person into self-­doubt or despair with the attendant physiological burden of becoming highly vigilant, can be



Jedi Public Health 173

turned around to suggest the ameliorative power of becoming attuned to how aspects of settings and situations affect people inequitably and, then, changing the settings so socially patterned disadvantage is not reproduced. Jedi Public Health

Those studying the impact of stereotype threat both in labs and classrooms (K–­12 and college) have concluded that large and enduring improvements in performance can be realized in youth by changing situational identity contingencies, the cues that signal them, and the narratives that students use to interpret them (for a review, see Steele 2010). In the school context, they have shown positive results in improving performance through enhancing feelings of identity safety through simple methods, including • Improving a minority group’s critical mass (number) in an integrated setting; • Changing ways of giving critical feedback to show confidence in the student’s ability to succeed, • Framing the ability to meet a challenge as learnable and expandable rather than as a fixed capacity; • Fostering intergroup conversations that substitute familiarity and firsthand knowledge for stereotype-­driven assumptions; • Having students affirm their most valued sense of self, helping inoculate them from threats; • Helping students develop a narrative about the setting that explains their frustrations. Many of these guiding principles for action can be applied to other institutional settings such as workplaces, health care facilities, or health promotion interventions. More broadly, they may suggest policy applications related to neighborhood environments and representations of social settings and interactions in the media. The more settings in which a person experiences identity safety instead of stereotype threat, and the fewer insults that are accumulated on a daily basis to trigger the physiological stress process, the greater the chance that positive recursive processes are set in motion instead. For example, a girl who is not exposed to threatening situational cues in the school setting about girls and math will not only perform better on a specific math test but may also enlarge her future options if she is subsequently more likely to continue math study than a girl exposed to stereotype threat in math. A talented and

174

Arline T. Geronimus

committed Black college student who is not challenged by stereotype threat in his organic chemistry course may be more likely to realize his potential and ambition to become a physician when it is not unilaterally dashed by a low grade in this gateway course (Steele 2010). Moreover, practices that reduce stereotype threats to underrepresented group members in integrated settings such as schools not only improve the health and life prospects of individuals, but they help diversify the academy and the labor force, naturally improving the prospects for continued identity safety. The power of speech that Amaechi details can also be used to provide new self-­affirming narratives, instill confidence in vulnerable groups, and quell physiological stress activation. As a seven-­year-­old, accompanying his mother on house calls in England in 1977—­the year the first Star Wars movie was released—­Amaechi observed, I would sit in the living room of families who are really, really stressed because somebody in their house is really, really sick. And my mother would . . . do doctor stuff, and then she would come downstairs and she would always make time to talk to these families . . . She would look at these families . . . the tension in the room was clawing at them. She would look at them and she would say, “You can cope. I’ll be back in two weeks and you will be fine . . . You can manage . . . This is what you are going to do . . . and then I will be back.” You could just feel the tension . . . drop in the room. . . . . . . I thought, I’ve seen this before . . . in Star Wars. Obi-­Wan Kenobi, he’s got the droid in the back and he is being stopped by the storm trooper and Obi-­Wan says to the storm trooper, “These aren’t the droids you are looking for.” And the storm trooper said, “These aren’t the droids we are looking for. Move along, move along.” This is what my mother is doing. I remember going back to my room, shutting the door and sitting on my bed and thinking, my mother is a Jedi.

We can all be Jedi in this way. While doing our best to design settings and interactions to maximize all participants’ feelings of identity safety and to eliminate many episodes of health-­threatening stress process activation, we can also use words once a stress-­inducing challenge has occurred, to shut off the stress response and allow the feeling of threat to recede, thereby reducing overexposure to stress hormones. Actions that reduce stereotype threat and threats to identity safety more broadly—­that expand the ways that diverse people feel valued and safe from the contingencies of their racial identities in integrated settings and that



Jedi Public Health 175

reassure all individuals of their competence and belonging in a variety of situations—­would be steps toward Jedi public health. Considerable evidence now suggests that taking such actions would improve the academic and work performance of members of underrepresented groups. Even if that is all such actions accomplished, they would contribute to reducing health disparities by mitigating structural barriers to socioeconomic resources for members of oppressed groups. But if, as one might suspect, these actions also importantly diminish the frequency or duration of acute episodes and chronic periods of physiological stress process activation and set positive recursive processes in motion instead, great and rapid gains toward eliminating racial health inequality might be achieved. While we await returns on high-­stakes, speculative, molecular, and pharmaceutical translational research investments—­knowing also that their products are likely to be distributed inequitably and may even reify racial health inequality (Kahn, this volume; Link and Phelan 1996; Rubin et al. 2010)—­it would be relatively straightforward and inexpensive to explore the possibilities of Jedi public health. Acknowledgments

The author is grateful to the Center for Advanced Study in the Behavioral Sciences at Stanford for a residential fellowship and for providing a setting where many helpful conversations took place as these ideas developed, especially with Claude Steele, Larry Bobo, Anne Petersen, Sterling Stuckey, and Amy Stuart Wells. National Institute of Aging Grant # R01 032632 provided additional financial support. I am also grateful to Sherman James for his scholarship and encouragement; to Miriam Geronimus and participants at the National Institutes of Health (NIH) workshop, “Mapping ‘Race’ and Inequality: Best Practices for Theorizing and Operationalizing ‘Race’ in Health Policy Research” for helpful comments; and to N.E. Barr and Jenny Crawford for assistance with production of the manuscript. The views expressed are my own. Notes 1. Epigenetic mechanisms influence what proteins cells make under specific conditions. To the extent that the environmental conditions that impact them are experienced at critical periods, especially in utero, they may have effects on later health, although understanding of epigenetic influences and mechanisms, as well as the degree to which they may be permanent or transgenerational, is in its infancy. Similarly, human stress genomics research suggests that stressful life experiences can influence gene expression. Both of these emerging approaches require continued elaboration and critique beyond the scope of this chapter. 2. Telomeres are the stabilizing caps of chromosomes that shorten with each mitotic division until the cell either dies or enters reproductive senescence. Research

176

Arline T. Geronimus

indicates that, because breaks in the DNA structure due to oxidative stress are not easily repaired in PBMC telomeres, oxidative stress is an important mechanism by which telomeres are shortened. As oxidative stress is an important mechanism linking aging, psychosocial stress, biological stress activation, inflammation, and disease development, PBMC telomere length may serve as a powerful marker of the toll that cumulative stress takes on the body—­a biological rather than chronological age marker (see Geronimus et al. 2010 for a review). The validity and utility of using telomere length as a biomeasure of aging is still being investigated. 3. These estimates were robust to controlling for measured health behaviors and health care use, and they occurred despite higher economic profiles of immigrants with increased length of residence in the United States. 4. This is different from the idea that one’s reported racial label may change according to situation—­although it is also true that people’s self-­reported racial identity in surveys such as the census can be fluid; some members of marginalized racial or ethnic groups “pass” part or full time; or their racial identity might differ from their racial assignment in the eyes of others. References Amaechi, J. 2011. “Stereotypes and the Power of Words.” Speech presented at Diversity Inc.’s Special Awards dinner (November 9, 2011). Washington, DC. Ambady, N., M. Shih, A. Kim, and T. L. Pittinsky. 2001. “Stereotype Susceptibility in Children. Effects of Identity Activation on Quantitative Performance.” Psychological Science 12 (5): 385–­90. Aronson, J., and C. M. Steele. 2005. “Stereotypes and the Fragility of Academic Competence, Motivation, and Self-­Concept.” In Handbook of Competence and Motivation, edited by C. Dweck and E. Eliot, 436–­56. New York: Guilford. Blanchard, R. J., C. R. McKittrick, and D. C. Blanchard. 2001. “Animal Models of Social Stress: Effects on Behavior and Brain Neurochemical Systems.” Physiology and Behavior 73 (3): 261–­71. Blascovich, J., S. J. Spencer, D. M. Quinn, and C. M. Steele. 2001. “African Americans and High Blood Pressure: The Role of Stereotype Threat.” Psychological Science 12: 225–­29. Burton, L. M., and K. E. Whitfield. 2003. “Weathering towards Poorer Health in Later Life: Co-­Morbidity in Urban Low-­Income Families.” Public Policy and Aging Report 13 (3): 13–­18. Collins, J. N., and D. K. Shay. 1994. “Prevalence of Low Birth Weight among Hispanic Infants with United States-­Born and Foreign-­Born Mothers: The Effect of Urban Poverty.” American Journal of Epidemiology 139: 184–­92. Cooper, R. S., J. S. Kaufman, and R. Ward. 2003. “Race and Genomics.” New England Journal of Medicine 348 (12): 1166–­70. Dallman, M. F., N. Pecoraro, S. F. Akana, S. E. La Fleur, F. Gomez, H. Houshyar, et al. 2003. “Chronic Stress and Obesity: A New View of ‘Comfort Food.’” Proceedings of the National Academy of Sciences of the United States of America 100 (20): 11696–­701. Epel E. S., E. H. Blackburn, J. Lin, F. S. Dhabhar, N. E. Adler, J. D. Morrow, R. M. Cawthon. 2004. “Accelerated Telomere Shortening in Response to Life Stress.” Proceedings of the National Academy of Sciences of the United States of America 101 (49): 17312–­15. Geronimus, A. T. 1992. “The Weathering Hypothesis and the Health of African American Women and Infants.” Ethnicity and Disease 2 (3): 207–­21.



Jedi Public Health 177

———. 2000. “To Mitigate, Resist, or Undo: Addressing Structural Influences on the Health of Urban Populations.” American Journal of Public Health 90: 867–­72. ———. 2001. “Understanding and Eliminating Racial Inequalities in Women’s Health in the United States: The Role of the Weathering Conceptual Framework.” Journal of the American Medical Women’s Association 56 (4): 133–­36. Geronimus, A. T., and J. P. Thompson. 2004. “To Denigrate, Ignore, or Disrupt: The Health Impact of Policy-­Induced Breakdown of Urban African American Communities of Support.” Du Bois Review 1 (2): 247–­79. Geronimus, A. T., M. Hicken, D. Keene, and J. Bound. 2006. “Weathering and Age-­ Patterns of Allostatic Load Scores among Blacks and Whites in the United States.” American Journal of Public Health 96: 826–­33. Geronimus, A. T., D. Keene, M. Hicken, and J. Bound. 2007. “Black-­White Differences in Age Trajectories of Hypertension Prevalence among Adult Women and Men, 1999–­ 2002.” Ethnicity and Disease 17 (1): 40–­48. Geronimus, A. T., M. T. Hicken, J. A. Pearson, S. J. Seashols, K. L. Brown, and T. D. Cruz. 2010. “Do U.S. Black Women Experience Stress-­Related Accelerated Biological Aging? A Novel Theory and First Population-­Based Test of Black-­White Differences in Telomere Length.” Human Nature 21: 1938. James, S.A. 1993. “Racial and Ethnic Differences in Infant Mortality and Low Birth Weight: A Psychosocial Critique.” Annals of Epidemiology 3: 130–­36. James, S. A. 1994. “John Henryism and the Health of African Americans.” Culture, Medicine, and Psychiatry 18: 163–­82. Jayo, J. M., C. A. Shively, J. R. Kaplan, S. B. Manuck, and J. M. Jayo. 1993. “Effects of Exercise and Stress on Body Fat Distribution in Male Cynomolgus Monkeys.” International Journal of Obesity Related Metabolic Disorders 17: 597–­604. Kaestner, R., J. A. Pearson, D. Keene, and A .T. Geronimus. 2009. “Stress, Allostatic Load, and Health of Mexican Immigrants.” Social Science Quarterly. Kaufman, J.S., R. S. Cooper, D. L. McGee, et al. 1997. “Socioeconomic Status and Health in Blacks and Whites: The Problem of Residual Confounding and the Resiliency of Race.” Epidemiology 8: 621–­28. Kaufman, J., A. T. Geronimus, and S. A. James. 2007. “Letter to the Editor.” American Journal of Obstetrics and Gynecology 327 (September). Krendl, A. C., J. A. Richeson, W. M. Kelley, T. F. Heatherton. 2008. “The Negative Consequences of Threat: A Functional Magnetic Resonance Imaging Investigation of the Neural Mechanisms Underlying Women’s Underperformance in Math.” Psychological Science 19 (2): 168–­75. Kuzawa, C. W., and E. Sweet. 2009. “Epigenetics and the Embodiment of Race: Developmental Origins of U.S. Racial Disparities in Cardiovascular Health.” American Journal of Human Biology 21: 2–­15. Lauderdale, D. S. 2006. “Birth Outcomes for Arabic-­Named Women in California before and after September 11.” Demography 43 (1): 185–­201. Lewontin, R. C. 1972. “The Apportionment of Human Genes.” Journal of Evolutionary Biology 6: 381. Link, B. G., and J. C. Phelan. 1996. “Understanding Sociodemographic Differences in Health. The Role of Fundamental Social Causes.” American Journal of Public Health 86: 471–­73. Markides, K. S., and K. Eschbach. 2005. “Aging, Migration, and Mortality: Current Status of Research on the Hispanic Paradox.” Journals of Gerontology, Series B: Social Sciences and Psychological Sciences 60 (special issue no. 2): S68–­S75.

178

Arline T. Geronimus

McEwen, B. S. 1998. “Protective and Damaging Effects of Stress Mediators.” New England Journal of Medicine 338 (3): 171–­79. Northridge, M. E., A. Morabia, M. L. Ganz, M. T. Bassett, D. Gemson, H. Andrews, and C. McCord. 1998. “Contribution of Smoking to Excess Mortality in Harlem.” American Journal of Epidemiology 147: 250–­58. Pearson, J. A. 2008. “Can’t Buy Me Whiteness.” Du Bois Review 5 (1): 27–­48. Pearson, J. A., and A. T. Geronimus. 2011. “Race/Ethnicity, Socioeconomic Characteristics, Co-­Ethnic Social Ties, and Health: Evidence from the National Jewish Population Survey.” American Journal of Public Health 101 (7): 1314–­21. Purdie-­Vaughns, V., C. M. Steele, P. G. Davies, R. Ditlmann, and J. R. Crosby. 2008. “Social Identity Contingencies: How Diversity Cues Signal Threat or Safety for African Americans in Mainstream Institutions.” Journal of Personality and Social Psychology 94 (4): 615–­30. Rubin, M. S., C. G. Colen, and B. G. Link. 2010. “Examination of Inequalities in HIV/ AIDS Mortality in the United States from a Fundamental Cause Perspective.” American Journal of Public Health 100: 1053–­59. Sapolsky, R. M. 1998. Why Zebras Don’t Get Ulcers—­A Guide to Stress, Stress-­Related Disorders, and Coping. 2nd ed. New York: W. H. Freeman. Sapolsky, R. M., L. M. Romero, and A. U. Munck. 2000. “How Do Glucocorticoids Influence Stress Responses? Integrating Permissive, Suppressive, Stimulatory, and Preparative Actions.” Endocrine Review 21: 55–­89. Schmader, T., and M. Johns. 2003. “Convergent Evidence that Stereotype Threat Reduces Working Memory Capacity.” Journal of Personality and Social Psychology 85: 440–­52. Schulz, A. J., S. Kannan, J. T. Dvonch, B. A. Israel, A. Allen III, et al. 2005. “Social and Physical Environments and Disparities in Risk for Cardiovascular Disease: The Healthy Environments Partnership Conceptual Model.” Environmental Health Perspectives 113 (12). Seckl, J. R., and M. J. Meaney. 2004. “Glucocorticoid Programming.” Annals of the New York Academy of Science 1032: 63–­84. Shapiro, T. M. 2004. The Hidden Cost of Being African American. Oxford: Oxford University Press. Smedley, Brian D., Adrienne Y. Stith, and Alan R. Nelson, eds. 2003. Unequal Treatment: Confronting Racial Ethnic Disparities in Health Care. Washington, DC: National Academies Press. Steele, C. M. 2010. Whistling Vivaldi: How Stereotypes Affect Us and What We Can Do. New York: Norton. Steptoe, A., A. E. Donald, K. O’Donnell, M. Marmot, and J. E. Deanfield. 2006. “Delayed Blood Pressure Recovery after Psychological Stress Is Associated with Carotid Intima-­ Media Thickness.” Arteriosclerosis, Thrombosis, and Vascular Biology 26 (11): 2547–­51. Viruell-­ Fuentes, E. 2007. “Beyond Acculturation: Immigration, Discrimination, and Health Research among Mexicans in the United States” Social Science and Medicine 65 (7): 1524–­35. Williams, D. R., H. W. Neighbors, and J. S. Jackson. 2003. “Racial/Ethnic Discrimination and Health: Findings from Community Studies.” American Journal of Public Health 93 (2): 200–­208.

Chapter 12

Nancy López

Contextualizing Lived Race-­Gender and the Racialized-­Gendered Social Determinants of Health

When I take my two daughters and other family members to the local hospitals in Albuquerque, New Mexico, I am sometimes asked to fill out forms regarding “race,” ethnicity, and language at the registration desk.1 As I fill out these forms, I make note of the large bright posters lining some of the registration cubicles, which feature smiling patients from a variety of backgrounds. Several captions attempt to reassure patients by explaining why it is important to collect race, ethnicity, and language data in the hospital setting: “We ask because we care. By asking your race, ethnicity, and language, we are able to deliver health care equally to all patients. What is your race? What is your ethnicity? What is your primary language?” Toward the end of the placard another heading affirms: “Respecting every difference, treating each equally. Get REAL: Race, Ethnicity, and Language.”2 As an Afro-­Latina and a sociologist of racial and gender stratification, I am viscerally aware of the importance of collecting data and analysis of data on “race” and ethnicity. As several of my colleagues have pointed out in this volume, one way of pursuing high-­quality research on race and inequality in a variety of domains including health, education, and beyond is to take the social construction of race seriously (Gómez, this volume). While it is tempting to equate ethnicity with racial status, the conceptual and analytical distinction between race and ethnicity is of particular importance, as studies have found qualitatively different treatment and health outcomes for Latinos who self-­ identify or are socially defined as Black as opposed to White, or “some other race” (LaVeist-­Ramos et al. 2011; Jones et al. 2008; Gravlee and Dressler, 2005). For example, I was born and raised in a New York City public housing project

179

180

Nancy López

and Spanish is my first language. Although I share the same ethnic background of my immigrant Dominican parents, my father, who is light-­skinned, and not of discernable so-called African phenotypes, occupies a very different racial status than my mother and me. In most social circumstances in the U.S. my mother and I are classified as Black (Bonilla-­Silva 1999; Rodriguez 2000; Vidal-­ Ortiz 2004).3 The distinction between ethnicity and “race” is not trivial. As argued by Griffith (2012, 110), “In the context of men’s [and women’s] health, distinguishing between race and ethnicity can help researchers disentangle health outcomes that may be due to environmental constraints and contexts that vary by race from the cultural traditions, beliefs and habits and practices that vary by ethnicity.” In an effort to explore the separate effects of ethnicity from “race” in health disparities research, LaVeist-­Ramos et al. (2011) used the National Health Interview Survey to disentangle whether Black Hispanics are more similar to their co-­ethnics or to Black non-­Hispanics. They found that co-­ethnics regardless of race shared similar health outcomes; however, for health services outcomes, Black Hispanics occupy the same stigmatized racial status as U.S.-­born Blacks. This means that Black Hispanics did not receive the same type of treatment as their White Hispanic counterparts when they access health care: “The common cultures among black and white Hispanics people may motivate similar values, beliefs, attitudes, behaviours. On the other hand, that race exerts greater influence on both health status and health services of black Hispanics may reflect the impact of societal forces. Black Hispanics visual similarity with non-­ Hispanic blacks may lead to similar social status and subject them to similar levels of discrimination” (LaVeist-­Ramos 2011, 5).4 Here LaVeist-­Ramos et al. underscore the value-­added to health disparities research by disentangling ethnicity (culture, values, behaviors, and so on), from “race” as a social status that is analytically distinct from ethnicity or cultural background. How can we go beyond merely complying with federal guidelines to collect race and ethnicity data, to improve health care and ultimately eliminate racial and ethnic health disparities? Since 2011 I have had the privilege of serving as a member of the Race and Ethnicity Advisory Committee of the New Mexico Hospitals Association. Part of our task is to create systematic data collection that would allow us to improve the delivery of services to the diverse communities in New Mexico. At just over 2 million residents, New Mexico has a relatively small population. A harbinger of the changing demographics in the U.S., New Mexico has the highest percentage of Latinos (47 percent) and one of the highest percentages of American Indians in the country (10 percent); only 3 percent and 2 percent of the population is Black and Asian respectively. Less



Contextualizing Lived Race-­G ender

181

than 10 percent of the population is foreign born (U.S. Census 2010). New Mexico was one of a handful of states to receive a federal grant from the Agency for Healthcare Review and Quality (AHRQ). A primary goal of the grant is to assist hospitals in complying with the 1997 and 2003 Office of Management and Budget (OMB) guidelines for the collection of ethnic and racial data as two separate questions. First, all hospitals should ask about Hispanic ethnicity, and second they should instruct patients to self-­identify their race by marking one or more from the following list: American Indian or Alaska Native; Asian; Black or African American; Native Hawaiian or Other Pacific Islander; and White. A second goal of the grant was to pilot best practices for the collection of racial and ethnic data in hospital discharges and emergency department databases that would address the unique contexts and needs of diverse populations within a given state. What is innovative and potentially transformative about this grant is that it actually encourages data users to develop contextualized policy changes that go beyond mere compliance but takes a step forward to advance context-­specific data collection and action plans that advance social justice and health equity. Meaningful use means contextualizing data for a given community and context. This means that the New Mexico Hospitals Association is collecting detailed tribal affiliation data on the twenty-­two sovereign nations (nineteen pueblos, two Apache tribes, and the Navajo nation). This contextualized data will allow administrators and tribal authorities to address the challenges facing the diverse Native American communities in the state, while at same time this data can still be aggregated up to the standard OMB categories for national comparisons.5 My experiences serving on the New Mexico Hospitals Association Race and Ethnicity Advisory Committee and accessing health care for my family led me to consider several empirical, conceptual, methodological, and epistemological questions: While collecting data on gender is fairly straightforward, how can we “keep it real” when collecting “race,” ethnicity, and class data? What are some promising practices for collecting “race,” ethnicity, gender, language and class? How can this data inform public policies that advance the elimination of inequalities in health, particularly for communities that have experienced historic and contemporary discrimination in the U.S. context? And finally, how can the insights of intersectionality or a deliberate conceptual focus on the connections between race, gender, ethnicity, and class as interlocking axes of stratification shed light on the social forces that are manifested as health disparities? I make several related arguments about how to collect meaningful data in health disparities research. First, I argue that in order to understand the historic and ongoing health disparities among racially stigmatized groups, we must

182

Nancy López

anchor our analysis in an examination of what I term “lived-­race gender” and the “racialized-­gendered social determinants of health.” The racialized-­gendered social determinants of health is a framework that interrogates intersecting systems of stratification at multiple levels, including the micro/individual level or what I call lived race-­gender, the meso/institutional level, for example neighborhoods, schools or other local social contexts, and the macro/structural level of society, including state and federal policies and political economic structures at the national and global levels. Second, I argue that when we are collecting data to study health disparities, we need to be attentive, self-­reflexive, and transparent about our own positionality as well as what dimension and level of the social construction of race and/or ethnicity we are collecting data on.6 The uncritical use of race and ethnicity as interchangeable concepts without any conceptual justification may impede our ability to interrogate pathways of embodied health disparities. Race and ethnicity are analytical distinct concepts that are not interchangeable. And finally, the so-­called “gold standard” of using self-­identified “race” as the only data collected in health disparities research should be abandoned in favor of multidimensional and multilevel models. This chapter is organized into several major sections. I begin by bringing the social determinants of health, racial formation theory, critical race theory, and intersectionality into a productive dialogue for developing new ways of conceptualizing what I call the “racialized-­gendered social determinants of health” (see figure 12.2). Next, I provide autobiographical snapshots of my own “lived race-­gender,” including my experience with “race-­gender profiling” in health and being “pregnant while Black.” I detail these experiences to illustrate how race-­gender are inseparable, dynamic, and context-­specific, or what Geronimus calls “contingent” social constructions. I then provide conceptual models of “race” (figure 12.1) and ethnicity (figure 12.2), as well as the racialized-­gendered social determinants of health (figure 12.3). Key to these conceptual models is the idea of multidimensional and multilevel conceptualizations that intentionally link individual micro-­level data to meso-­level and macro-­level data on race-­gender inequality to identify the mechanisms and pathways of embodiment in a given context. I end by inviting health disparities researchers to clarify what aspect of the social construction of “race” they are collecting, analyzing, and presenting in their research. I also urge disparities researchers to, whenever possible, compare the experiences of men and women in a given racial or ethnic group. These steps can help advance research that has the potential illuminate the distinct pathways of embodied health disparities



Contextualizing Lived Race-­G ender

183

and in turn illustrate how we can advance important policy changes to eradicate health disparities. Conceptualizing the Racialized-­G endered Social Determinants of Health

The literature on the social determinants of health is anchored in unpacking how the vast majority of disease and illness is due to social forces or fundamental causes in society in the form of social policies and political decisions that create conditions of poverty and inequality (Marmot 2005; Marmot et al. 1991; Phelan, Link, and Tehranifar 2010). Accordingly, Marmot argues, “The challenge is to understand how position in the social hierarchy is related to health” (2006, 6). The “social determinants of health” paradigm stands in stark contrast to the popular biomedical model that presumes that individual-­level genetic or biological differences are the fundamental cause of health disparities by race. The biomedical model has been the cornerstone of the medical training that medical doctors are exposed to; however, the social determinants of health paradigm is beginning to gain traction. Beginning in 2015, the Medical College Admissions Test (MCAT) will include questions about the about social determinants of health: “This new section recognizes recent findings—­highlighted in the AAMC [Association of American Medical Colleges] report ‘Behavioral and Social Science Foundations for Future Physicians’—­that integrating social and behavioral sciences into medical education can improve health care” (AAMC 2012). While the social determinants of health research meticulously maps social inequality in terms of social gradients in health along traditional socioeconomic status (SES) measures, including education, income, and occupational status, the impact of lived race-­gender and the racialized-­gendered social determinants of health is conspicuously absent (Krieger 2000). According to this logic, if we see racial disparities in health (as well as in education, criminal justice, housing, employment, and so on), these inequalities are really just due to variations in socioeconomic status (for example, income, educational attainment, occupational status, net worth, and so on). It also erroneously assumes that people who are racialized as White or Black or gendered as feminine or masculine have the same lived experiences because they share the same social class status and or gender status (Feagin and Sikes 1994; Oliver and Shapiro 1994). Anchored in the social constructionist conception of race, I place “race” in quotation marks to call into question “common sense” ideas about “race” as so-­ called natural divisions in the human family (AAA 1998; AAPA 1996; Human Genome Project 2012; Mays et al. 2003; Morning 2009, 2011; Omi and Winant 1994; Takeuchi and Gage 2003; Trans-­ disciplinary “Race” Working Group

184

Nancy López

2010).7 I challenge essentialist understandings of “race” that assume there are deep and innate genetic and biological differences between people of different “races.” The Human Genome Project (HGP) has found that there is no genetic or biological basis for the social distinctions that are often invoked to classify people into “races” (Graves, this volume; Human Genome Project 2012; Morning 2011). In other words, there is no concordance between physical appearance and genetic makeup (AAA 1998; AAPA 1996; Human Genome Project 2012). Some have argued that since race does not exist as a biological or genetic reality in human society, we should aim to be color-­blind and end the collection of data on race and instead replace the word “race” with “ethnicity” (AAA 1997). I take the position that although well-­intentioned, this approach is problematic. Just because something is not biologically real does not mean that it does not exist (ASA 2003; Krieger 2000). For example, one’s class position in society is not biologically or genetically determined; however, social class is an important force shaping the life chances of individuals and entire communities. For these reasons, we need to continue collecting data on race as distinct from ethnicity. To ignore the reality of racial stratification would at best maintain the status quo (ASA 2003). Critical race theory is of particular relevance in health disparity research because it departs from the premise that racism is ingrained in the institutions of U.S. society and that White privilege (unearned advantages shared by individuals and entire groups of people that are racialized as White) and White supremacist ideologies (for example, the assumption that becoming American means approximating some so-­called, mainstream White ideal) are still very important dynamics in the United States context (Bonilla-­Silva 2003; Gómez 2007; Gotanda et al. 1995; Jones et al. 2008). Accordingly, just as power dynamics are constitutive of gender and class inequalities in our social institutions and social interactions, racial formation theory acknowledges that power and dominant ideologies that become unquestioned by society are a key dimension of the social construction of racial hierarchies (Omi and Winant 1994). The power to self-­define and name one’s own reality is key to unraveling ongoing dynamics of oppressions and racial inequality and advancing social justice in health as well as in other arenas (Bonilla-­Silva 2003; Delgado and Stefancic 2011; Feagin 2006; Gotanda et al. 1995). I also conceptualize gender as a fundamental axis of inequality in society that assigns different societal expectations to males and females along a continuum of what is defined as “masculine” or “feminine” in a given sociohistorical context (Lorber 1994; West and Zimmerman 1987). Lorber explains that gender is a multidimensional and multilevel dynamic social construction:



Contextualizing Lived Race-­G ender

185

“the analysis of gender as a social structure has its origins in the development of human culture, not biology . . . as true of other institutions, gender’s history can be traced, its structure examined and its change effect researched” (1994, 6). Micro-­, meso-­, and macro-­level gender processes are co-­constructing. At the macro level, we see gender representations in the media and other cultural institutions that continually deem women to be primarily valuable for their sexuality and physical appearance. We also see that women are absent or marginalized in major political and economic power structures. At the micro level, West and Zimmerman (1987) point out how gender performances are accomplished through the act of “doing gender.” Other facets of gender at the micro level include self-­identity, socially defined gender status, lived experience, and so forth (Frankenberg 1993; Collins 2009). “The social reproduction of gender in individuals reproduces the gendered social order; as individuals act out gender norms and expectations in face-­to-­face interaction, they are constructing gendered systems of domination and power” (Lorber 1994, 6). Theories of intersectionality offer important conceptual and methodological tools for understanding and mapping the contemporary social dynamics contributing to health disparities in the United States (Collins 2009; Crenshaw 1991; Ford and Airhihenbuwa 2010; Ford and Harawa 2010; Griffith 2012; Hurtado 1996; Landry 2006; Schultz and Mullings 2005). “The goal of an intersectional approach is to simultaneously examine the social and health effects of several key aspects of identity and contexts in ways that create new understandings of these factors and that are a more accurate reflection of the lived experiences of the populations of interest” (Griffith 2012, 106). To this end, Crenshaw urges us to “map the margins” by focusing our work on groups that are often invisible when one examines “race” as separate from gender. This happens often when universal policies for health disparities among women are shaped by the experiences of White women and universal policies for Blacks are shaped by the experiences of Black men. Mapping the margins means that policies to address health disparities among Latinos assume that the experiences of Latinos that are racialized as Black, White or “some other race” are not necessarily equivalent by gender, class, ethnicity, or generational status (Griffith 2012; King 2006). It is also important that we examine the connections between race, gender, and class and that we interrogate the pathways of embodiment or make the connections between lived race-­gender at the micro, meso, and macro levels (Collins 2009; Crenshaw 1991; López 2003, 2011). Collins (2009) outlines two major multidimensional and multilevel concepts in the matrix of domination

186

Nancy López

theory: first, intersecting systems of oppression, which include race, gender, class, sexual orientation, age, and ethnicity; and second, a particular arrangement of hierarchical power relations along the macro/structural, meso/institutional domain, and micro/individual and interpersonal levels. Collins pays particular attention to the hegemonic (ideological) domain of power as the crosscutting “glue” that connects all these levels of power through the use of ideas and ideology. While the hegemonic domain of power aims to achieve social control and suppress dissent through dominant ideologies, there are always moments and social movements that engage in counter-­hegemony and resistance to domination and oppression in the form of social justice projects. For example, the civil rights movement and recently the movements for lesbian, gay, bisexual, transgendered, and questioning (LGBTQ) individuals and communities are examples of counter-­hegemonic projects that advance social justice in our society. I propose the concept of “racialized-­ gendered social determinants of health” as a key concept for understanding and ameliorating health disparities. This framework departs from the premise that intersecting racial and gendered inequalities are fundamental axes of inequality in their own right that cannot be subsumed as epiphenomena of something else, such as class, culture, or ethnicity (Collins 2009; Crenshaw 1991; Feagin 2006; Griffith 2012; Hurtado 1996; Krieger 1990; López 2003; Lorber 1994; Omi and Winant 1994; Roberts 1997; Schultz and Mullings 2005; Zuberi and Bonilla-­Silva 2008).8 Another fundamental assumption of this framework is that understanding and interrupting the race-­gender gaps in health require what Chapman and Berggren (2005) refer to as “radical contextualization,” or a meticulous attention and thick descriptions of the social contexts that produce inequalities in health for entire categories of people. This will require that we link micro-­level individual experiences to multi-­level phenomena in context. The racialized-­gendered social determinant of health framework consists of two major concepts: “lived-­race gender” and “racialized-­gendered pathways of embodiment.” Lived-­race gender refers to the everyday experiences related to one’s intersecting ascribed racial and gender social status in society. Examining the unearned privileges or disadvantages related to one’s intersecting race-­gender social status in a given context can capture lived race-­gender. The racialized-­gendered pathways of embodiment refer to the cumulative and life course effect of everyday microaggressions as well as the impact of racialized-­ gendered contexts in shaping health status and health outcomes.9 It is important to clarify that the concept of racialized-­gendered social determinants of health does not dispute the roles of social forces in shaping



Contextualizing Lived Race-­G ender

187

biological pathways of embodiment, such as the real biological changes in stress levels related to the contingent experiences of race-­gender stigmatization that can have a cumulative effect and result in negative health outcomes (see Feagin and Sikes 1994; Geronimus, this volume; Gravlee 2009; Krieger 1990; López 2003; Schultz and Mullings 2005; Williams et al. 1997). Nor does this framework question the real physiological differences between males and females, or what Fausto-­Sterling (1993) refers to as the five sexes. Instead, the racialized-­gendered determinants of health framework simply acknowledges that there are social forces that that result from intersecting racialized and gendered social hierarchies at the individual/micro level, the institutional/ community/meso level and the societal/structural level that have implications for health outcomes and disparities (Harris and Sim 2002; Hurtado 1996; Jones et al. 2008; Omi and Winant 1994; Richardson et al. 2011; Roberts 1997; Schultz and Mullings 2005; Weber 2007). Given that everyone is simultaneously racialized and gendered in a given society, whether these racialized and gendered inequalities translate into cumulative unearned social advantages or cumulative disadvantages in either a particular social setting or over the life course requires empirical scrutiny and meticulous contextualization of pathways of embodiment (Chapman and Berggren 2005). And now we turn to my own experiences with lived race-­gender as examples of the social forces of intersecting race and gender hierarchies in a given context to shape access to health care and health outcomes. Autobiographical Encounters with Lived Race-­G ender

Over the course of the last few decades, I have often engaged in conversations about the meaning of race with the physicians and medical practitioners I have encountered. I inquired about medical practitioners’ personal views as well as what was relayed to them about the “race” concept in their formal medical training. To my chagrin, I often encountered troubling essentialist and biodeterministic conceptualizations. For example, in the spring of 1988, I completed my first year at Columbia University in New York City and, a causality of overindulgence in dorm food, I embodied the infamous “freshman fifteen.” By the beginning of the fall semester of my sophomore year, I had successfully lost the weight, but I began to experience acute abdominal pain in my upper-­right abdomen. After several months of reporting this chronic discomfort, I finally ended up in an emergency room where a sonogram revealed that I had many small gallstones. When I returned to my primary care physician, a middle-­aged White man, to discuss my treatment options, he joked that he had not suggested a sonogram for gallstones because I was not a typical textbook case—­in his

188

Nancy López

words, “I was not fat, fair, and forty.” Because I was a dark-­skinned nineteen-­ year-­old who was not overweight, my physician could not “see” the clear-­cut, classic symptoms of gallstones I was manifesting. Had the physician inquired about other triggers for gallstone formation, such as rapid weight loss, I might have received better treatment: “race-­gender profiling” impeded my timely access to quality health care (see Epstein 2007 and Helms and Mereish, this volume, for more on the pitfalls of racial profiling in medicine and the simultaneous nondebate on sex profiling in medicine). One day in my early thirties and in the third trimester of my first pregnancy, I reached the door of the apartment complex where I lived at the same time as an older White man. Since I had multiple grocery bags in tow, I thanked him for opening the door and proceeded to enter the building. The man berated me for not ringing the doorbell so I could be buzzed in. Smiling, I said, “What makes you think that I don’t live here?” The man was speechless and apologetic. Again, it may be the case that in the apartment building where I lived in Albuquerque, there were few if any people of African descent or dark-­skinned Latinos, and this man was simply hoping to “protect” his home from potential intruders. However, this example also brings into sharp relief the reality that middle-­class privilege does not protect people of color from race-­gender microaggressions in their own homes (Feagin and Sikes 1994; López 2003; see also Bonilla-­Silva 2003 for an example of a Black Latino professor being mistaken for a construction worker while gardening in his own home or the national coverage on Professor Henry Louis Gates’s experience of being mistaken for a thief as he entered his home in Massachusetts). This reality has particular relevance for women of color of all class backgrounds who are pregnant, and have to deal with daily microaggressions that may be related to being “pregnant while Black” or “pregnant while Brown.” Another story that serves as a window to lived race-­gender is an experience I had when I delivered my first daughter in 2002 at one of the hospitals in Albuquerque. Several hours after my prolonged early morning delivery, the nurses, all of whom were White, instructed me to go to the basement of the hospital to fill out paternity papers. I assumed that this was standard protocol, so my husband, a Chicano whose is frequently assumed to be a recent immigrant from Mexico despite the fact that his family has lived in New Mexico for several centuries, and I dutifully made our way to the basement.10 When we got there, the clerk, a woman of color, asked us if we were married. When we said yes, she informed us that we didn’t have to fill out the papers after all. Apparently, the nursing staff had assumed that I was an unwed mother. The controlling



Contextualizing Lived Race-­G ender

189

image of women of color, and Black women in particular, as unwed mothers has become so pervasive that no one bothered to ask (Collins, 2009). These autobiographic vignettes bring into sharp relief one of the key dimensions of contextualized “lived race-­gender” and the “racialized-­gendered social determinants of health.” What is relevant here is that lived race-­gender whether in the form of privilege or stigma depends on context and adds up across time as life course embodiment (Walters et al. 2011). The added stresses faced by women of color, and Black women in particular, may be part of the puzzle of why even middle-­class Black women give birth to children with lower birth weights than their White counterparts, as pregnancy is a site of racialization (Bridges 2012; Geronimus et al. 2006). As discussed by Geronimus (this volume), the studies of the birth outcomes of Arab-­named women in California after 9/11 were far more negative than before the terrorist attacks in the United States and the increased surveillance (Lauderdale 2006). What is so compelling about this finding is that the genetic or biological makeup of these women did not change, but the social forces surrounding their lived race-­gender experiences took a turn for the worse as they were subjected to new race-­gender microaggressions (see also Zaal and Fine 2007). More research is needed to assess and contextualize “lived race-­gender,” and specifically what I call “pregnant while Black” or “pregnant while Brown” (Bridges 2011; Gómez 1997; Roberts 1997). I also encountered race-­gender profiling in medicine when I queried my pediatrician about my daughter’s umbilical hernia. The pediatrician, whom I believe was Asian American, said that umbilical hernias were more common among people of African ancestry than among people of European ancestry. But how did my pediatrician reach this conclusion? She relied on my daughter’s and my physical appearance, which was enough for her to summarily deduce that there was concordance between our phenotype and our genetic profiles. Most importantly, even if she had evidence of my so-­called African “ancestry informative markers (AIMs),” these AIMs would have only included select segments of populations in Africa (Graves, this volume). Even if my daughter’s pediatrician had the AIMs that traced my daughter’s genetic history to Africa, correlation does not indicate causation (Zuberi and Bonilla-­Silva 2008). It would still be unlikely that my daughter’s genes alone “caused” her to have an umbilical hernia. It would be equally important to explore the environment and social forces shaping my daughter’s embodied health status, including during my gestation and labor. Could umbilical hernias be more common among people of African ancestry in the U.S. context because of concrete differences in race-­gender experiences of stigma and privilege and pathways and mechanisms

190

Nancy López

of embodied health, such as induced labor, prolonged delivery, cumulative race-­gender microaggressions and discrimination during pregnancy and delivery? We don’t know the answers to these questions because physicians and researchers have not asked them: they revert to thinking of race in essentialist, biological terms or what Montoya (2011, 184) terms “bioethnic conscription.” In short, genetic reductionist explanations generally fail to account for the interaction of genes with other social forces, such as race-­gender profiling, microaggressions and social context in shaping illness and disease (Duster 2006; Graves, this volume; LaVeist 1994; Roberts 1997). Another increasingly common yet equally troubling conceptualization of race that circulates in the medical industry is the antiessentialist, color-­blind understanding of race. While seeking treatment for my daughter during an asthma attack, I mentioned to the resident doctor in the emergency room that I was organizing a multidisciplinary National Institutes of Health (NIH) Workshop on Best Practices for Conceptualizing and Operationalizing “Race” in Health Policy Research at the University of New Mexico. This physician, a young woman who I believe was Asian American, seemed perplexed about the premise of the workshop. She informed me that her medical training precluded her from inquiring about any patient’s race; she reminded me that under the Hippocratic oath she was sworn to treat “prisoners and millionaires” (read: people of color and Whites) the same.11 In other words, her training advocated a so-­called color-­blind approach to medical practice (Bonilla-­Silva 2003; Frankenberg 1993). Since the Human Genome Project discovered that “race” is genetically meaningless, this approach maintains that it makes no sense to invoke race as an analytical concept at all. In a similar vein, the American Anthropological Association (1997) asserts that we should eliminate the use of the term “race” from data collection and instead use the term “ethnicity.” The pitfalls of this seemingly antiessentialist, color-­blind approach to race is that, like the genetic reductionist understanding of race, it disregards the reality of historic and ongoing racialization processes that stigmatize entire groups of people over long periods of time that cannot be deduced to ethnicity. A more productive conceptualization of race would begin from the social constructionist approach (Morning 2011). And finally, perhaps the most common conceptualization of race I have come across when interacting with physicians and other medical personnel is that idea that race is an uncomplicated proxy for ethnicity and or social class. What these two hegemonic discourses share is that belief that racial disparities in health are really just due to “cultural” and “behavioral differences” or variations that vary according socioeconomic status (income, educational attainment, occupational status, net worth, and so on) or ethnicity, national



Contextualizing Lived Race-­G ender

191

origin, and culture. The problem with these conceptualizations is that they reduce race to an epiphenomenon of something else, either ethnicity or class (Omi and Winant 1994). The conceptualization of race as simply a proxy for ethnicity also ignores the differences in power and lived experiences among entire categories of people by race that cannot be reduced to cultural practices or national origin or social class and wealth. Moving toward Conceptual Complexity: Multidimensional and Multilevel “Race” and Ethnicity

If our goal is to fully comprehend the social forces that are eventually manifested as health disparities, as researchers we should start by posing the following questions: What dimension of the social construction of “race” am I studying? What level of “race” am I excavating? For example, if I am interested in individual, micro-­level “race” data, am I collecting data on how people self-­ identify or ascribed racial status? Am I interested in capturing lived race by asking questions about everyday discrimination? Moving one level up, if I’m interested in institutional, meso-­level data, am I collecting data on neighborhood context, local policing practices, and school policies? If I’m interested another level of the social construction of race, I may be collecting structural data, such as state or national policies that may be contributing to structured inequalities for entire categories of people. For example, race-­gender profiling in policing and the prision industrial complex or “papers please laws” being enforced in Arizona, as well as laws requiring particular ID or barring felons, disproportionately disenfranchise communities of color and particularly men in these communities. Regardless of what level of race data one is collecting, one should aim to for transparency by specifying how one is linking individual level to institutional level and structural levels of inequality (Williams 1975; Williams et al. 1997, 2010). One should also aim to engage the simultaneity and connections between “race,” gender and class inequality. Below I schematically represent a conceptual model for “race” as a dynamic and multidimensional social construction (see figure 12.1). I identify at least four distinct dimensions of “race” at the individual level: (1) racial self-­ identity, (2) ascribed racial status (folk race or social race), (3) lived race-­gender and life course embodiment, and (4) tribal status/political status. All of these facets of “race” are important for understanding health disparities.12 Racial Self-­I dentity

In order to facilitate national and interagency comparisons, racial self-­identity should be assessed using the 1997 and 2003 OMB guidelines. Specifically, we

192

Nancy López

Figure 12.1   Multidimensional “Race” Data at the Individual,

Micro-level

should first collect information on Hispanic ethnicity as two mutually exclusive choices (Hispanic versus non-­Hispanic). Next, for the “race” question, respondents should be given the option to mark more than one “race” from the official OMB defined categories. Although OMB guidelines do not specifically mention the use of “some other race” in their guidelines, the U.S. Census regularly includes the “some other race” category. Whenever possible, “some other race” should be included as it could be analytically valuable for discerning differences in health outcomes, and people should be able to mark more than one race. Another follow-­up question that can be used for those that mark more than one race is, “What is your primary racial identity?” Ascribed Racial Status (aka Social Race, Folk Race, Cultural Race, Socially Determined “Race”)

Perhaps one of the most important dimensions of “race” that needs to be captured in health disparities research is socially defined race (see also Gravlee and Dressler 2005 for more on multidimensional measures of race and the importance of folk race in shaping blood pressure in Puerto Rico; Wagley 1968 for a description of social race in Latin America and the Caribbean). Using the Reactions to Race module of the 2004 Behavioral Risk Factor Surveillance System



Contextualizing Lived Race-­G ender

193

developed by the Centers for Disease Control and Prevention, Jones et al. (2008) and her colleagues measured socially defined race through the following question: “How do others usually classify your ‘race’ in this country?” They found that socially defined “race” measures trumped self-­identified “race” in terms of predicting health status, even after controlling for socioeconomic variables such as income and educational attainment. In other words, if I identify as Black Hispanic but reported that others usually classified me as White, my health status would approximate that of self-­identified Whites, who had the best self-­reported health status of any of the racial or ethnic groups. This finding is of particular importance for health disparities researchers (see also Gravlee and Dressler 2005). Whenever possible, researchers should include measures of ascribed racial status as a key dimension of social stratification with particular relevance for health outcomes. Other questions that can capture ascribed racial status include “Has anyone ever mistaken you to be of a different race? Yes or No.” If “yes,” the instructions may include the following question: “Rank the top three racial group(s) you have been confused for most often in descending order, starting with the one you are most often mistaken for and so on.” For each listed group, one might include a chart with check boxes asking respondents to check the context(s) where this misidentification happens most often for each reported race (for example, when walking down the street in my neighborhood, when shopping in a store, in a social gathering at work, in a hospital, in school, in interacting with law enforcement, when applying for a job, while at work, while boarding an airplane, while traveling abroad, or other context). While adding these questions to any data collection will surely add to the cost, the goal of all data collection should be guided by meaningful use for civil rights enforcement officials, researchers, and the communities in question. This value-­added question on ascribed race is especially important in a state like New Mexico, where, according to the U.S. Census, almost half of the population identifies as Hispanic but the majority identify as White; this seemingly paradoxical pattern may be explained by examining the unique history of New Mexico. When New Mexico was annexed to the United States, all of the Hispanic residents were legally defined as White as racial status was a precondition for citizenship. This is why in a recent census, close to two-­thirds of New Mexicans identify as White although socially they occupy what Gómez calls an off-­White racial status whereby they are not socially defined as White (Gómez 2007; U.S. Census 2012). Many Hispanics can report numerous occasions where they are mistaken for “some other race,” particularly Mexican or Native American, when seeking housing, while in public

194

Nancy López

institutions such as hospitals and schools, while shopping or in public spaces or when interacting with law enforcement. Other questions that could be used for capturing ascribed racial status could include combined measures of skin color and phenotype and folk race, as well as interviewer-­identified race (Candelario 2007; Garcia, this volume; Gravlee and Dressler 2005; Sanchez and Ybarra, this volume; Saperstein, this volume). All of these value-­added questions could provide insight to the ways in which racial status gets translated into health inequality. Lived Race-­G ender

Perhaps the most challenging yet most important dimension of “race” to that requires systematic data collection is lived race-­gender. (See also López 2003 for more on race-­gender experiences.) In order to capture lived race-­gender in survey format, one can build on the everyday racism scale generated by Williams et al. (1997). The everyday racism scale provides a tool for survey researchers interested in capturing the frequency and context of discrimination perceived by individuals. Questions from this scale explore the frequency and specific instances of discrimination (for example, respondants are asked about being treated with less courtesy/respect than other people, as well as being harassed/threatened). Respondents on this scale are then prompted to indicate whether they feel they are treated this way because of their national origin, gender, race, tribe, sexual orientation, and so forth (Williams et al. 1997). To capture what I call gendered racism, and in particular the lived race-­ gender microaggressions that women of color experience differently from their male counterparts, I would add questions such as, “How often do people assume you are an unwed parent? Do people assume that you are receiving public assistance? Do people assume that you have a lower education or occupational status than others? Do people assume that you are uninsured when accessing health care? How often do you experience sexual harassment that is tied to your race? How often do people assume that you are a drug dealer? How often are you profiled as a potential thief and followed in stores?” These types of questions are crucial for capturing unequal treatment faced by women of color in the United States and their experiences with lived race-­gender. Repeated microaggressions, such as being mistreated when seeking public services, being profiled as a potential criminal when shopping, and having to show proof of legal status because one is perceived to be an “illegal alien,” can cause physiological harm that can be translated into health disparities such as advanced biological age, elevated blood pressure, and higher general stress



Contextualizing Lived Race-­G ender

195

levels (Geronimus 1996; Hudson et al. 2012; Krieger and Sidney 1996; LaVeist 2005; Oliver and Shapiro 1995; Williams et al. 2010). Participant observation and ethnography are particularly important methodological tools for capturing lived race-­gender. Other data collection options include qualitative in-­depth interviews or focus groups in a neutral place where participants may feel more able to talk freely about their experiences (Chapman and Berggren 2005; Feagin and Sikes 1994; Geronimus, this volume; Gravlee, this volume; Gravlee and Dressler 2005; Morris 2007). Geronimus (this volume) emphasizes the need for more radical contextualization of the social construction of “race” as a lived experience that is context-­and situation-­specific and can eventually become embodied (Chapman and Berggren 2005; Walters et al. 2011). Part of this data collection requires that researchers, scholars, and policy makers engage in self-­reflexivity by embodying themselves and consistently asking themselves how their own lived race-­gender and social class locations as well as their academic training shapes their understandings of inequality. This transparency about positionality as well data collection and analysis can provide the missing link for dismantling racial discrimination across a variety of institutions and in turn contribute to ending health disparities (Bridges 2012; Fine et al. 2000; Foster et al. 2000; Geronimus et al. 2006; Hoberman 2012; Hudson et al. 2012; López 2003; Williams et al. 1997). Tribal Status

Because Native Americans are sovereign nations located within the United States, it is important to conceptually distinguish tribal status as a separate dimension of race distinct from racial self-­identity or ethnicity (Garroutte 2001; Huyser et al. 2009; Tallbear 2008). Currently the U.S. Census inquires about Native American identity as a race not a political status. As a conceptually distinct dimension of “race,” tribal status goes beyond just capturing racial identity or ancestry. Specifically, tribal status inquires about tribal enrollment and tribal residence.13 Why is this important? Using 2000 American Community Survey data, Huyser et al. (2009) found stark differences in terms of the educational and employment outcomes among Native Americans who mark that they are solely Native American and those who identify as Native American and any other race. Native Americans who mark that they are only Native American have the lowest educational outcomes and employment rates of all Native Americans. Males in particular had the lowest educational attainment and labor market status of any other group. Given these complexities, how do we collect meaningful data on health disparities in Native American communities in the United States? In the case

196

Nancy López

where a respondent indicates that her racial self-­identity is Native American, four additional questions should be asked: (1) Which tribe, pueblo, or Indian tribe do you identify with (you may mark more than one)? (2) For each tribe you identify with, please identify your tribal status (for example, enrolled or not enrolled). (3) Did you grow up on tribal lands? If so, please specify tribe(s) and number of years you lived in each tribe. And (4) Please indicate where you reside currently (for example, tribal lands or urban setting). By adding these conceptually important questions, we could discern if the patterns of health disparities—­for example cancer, diabetes, heart disease, obesity, and suicide rates among heterogenous Native Americans in the United States—­are unique to particular tribal communities (both enrolled or not enrolled) or unique residential contexts (for example, differences among those individuals who reside/grew up on tribal lands versus those who reside in urban settings; see also Garroutte 2001; Huyser et al. 2009; Snipp 2003; Tallbear 2008; Walters et al. 2011). This is important for research on health disparities in Native American communities because in the 2010 Census nearly half (44 percent) of Native Americans identified with more than one race. Again, the guiding question for data collection on racial disparities by race should be, “does this data help me understand the health disparities that indigenous communities experience?” Including the aforementioned questions in the U.S. Census or the American Community Survey or any other research instrument will help us discern whether these health disparities are more pronounced among Native Americans who live on tribal lands versus those who live in more urban settings. Furthermore, information on specific tribal identities and residence will help elucidate the complexities of intertribal heterogeneity in terms of socioeconomic status, including income, educational attainment, and labor market status and sociohistorical context. Of course, ethical considerations that are anchored in Native American tribal sovereignty and that are attentive to the potential for individual and tribal harm should guide all data collection, analysis, and publication (Indigenous Education Study Group 2010; Trans-­disciplinary “Race” Working Group 2010). “Some Other Race” as Analytically Important

Some policy makers and scholars have argued that it is important to include the term Hispanic as one of the “races” listed on the U. S. Census (Brown et al. 2006b; Prewitt 2010; Roth 2012). The 2010 Alternative Questionnaire Experiment Report found that adding Hispanic to the list of races can reduce the number of “some other race” answers (Compton et al. 2012). Regardless of intention, this proposal is problematic because it may inadvertently encourage



Contextualizing Lived Race-­G ender

197

all Hispanics (regardless of their racial identity) to mark just one race. This might mean that Hispanics who under the previous format marked Hispanic and White would now just mark White, thereby reducing the data available for interrogating inequalities in a variety of domains including housing segregation, education, income, and so forth. The current format, as per OMB guidelines, allows for rich multidimensional analysis among Hispanics and as Saperstein (this volume) points out, this is rich data that has not adequately been analyzed (see Rodriguez et al. 2011; Saperstein, Garcia, and Sanchez and Ybarra, this volume). Any attempt to collect meaningful data on Hispanics should continue to collect Hispanic origin and “race” as two separate questions as indicated in the 1997 and 2003 OMB standards (for a multidimensional conceptual model of the dimensions of ethnicity, see figure 12.2). In addition, the Hispanic origin question instruction should prompt individuals to further designate their national origin or ancestry as Mexican, Chicano; Puerto Rican; Cuban; Dominican; Manito (Hispano Nuevomexicano); or any other Latino ethnic or national origin and indicate that they can mark more than one more ethnic origin. And finally, the instructions preceding the ethnic origin and race questions for the 2010 census, stating that “for this census, Hispanic origins are not races,” should be dropped. These instructions were added for the first time in the 2010 census and they cause much consternation among Hispanic respondents. Instead definitions of the difference between race and ethnicity should be included. Collecting data on ethnicity and racial status via two separate questions adds value to the research conducted by health disparities researchers, scholars, policy makers, and communities interested interrogating inequalities. Logan (2003) found that there are meaningful differences in the educational attainment and employment outcomes among those Hispanics that identify their “race” as White, those who identify their “race” as “some other race,” and those who identify their “race” as “Black” even after controlling for socioeconomic status. All of these differences followed a pigmentocratic logic, whereby Latinos identifying as White fared better than the others. The “some other race” category in the U.S. Census is particularly important for Latinos/Hispanics, as 97 percent of individuals who mark “some other race” in the 2000 census are Hispanic; however, this differs significantly by the ethnicity or national origin as well as generational status of a given Hispanic group. In the 2010 census, just over half (53 percent) of Latinos marked their race as White; 37 percent marked their race as “some other race” and 3 percent marked Black. Cubans were the outliers in terms of having the highest number of people who identify as White (85 percent), while Dominicans were the Latino group with the largest number of people identifying their “race” as Black (13 percent; Cobas et al. 2009; Ennis

198

Nancy López

et al. 2011; Rodriguez et al. 2011). While some may view the “some other race” racial category as a problem that needs to be addressed because it is selected by over a third of all people who identified themselves as Hispanics in the 2010 census, I view it as important source of data for understanding health disparities among diverse Hispanics groups who may have members of the same family that nevertheless occupy very different racial statuses (Brown et al. 2006a; Rodriguez 2000). Ethnicity as Multidimensional

Ethnicity refers to one’s cultural background, language, national origin, customs, food preferences, ancestry, behavior, assimilation and acculturation, and sometimes religious background (Ford and Harawa 2010; Waters 1990). Health disparities researchers should aim to include questions that capture the different facets of ethnicity represented in figure 12.2. (See Helms and Mereish, this volume, for the need for constructs grounded in psychology research on depression and the pitfalls of essentializing individuals based on “ethnic” or racial traits.) The additional dimensions of ethnicity we should consider collecting data on include self-­identified ethnic identity, ancestry and national origin,

Figure 12.2   Multidimensional Ethnicity as Distinct from “Race”



Contextualizing Lived Race-­G ender

199

primary language and heritage language, and cultural practices, beliefs, values, food, and religion. These conceptual distinctions are particularly important for Hispanic/Latino groups that do not constitute a homogenous culture. For this chapter, I focus on the importance of generational status, legal status, and citizenship. Generational Status

Parental place of birth was last collected in the U.S. Census in 1980. Since 1994 only the Current Population Survey includes a measure of parental place of birth. To assess the impact of generational status, one major change that would have to occur in the U.S. Census and especially the American Community Survey would be to include questions on parental place of birth. Currently the American Community Survey has an ancestry question, but this does not allow us to separate the experiences of immigrants who came to the U.S. as adults (first generation), the U.S.-­born children of immigrants (second generation) or the grandchildren of immigrants (third generation). This is particularly important information for health disparities researchers in that there may be substantial differences in health, educational attainment, and labor market outcomes among people who share the same ethnicity but vary by generational status (Telles and Ortiz 2008). While collecting data on citizenship and legal status can be problematic, health disparities researchers need to be conceptually attentive to the dynamics that may be affecting those immigrants that lack legal status as they may be subjected to physical and mental health contexts that result in health disparities due to their legal status and or generational status. For example, several municipal and state governments have instituted law and surveillence systems that target undocumented immigrants and may in turn cause higher stress levels among these populations. Making the Connections: Micro-­, Meso-­, and Macro-­L evel Pathways of Embodiment

To take seriously that “race” is a multilevel social construction would mean that health disparities researchers would examine structural racism and systemic racism, and map racialized power and privilege dynamics and hierarchies as manufactured through formal and informal practices in our all of our social institutions (Bonilla-­Silva 2001, 2003; Gómez, this volume; López 2003, 2011; Omi and Winant 1994). Rather than accept biodeterministic models that assume that human disease and experience are a function of innate genetic or biological phenomenon, we need to map the constellation of contexts, whether situations, environments, and social structures that overlap and produce

20 0

Nancy López

inequality. As decribed by Montoya (2011, 184): “Genes do not cause chronic disease. Genes in certain bodies under certain conditions contribute to disease susceptibility.” Figure 12.3 visually represents a conceptual model of the racialized-­ gendered social determinants of health as an inverted a pyramid to allude to the connections via micro, meso and macro social forces. By necessity, this model is heuristic; it seeks to sketch the types of research that we should pursue in order to develop a more comprehensive strategy for studying multilevel race in the health disparities area. Each stratum delineates the different levels of social reality where health disparities researchers can highlight the social construction of racialized social structures at the macro level, including law, social policy, ideological representations, and so forth; at the meso level in the form of techniques of surveillance in bureaucratic institutions and community and local context; and at the micro level of individual identity and lived race-­gender. Each of these levels of the racialized-­gendered social determinants

Macro-Level: Structural Arrangements/Structural Domain of Power; Federal, State Laws; Hegemonic Ideology; Controlling Images; Social Movements Meso-Level: Institutional Practices and Disciplinary Domain of Power Rules, Surveillance, Institutional Discretion Micro-Level: Individual Data

(see previous figure 12.1)

Figure 12.3  Racialized and Gendered Social Determinants of

Health: Multilevel “Race” Data at the Micro-, Meso-, and Macrolevels Note: Intersectionality is a fundamental organizing principle of this model. “Intersectionality” refers to “analysis claiming that systems of race, social class, gender, sexuality, ethnicity, nation, and age form mutually constructing features of social organization” (Collins 2009, 320).



Contextualizing Lived Race-­G ender

201

of health underscore that “race,” ethnicity, and gender are dynamic, multifaceted social constructions that are visible at the individual, institutional, and structural levels of society and that intersect with other systems of oppression including gender, class, and sexual orientation to produce distinct social locations for entire categories of people (Collins 2009; Crenshaw 1991; Jones 2000; LaVeist 1992, 1993, 1994). At the macro level, we need data at the national level and at the global level, for example, international comparisons of race-­gender inequality across the globe. In the U.S. context, we need comparative data that will allow us to excavate federal laws and policies (both formal and informal) that form the backdrop for ongoing structural gendered racism and its manifestations in health outcomes. For example, Williams et al. allude to the importance of examining the simultaneity of race and gender to understand and intervene in health disparities: We will be attentive to the role of gender and present gender differences whenever the data are available. It has recently been argued that although Black women lag behind other social groups on some societal indicators, they are nonetheless rapidly becoming a “model minority” on a broad range of indicators. For example, Black women have a higher rate of college enrollment than Black males but also than Whites and Hispanics. In addition, they also have lower suicide rates than Black males and Whites and low rates of crime, cigarette smoking, alcohol use, and the use of illegal drugs. (2010, 70)

In another study, Williams and Jackson (2005) find that homicide rates for Blacks was six times higher than for Whites. They argue that death from heart disease, cancer, and stroke are far more important contributors to high Black mortality rates among both men and women than homicide, which mostly affects men. Studies show that more education extends longevity for all demographic groups except Black males and that middle-­class Black men have higher rates of depression than their non-­Black peers (Griffith 2012; Husdon 2012; Olansky et al. 2012). Williams and Jackson (2005) also point to the high impact of residential segregation on health disparity. This means health disparities researchers should excavate how current housing and zoning laws allow environmental racism and toxic dumping in low-­income communities, and disproportionately in communities of color. Informal practices in the mortgage industry encourage higher rates for Black and Latino communities. To unravel the controlling images (Collins 2009) of people of color, we need to unmask the prevailing ideologies circulated through the mass media about the root cause of health

202

Nancy López

disparities by race. The media often portray health disparities as genetically caused or a matter of individual choices rather than examining how federal policies in housing, employment, and education may contribute to structured inequality among entire categories of people. At the meso level, we need data that would allow us to contextualize and map local-­and state-­level, and even community-­level dynamics. To accomplish this, we need to engage in radical contextualization (Chapman and Berggren 2005) of the environments that create racial inequalities, whether they are in schools or local neighborhood association zoning committees, local governments or hospital administrations (Bridges 2011). In other words, we need to highlight what Collins (2009) defines as the techniques of surveillance in bureaucracies that regardless of intent may be contributing to racial disparities in health. We need to connect micro-­level “race” data to meso-­level and macro-­level race processes. At the micro level, we need to explore individual-­level identities and lived race-­gender and experiences. This work will require longitudinal data that captures life course embodiment (Saperstein, this volume), as well as more data on the contexts that produce inequalities in health status (Geronimus, this volume). In short, researchers, scholars, policy makers and community members interested in excavating structural racial inequality necessarily must include not only the individual multidimensional “race” data I outlined in the previous section (see figure 12.1 and 12.2) but, whenever possible, data on the contexts that may be shaping these outcomes, including federal and state laws, bureaucratic policies, neighborhood resources or disadvantages, housing quality, school characteristics, microaggressions, and so on (Grady 2006). This mapping can take place by linking individual-­level data on race, ethnicity, gender, and so forth to meso-­level data (both quantitative and qualitative) on residential and social segregation in neighborhoods, schools, and other context measures of structural racism, including local measures of racialized inequalities such as race-­gender profiling and environmental racism. Federal funding for research that is anchored in genetic reductionist theory promotes the falsehood of “race” as genetics and biology and channels funding into this domain rather than mapping the racialized-­gendered social determinants of health. Only in this way can we identify the pathways and mechanisms that, although social in origin, become embodied as health disparities. Conclusion: Mapping and Contextualizing Race-­G ender and Health Disparities

How can we do a better job of collecting and analyzing data on racial disparities in health and beyond that can help us eradicate health disparities? Among the



Contextualizing Lived Race-­G ender

203

most important changes that can occur in the collection of “race” data for mapping health disparities is a paradigm shift that contextualizes the racialized-­ gendered social determinants of health. This will mean that all health providers will have to be trained in the social determinants of health as well as the social construction of race and gender as intersecting systems of inequality such as social class. While I recognize that there will never be a perfect system for collecting “race” data for health disparities, at the very least we should clarify what dimension or level of the social construction of race we are researching. Although the conceptual models I propose are focused on health disparities research, they can travel across other social institutions or policy-­making arenas, including education, the criminal justice system, housing and voting rights, and the media and the arts. Anchored in theories of intersectionality, Schultz and Mullings articulate our task: “The challenge we now encounter is how to understand the ways in which gender, race and class relations intertwine and are expressed in disparate chances for health, illness and well being” (2005, 6). Two questions remain: Do we have the political will to not only revise the “gold standard” of collecting just self-­identified “race” data to add other dimensions of race in order to achieve more meaningful and complex understandings of lived race-­ gender and the racialized-­gendered social determinants of health? Do we have the moral conviction to construct macro-­and meso-­level federal, state, and local policies that aim to eradicate the racialized and gendered social determinants of health? It is my hope that the work in this volume identifed some promising strategies for the collection and analysis of “race” data that can shed light on the contexts and policies that can advance social justice and eradicate health disparities. Acknowledgments

I am forever grateful for the extensive feedback and provocative comments provided by Dr. Laura Gómez through multiple iterations of this chapter. I also appreciate the thoughtful feedback provided by Dr. Raquel Rivera. Thank you both for pressing me to clarify my ideas. ¡Gracias colegas queridas! Notes 1. I place “race” in quotation marks to call attention to common-­sense understandings anchored in the myth of “race” as a genetic or biological reality. My use of the quotation marks in no way is meant to diminish the importance of understanding the origins and consequences of the social construction of “race” as a fundamental axis of social stratification in the United States and beyond.

204

Nancy López

2. The forms usually list the standard “race” and ethnicity options set by the Office of Management and Budget (OMB). In addition, clerks also regularly inquire about my religious affiliation, but not my socioeconomic status. Future forms should also include measure of socioeconomic status (Marmot 2006). 3. In New Mexico, a state with a Black population of less than 3 percent, people of so-­called African phenotype are seen as a novelty. For example, at a new faculty reception in 2001, a white male administrator pointed at me and smiled from across the room. Upon meeting, he remarked how much I resembled Secretary of State Condoleezza Rice. While I can imagine that this man was trying to create conversation and connect with me by comparing me to a highly accomplished African American woman, this episode reveals how race and racial status in particular is not equivalent to ethnicity. It also points to the reality that faculty of color have qualitatively different experiences than their White counterparts who may not be subjected to jokes about their racial status. 4. For more on the experience of Afro-­Latinos, see also Aparicio 2006; Candelario 2007; Cobas et al. 2009; Jiménez Román and Flores 2010; López 2003. 5. It is important to clarify that OMB guidelines do not preclude the collection of additional data on race and ethnicity; however, they do stipulate that whatever additional data collection occurs needs to aggregate the aforementioned standard OMB ethnic and racial categories. 6. Part of this transparency involves critical self-­reflection on the part of the researcher on her/his positionality, what I call “embodying the researcher.” To embody the researcher, we must always ask ourselves, how our own lived experiences and academic training shape how we understand “race” and ethnicity. 7. I do not place gender, ethnicity, or class in quotes because generally these concepts have not been conceptualized as biological realities (that is, as innate and unchanging genetic realities). Yet “race” continues to be conceptualized in both popular and so-­called scholarly circles as an essential difference among entire categories of people (Morning 2011). The only exception would be gender, which is often erroneously conceptualized as interchangeable with the concept of “sex,” which is also socially constructed (Epstein 2007). While I recognize that both gender and race are viewed as biological essences, for the purposes of this chapter, I focus on the social construction of race-­gender, which I do not place in quotes. 8. The racialized-­gendered social determinants of health framework departs from the premise that class, sexual orientation, age, disability status, and other axes of stratification are always important for understanding and dismantling health disparities; however, for the purposes of this chapter I highlight race and gender as salient categories of analysis for health disparity. 9. The concept of “lived race-­gender” builds on my previous “race-­gender experience framework” and the concepts of “race-­gender experiences” and “race-­gender outlooks,” which more directly relate to the realm of education (López 2003, 6). The concepts of “lived race-­gender” and “racialized-­gendered social determinants of health” more directly capture the impact of the intersection of race-­gender processes in the field of health (see also Brown et al. 2006 for a discussion of lived race and health status among Latinos). 10. Although my husband’s family can trace its genealogy over several centuries in New Mexico before it became a U.S. state in 1912, in many social circumstances in the United States, he is assumed to be an immigrant. This is in large part due to racist media projects depicting dark-­skinned Mexicans as criminal “illegal aliens.” My



Contextualizing Lived Race-­G ender

205

husband recalls being stopped and frisked by police officers on a number of occasions because he fit the profile of suspected criminals. 11. Since this is a main hospital in Albuquerque, one often sees prisoners, most of whom are people of color, mostly Hispanic and Native American, both men and women, being brought to appointments in handcuffs and ankle shackles. In addition, the local newspaper regularly includes the mug shots of people who have been arrested for driving under the influence of alcohol or drugs, and most of these people would be racialized as people of color, again, mostly Hispanic and Native American. 12. I intentionally do not include “genetic ancestry” as a dimension of race (or ethnicity) at the individual or group level because human genetic differences do not map on to “race” (Graves, this volume; Williams et al. 2010). 13. In the U.S. Census, one can mark Native American “race” and even identify one’s specific tribe by writing it in. One need not be formally enrolled, recognized, or meaningfully connected to a Native American tribe in order to designate one’s race as Native American. Moreover, ancestry is also not interchangeable with racial status. References American Anthropological Association (AAA). 1997. “American Anthropological Association Response to OMB Directive 15: Race and Ethnic Standards for Federal Statistics and Administrative Reporting.” Washington, DC: AAA. http://www.aaanet.org/ gvt/ombdraft.htm, retrieved July 19, 2008. ———. 1998. “Statement on Race.” Accessed December 29, 2009. http://www.aaanet.org/ stmts/racepp.htm. American Association of Medical Colleges (AAMC). 2012. “AAMC Approves New MCAT Exam with Increased Focus on Social, Behavioral Sciences.” AAMC Reporter, March 2012. Last accessed December 17, 2012. https://www.aamc.org/newsroom/reporter/ march2012/276588/mcat2015.html. American Association of Physical Anthropologists (AAPA). 1996. “AAPA Statement on Biological Aspects of Race.” American Journal of Physical Anthropology (101): 569–­ 70. Accessed August 15, 2010. http://physanth.org/association/position-statements/ biological-aspects-of-race. American Psychological Association (APA). 2002. Guidelines on Multicultural Education, Training, Research, Practice, and Organizational Change for Psychologists. Washington, DC: American Psychological Association. Accessed August 15, 2010. http://www.apa.org/pi/oema/resources/policy/multicultural-guidelines.aspx. American Sociological Association (ASA). 2003. The Importance of Collecting Data and Doing Social Scientific Research on Race. Washington, DC: American Sociological Association. Aparicio, A. 2006. Dominican Americans and the Politics of Empowerment. Gainesville, FL: University of Florida. Bonilla-­Silva, E. 2001. White Supremacy in the Post-­Civil Rights Era. Boulder, CO: Lynne Reinner. ———. 2003. Racism without Racists: Color-­Blind Racism and the Persistence of Racial Inequality in the United States. 2nd ed. Lanham, MD: Rowan and Littlefield. ———. 2004. “From Bi-­racial to Tri-­racial.” Ethnic and Racial Studies 27 (6): 931–­50. Bridges, K. 2011. Reproducing Race: Pregnancy as a Site of Racialization. Berkeley: University of California.

206

Nancy López

Brown, S. J., S. Hitlin, G. H. Elder Jr. 2006a. “The Importance of Being ‘Other’: A Natural Experiment about Lived Race Over Time.” Social Science Research 36 (1): 159–­74. ———. 2006b. “The Greater Complexity of Lived Race: An Extension of Harris and Sims.” Social Science Quarterly 87 (2): 411–­13. Burgess-­Proctor, A. 2006. “Intersections of Race, Class, Gender, and Crime: Future Directions for Feminist Criminology.” Feminist Criminology 1 (27): 27–­47. Burgos, Giovani and Fernando Rivera. 2012. “Residential Segregation, Socioeconomic Status, and Disability: A Multilevel Study of Puerto Ricans in the United States.” Centro: Journal of the Center for Puerto Rican Studies 24 (2). Candelario, G. 2007. Black Behind the Ears: Dominican Racial Identity from Museums to Beauty Shops. Durham, NC: Duke University Press. Carter, R. 2007. “Genes, Genomics, and Genealogies: The Return of Scientific Racism.” Ethnic and Racial Studies 30 (4): 546–­56. Chapman, R., and J. Berggren. 2005. “Radical Contextualization: Contributions to an Anthropology of Racial/Ethnic Health Disparities.” Health: An Interdisciplinary Journal for the Social Study of Health, Illness, and Medicine 9 (2): 145–­67. Cobas, J., J. Duany, and J. Feagin, eds. 2009. How the United States Racializes Latinos: White Hegemony and Its Consequences. Boulder, CO: Paradigm Publishers. Collins, P. H. 2009. Black Feminist Thought: Knowledge, Consciousness, and the Politics of Empowerment. 3rd ed. New York: Routledge. Compton, E. Bentley, M. Ennis, S., Rastogi, S. 2012. “2010 Census Race and Hispanic Origin Alternative Questionnaire Experiment Report. Decennial Statistical Studies Division and Population. Washington, DC: U.S. Census. Last accessed November 18, 2012. http://2010.census.gov/2010census/pdf/2010%20Census%20Race%20and%20 Hispanic%200rigin%20Alternative%20Questionnaire%20Experiment.pdf. Crenshaw, K. 1991. “Mapping the Margins: Intersectionality, Identity Politics, and Violence against Women of Color.” Stanford Law Review 43 (6): 1241–­99. Delgado, R., and J. Stefancic. 2011. Critical Race Theory: An Introduction. New York: New York University Press. Duster, T. 2006. “Comparative Perspectives and Competing Explanations: Taking on the Newly Configured Reductionist Challenge to Sociology.” American Sociological Review 17: 1–­5. Emirbayer, M., and M. Desmond. 2012. “Race and Reflexivity.” Ethnic and Racial Studies 35 (4): 574–­99. Ennis, S., M. Rios-­Vargas, and N. Albert. 2011. “The Hispanic Population: 2010.” 2010 Census Briefs. U.S. Department of Commerce, Economics and Statistics Administration. Washington, DC: U.S. Census Bureau. Last accessed December 18, 2012. http:// www.census.gov/prod/cen2010/briefs/c2010br-04.pdf. Epstein, Steven. 2007. Inclusion: The Politics of Difference in Medical Research. Chicago: University of Chicago Press. Fausto-­Sterling, A. 1993. “The Five Sexes: Why Male and Female Are Not Enough.” The Sciences: 20–­24. Feagin, J. 2006. Systemic Racism: A Theory of Oppression. New York: Routledge. Feagin, J., and M. Sikes. 1994. Living with Racism: The Black Middle-­Class Experience. New York: Routledge. Fine, M., et al. 2000. Off-­White: Readings in Race, Power, Privilege, and Resistance. New York: Routledge.



Contextualizing Lived Race-­G ender

207

Ford, C., and C. O. Airhihenbuwa. 2010. “Critical Race Theory, Race Equity, and Public Health: Toward Antiracism Praxis.” American Journal of Public Health 100 (Suppl. 1): S30–­S35. Ford, C., and N. Harawa. 2010. “A New Conceptualization of Ethnicity for Social Epidemiologic and Health Equity Research.” Social Science and Medicine 71: 251–­58. Foster, H., L. Wu, M. Bracken, K. Semenya, and J. Thomas. 2000. “Intergenerational Effects of High Socioeconomic Status on Low Birth Weight and Preterm Birth in African Americans.” Journal of National Medical Association 92 (5): 213–­21. Frankenberg, R. 1993. White Women, Race Matters. Minneapolis: University of Minnesota. Garroutte, E. 2001. “The Racial Formation of American Indians: Negotiating Legitimate Identities within Tribal and Federal Law.” American Indian Quarterly 25 (2): 224–­39. Geronimus, A. 1996. “Black/White Differences in the Relationship of Maternal Age to Birth Weight: A Population-­Based Test of the Weathering Hypothesis.” Social Science and Medicine 42 (4): 589–­97. Geronimus, A., M. Hicken, D. Keene, and J. Bound. 2006. “Weathering and Age-­Patterns of Allostatic Load Scores among Blacks and Whites in the United States.” American Journal of Public Health 96 (5): 826–­33. Golash-­Boza, T. 2010. “Does Whitening Happen? Distinguishing between Race and Color Labels in an African-­Descended Community in Peru.” Social Problems 57 (1): 138–­56. Golash-­Boza, T., and W. Darity. 2008. “Latino Racial Choices: The Effects of Skin Colour and Discrimination on Latinos’ and Latinas’ Racial Self-­identifications.” Ethnic and Racial Studies: 1–­36. Gómez, L. 1997. Misconceiving Mothers: Legislators, Prosecutors, and the Politics of Prenatal Drug Exposure. Philadelphia: Temple University Press. ———. 2007. Manifest Destinies: The Making of the Mexican American Race. New York: New York University Press. Gotanda, N. et al. 1995. Critical “Race” Theory: Key Writings That Formed the Movement. New York: New Press. Grady, S. 2006. “Racial Disparities in Low Birth Weight and the Contribution of Residential Segregation: A Multilevel Analysis.” Social Science and Medicine 63 (12): 3013–­29. Gravlee, C. 2009. “How ‘Race’ Becomes Biology: Embodiment of Social Inequality.” American Journal of Physical Anthropology 139: 47–­57. Gravlee, C., and W. Dressler. 2005. “Skin Pigmentation, Self-­Perceived Color, and Arterial Blood Pressure in Puerto Rico.” American Journal of Human Biology 17: 195–­206. Gravlee, C., and E. Sweet. 2008. “Race, Ethnicity, and Racism in Medical Anthropology, 1977–­2002.” Medical Anthropology Quarterly 22: 27–­51. Griffith, D. 2012. “An Intersectional Approach to Men’s Health.” Journal of Men’s Health 9 (2): 106–­12. Harris, D. R, and J. J. Sim. 2002. “Who Is Multiracial? Assessing the Complexity of Lived Race.” American Sociological Review 67 (4): 614–­27. Hoberman, J. 2012. Black and Blue: The Origins and Consequences of Medical Racism. Berkeley: University of California Press. Hudson, E., K. Bullard, H. Neighbors, A. Geronimus, J. Yang, and J. Jackson. 2012. “Are Benefits Conferred with Greater Socioeconomic Position Undermined by Racial Discrimination among African American Men?” Journal of Men’s Health 9 (2): 127–­36. Human Genome Project (HGP). 2012. “Genetic Anthropology, Ancestry, and Ancient Human Migration.” U.S. Department of Energy Genome Program’s Biological and

208

Nancy López

Environmental Research Information System (BERIS). Accessed July 2010. http:// www.ornl.gov/sci/techresources/Human_Genome/elsi/humanmigration.shtml. Hurtado, A. 1996. The Color of Privilege: Three Blasphemies on “Race” and Feminism. Ann Arbor: University of Michigan Press. Huyser, K., A. Sakamoto, and I. Takei. 2009. “The Persistence of Racial Disadvantage: The Socioeconomic Attainments of Single-­Race and Multi-­Race Native Americans.” Population Research and Policy Review 29 (4): 541–­98. Indigenous Education Study Group (IESG). 2010. “Indian Education in New Mexico, 2025.” Ted Jojola, Tiffany Lee, Adelmar Alcantara, Mary Belgarde, Carlotta Bird, Nancy López, Beverly Singer, co-­ principal investigators. Marie-­ Michele Jasmin-­ Belisle, Joi Burshia, Genevieve Giaccardo, Neomi Gilmore, Betty McCorkey, Shynoke Ortiz, Michael Vincent Romero, Leola Tsinnajinnie, Carolene Whitman, Natahnee Winder, student research assistants. Study contracted by New Mexico Public Education Department, Indian Education Division. Conducted by Eight Northern Indian Pueblo Council, Inc. Jiménez Román, M. and J. Flores. 2010. The Afro-­Latin@ Reader: History and Culture in the United States. Durham, NC: Duke University Press. Jones, C. 2000. “Levels of Racism: A Theoretic Framework and a Gardener’s Tale.” American Journal of Public Health 90: 1212–­15. Jones, C., B. Truman, L. Elam-­Evans, C. Y. Jones, R. Jiles, S. Rumisha, and G. Perry. 2008. “Using ‘Socially Assigned Race’ to Probe White Advantages in Health Status.” Ethnicity and Disease 18: 496–­504. King, D. 2006. “Multiple Jeopardy, Multiple Consciousness: The Context of Black Feminist Ideology.” In Race, Gender, and Class: Theory and Methods of Analysis, ed. Bart Landry. Upper Saddle River, NJ: Pearson Prentice Hall. Krieger, N. 1990. “Racial and Gender Discrimination: Risk Factor for High Blood Pressure?” Social Science and Medicine 30: 1273–­81. Krieger, N. 2000. “Refiguring ‘Race’: Epidemiology, Racialized Biology, and Biological Expressions of Race Relations.” International Journal of Health Services 30 (1): 211–­16. Krieger, N., and S. Sidney. 1996. “Racial Discrimination and Blood Pressure: The CARDIA Study of Young Black and White Adults.” American Journal of Public Health 86 (10): 1370–­78. Ladson-­Billings, G. J., and W. F. Tate. 1994. “Toward a Theory of Critical Race Theory in Education.” Teachers College Record 97: 47–­68. Landry, B. 2006. Race, Gender, and Class: Theory and Methods of Analysis. Upper Saddle River, NJ: Pearson Prentice Hall. Lauderdale, D. 2006. “Birth Outcomes for Arabic-­Named Women in California Before and After September 11.” Demography 43 (1): 185–­201. LaVeist, T. 1992. “The Political Empowerment and Health Status of African-­Americans: Mapping a New Territory.” American Journal of Sociology 97 (4): 1080–­95. ———. 1993. “Segregation, Poverty, and Empowerment: Health Consequences for African Americans.” Milbank Quarterly 71 (1): 41–­64. ———. 1994. “Beyond Dummy Variables and Sample Selection: What Health Services Researchers Ought to Know about Race as a Variable.” Health Services Research 29: 1. ———. 2005. “Disentangling ‘Race’ and Socioeconomic Status: A Key to Understanding Health Inequalities.” Journal of Urban Health: Bulletin of the New York Academy of Medicine 82 (2, suppl. 3): iii26–­iii34.



Contextualizing Lived Race-­G ender

209

LaVeist-­Ramos, T., J. Galarraga, R. Thorpe Jr., C. Bell, and C. Austin. 2011. “Are Black Hispanics Black or Hispanic? Exploring Disparities at the Intersection of Race and Ethnicity.” Journal of Epidemiological Community Health 10: 1136.1–­1136.5. Logan, J. 2003. “How ‘Race’ Counts for Hispanic Americans.” Working paper (July 14). Lewis Mumford Center for Comparative Urban and Regional Research University at Albany. López, N. 2003. Hopeful Girls, Troubled Boys: “Race” and Gender Disparity in Urban Education. New York: Routledge. ———. 2011. “Racially Stigmatized Masculinities: Conceptualizing Latino Male Schooling in the United States.” In Invisible No More: Understanding the Disenfranchisement of Latino Men and Boys, edited by Pedro Noguera, Edward Fergus, and Aida Hurtado, 235–­54. New York: Routledge. Lorber, J. 1994. Paradoxes of Gender. New Haven: Yale University Press. Marcus, G. 1998. Ethnography through Thick and Thin. Princeton, NJ: Princeton University Press. Marmot, M. 2005. “Social Determinants of Health Inequalities.” Lancet 365 (9464): 1099–­1104. ———. 2006. “Health in an Unequal World: The Haveian Oration.” Fellows of the Royal College of Physicians of London (Wednesday, October 18). Marmot, M., G. Davey Smith, and S. Stansfeld. 1991. “Health Inequalities among British Civil Servants: The Whitehall II Study.” Lancet 337: 1387–­93. Massey, D., and N. Denton. 1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge, MA: Harvard University Press. Mays, V., N. Ponce, D. Washington, and S. Cochran. 2003. “Classification of Race and Ethnicity: Implications for Public Health.” Annual Review of Public Health 24: 85–­110. Mills, C. W. 1956. The Power Elite. New York: Oxford University Press. Montoya, M. 2011. The Making of the Mexican Diabetic. Berkeley: University of California. Morning, A. 2009. “Toward a Sociology of Racial Conceptualization for the Twenty-­first Century.” Social Forces 87 (3): 1167–­92. Morning, A. 2011. The Nature of Race: How Scientists Think and Teach About Human Difference. Berkeley: University of California. Morris, E. 2007. “Researching Race: Identifying a Social Construction through Qualitative Methods and an Interactionist Perspective.” Symbolic Interaction 30 (3): 409–­25. Mullings, L., and A. Schultz. 2008. “Intersectionality and Health: An Introduction.” In Gender, “Race” and Class in Health: Intersectional Approaches, edited by A. Schultz and L. Mullings. San Francisco: Jossey-­Bass. Nakano Glenn, E. 2002. Unequal Freedom: How “Race” and Gender Shaped American Citizenship and Labor. Cambridge, MA: Harvard University Press. Olansky et al. 2012. “Differences in Life Expectancy Due to Race and Educational Differences Are Widening, and Many May Not Catch Up.” Health Affairs 31 (8): 1803–­13. Oliver, M., and T. Shapiro. 1995. Black Wealth/White Wealth: A New Perspective on Racial Inequality. New York: Routledge. Omi, M., and H. Winant. 1994. Racial Formation in the United States: From the 1960s to the 1990s. New York: Routledge. Penner, A., and A. Saperstein. 2008. “How Social Status Shapes Race.” Proceedings of the National Academy of Sciences 105 (50): 19628–­30.

210

Nancy López

Phelan J. C., B. G. Link, and P. Tehranifar. 2010. “Social Conditions as Fundamental Causes of Health Inequalities: Theory, Evidence, and Policy Implications.” Journal of Health and Social Behavior 51: S28–­S40. Prewitt, K. 2010. “The U.S. Decennial Census: Politics and Political Science.” Annual Review of Political Science 13: 237–­54. Richardson, L., J. Hussey, and K. Strutz. 2011. “A Life Course Perspective on Maternal and Child Health.” In Maternal and Child Health: Programs, Problems, and Policies in Public Health, edited by Jonathan B. Kotch, chap. 4. 3rd ed. Sudbury, MA: Jones and Bartlett. Roberts, D. 1997. Killing the Black Body: Race, Reproduction, and the Meaning of Liberty. New York: Pantheon. Rodriguez, C. 2000. Changing Race: Latinos, the Census, and the History of Ethnicity in the United States. New York: New York University Press. Rodriguez, C., G. Agueros, and M. Miyawaki. 2011. “Does Race and National Origins Influence the Hourly Wages that Latino Males Receive.” In Invisible No More: Understanding the Disenfranchisement of Latino Men and Boys, edited by Pedro Noguera, Edward Fergus, and Aida Hurtado, 207–­17. New York: Routledge. Roth, W. 2012. Race Migrations: Latinos and the Cultural Transformation of Race. Stanford, CA: Stanford University Press. Schultz, A., and L. Mullings. 2005. Gender, “Race,” and Class in Health: Intersectional Approaches. Indianapolis: Wiley. Smith, L. 2002. Decolonizing Methodologies: Research and Indigenous Peoples. New York: Palgrave St. Martin’s Press. Snipp, C. Matthew. 2003. “Racial Measurement in the American Census: Past Practices and Implication for the Future.” Annual Review Sociology 29: 583–­98. Takeuchi, D., and S.J.L. Gage. 2003. “What to Do with Race? Changing Notions of Race in the Social Sciences.” Culture, Medicine, and Psychiatry 27: 435–­45. Tallbear, K. 2008. “Native-­American-­DNA.coms: In Search of Native American Race and Tribe.” In Revisiting Race in a Genomic Age, edited by Barbara Koenig, Sandra Soo-­Jin Lee, and Sarah Richardson, 235–­52. New Brunswick, NJ: Rutgers University Press. Telles, E., and V. Ortiz. 2008. Generations of Exclusion: Mexican Americans, Assimilation, and Race. New York: Russell Sage. Trans-­ disciplinary “Race” Working Group. 2010. “Trans-­ disciplinary Guidelines for Researching ‘Race.’” Institute for the Study of “Race” and Social Justice, RWJF Center for Health Policy, University of New Mexico (August 2010). Accessed April 13, 2012. http://healthpolicy.unm.edu/node/486. U.S. Census Bureau. 2012. “State and County QuickFacts.” Data derived from Population Estimates, American Community Survey, Census of Population and Housing, State and County Housing Unit Estimates, County Business Patterns, Nonemployer Statistics, Economic Census, Survey of Business Owners, Building Permits, Consolidated Federal Funds Report. Accessed December 18, 2012. http://quickfacts.census.gov/qfd/ states/35000.html. Vidal-­Ortiz, S. 2004. “On Being a White Person of Color: Using Autoethnography to Understand Puerto Ricans’ Racialization.” Qualitative Sociology 27 (2): 179–­203. Wagley, C. 1968. “The Concept of Social Race in the Americas.” In The Latin American Tradition, 155–­78. New York: Columbia University Press. Walters, K., S. Mohammed, T. Evans-­ Campbell, R. Beltrán, D. Chae, and B. Duran. 2011. “Bodies Don’t Just Tell Stories, They Tell Histories: Embodiment of Historical



Contextualizing Lived Race-­G ender

211

Trauma among American Indians and Alaska Natives.” Du Bois Review: Social Science Research on Race 8 (1): 179–­89. Wang, Y. 2006. “How They Are Racialized? Racial Experiences of Chinese Students at UNM Graduate Programs.” MA thesis, University of New Mexico (March). Waters, M. 1990. Ethnic Options: Choosing Identities in America. Berkeley: University of California Press. Weber, L. 2007. “Future Directions of Feminist Research: New Directions in Social Policy—­The Case of Women’s Health.” In Handbook of Feminist Research: Theory and Praxis, edited by Sharlene Nagy Hesse-­Biber, 669–­79. Thousand Oaks, CA: Sage. West, C., and D. Zimmerman. 1987. “Doing Gender.” Gender and Society 1 (2): 125–­51. Williams, R., Jr. 1975. “Race and Ethnic Relations.” Annual Review of Sociology 1: 125–­64. Williams, D., and P. B. Jackson. 2005. “Social Sources of Racial Disparities in Health.” Health Affairs 24 (2): 325–­34. Williams, D., and M. Sternthal. 2010. “Understanding Racial-­Ethnic Disparities in Health: Sociological Contributions. Journal of Health and Social Behavior 51 (S15): S15–­S27. Williams, D., Y. Yu, J. Jackson, and N. Anderson. 1997. “Racial Differences in Physical and Mental Health: Socioeconomic Status, Stress, and Discrimination.” Journal of Health Psychology 2 (3): 335–­51. Williams D., S. Mohammed, J. Leavell, and C. Collins. 2010. “Race, Socioeconomic Status, and Health: Complexities, Ongoing Challenges, and Research Opportunities.” Annals of New York Academy of Sciences 1186: 69–­101. Williams, D., et al. 2012. “The Everyday Racism Scale.” http://scholar.harvard.edu/david rwilliams/pages/everyday-discrimination-scale-0. Accessed June 12, 2012. Zaal, Mayida, Tahini Salah, and Michelle Fine. 2007. “The Weight of the Hyphen: Freedom, Fusion, and Responsibility Embodied by Young Muslim-­American Women during a Time of Surveillance.” Applied Development Science 11 (3): 164–­77. Zuberi, T., and E. Bonilla-­Silva, eds. 2008. White Logic, White Methods: Racism and Methodology. Boulder, CO: Rowman and Littlefield.

Notes on Contributors

John A. Garcia  

is a research professor at the Institute for Social Research, University of Michigan, with appointments at the Inter-­University Consortium for Political and Social Research and the Center for Political Studies. His primary research areas are American politics, concentrating on minority group politics (especially Latinos), political mobilization and participation, urban governments, survey research, and public policy. He has published articles and book chapters in these fields for the past forty years. He is the co-­author of three books: Latino Lives in America: Making It Home (2009), Latinos in the New Millennium: An Almanac of Opinion, Behavior, and Policy Preferences (2012), and Latino Politics: Community Formation and Political Empowerment (2nd edition, 2012). is a professor in the School of Public Health and research professor and associate director of the Population Studies Center, Institute for Social Research, at the University of Michigan. She originated the “weathering hypothesis” as a framework for theorizing U.S. racial health inequality and has authored numerous articles published in social science, public health, and medical journals, including “To Denigrate, Ignore, or Disrupt: Racial Inequality in Health and the Impact of Policy- Induced Breakdown of African American Communities” and “The Mutability of Women’s Health with Age: The Sometimes Rapid, and Often Enduring, Health Consequences of Injustice.” She teaches courses on the structural influences on population health, on health and poverty, and on women’s health and the timing of reproduction.

Arline T. Geronimus 

Laura E. Gómez  is a professor of law at the University of California, Los Angeles (and professor, by courtesy, in UCLA’s departments of sociology and Chicana/o studies). She is the author of Manifest Destinies: The Making of the Mexican American Race (2007) and Misconceiving Mothers: Legislators, Prosecutors, and the Politics of Prenatal Drug Exposure (1997). She teaches courses on critical race theory, law and society, constitutional law, criminal law, and civil procedure. She is the former president of the Law and Society Association (2009–­2011). Joseph L. Graves Jr.   is the associate dean for research and a professor of biological sciences at the Joint School of Nanosciences and Nanoengineering, North Carolina A&T State University and UNC-­Greensboro. He has served as chair of the Senior Advisory Board for the National Evolutionary Synthesis Center. In addition, he is an associate editor for the second edition of the Encyclopedia of Race and Racism, and his books include The Emperor’s New Clothes: Biological Theories of Race at the Millennium (2005) and The Race Myth: Why We Pretend Race Exists in America (2005). His research interests include the evolutionary

213

214

Notes on Contributors

genomics of postponed aging, computational evolutionary phylogenetics and molecular evolution, and biological concepts of race in humans. Janet E. Helms  is Augustus Long Professor of Measurement in Counseling Psychology at Boston College and founding director of the Institute for the Study and Promotion of Race and Culture. She is the author of A Race Is a Nice Thing to Have: A Guide to Being a White Person or Understanding the White Persons in Your Life (2008), Using Race and Culture in Counseling and Psychotherapy (with Donelda Cook; 1999) and many articles and book chapters. She teaches courses in testing, research, and history with a focus on social justice and racial factors. Derek Kenji Iwamoto  is

a research assistant professor at the Center for Addictions, Personality, and Emotion Research, University of Maryland–­College Park. He is the editor of Culturally Responsive Counseling Interventions with Asian American Men (2010) and the author of numerous articles in addictions, gender socialization, and multicultural psychology. He teaches courses on assessment, gender socialization, and multicultural psychology. Jonathan Kahn  is

a professor of law at Hamline University School of Law. He is the author of Race in a Bottle: The Story of BiDil and Racialized Medicine in a Post-­Genomic Age (2012). He teaches courses on constitutional law, torts, biotechnology and the law, bioethics, and public health law. Jay S. Kaufman  is

a professor and Canada Research Chair in Health Disparities at the Department of Epidemiology, Biostatistics, and Occupational Health at McGill University. He is author of over two hundred articles and book chapters, and is coeditor of the textbook Methods in Social Epidemiology (2006). He teaches courses in epidemiologic methods, statistical analysis, and social epidemiology.

Mai M. Kindaichi  is a staff psychologist at George Washington University’s Counseling Center. She has numerous publications in the areas of multicultural and identity development concerns (for example, race, sexual identity, social class, ethnicity), family conflict, and clinical health concerns. She teaches courses on psychopharmacology and multicultural issues. Simon J. Craddock Lee is a medical anthropologist and an assistant professor of clinical sciences at the University of Texas Southwestern Medical Center. He leads studies in cancer disparities funded by the American Cancer Society, the Cancer Prevention and Research Institute of Texas, and the National Cancer Institute. Former chair of the Committee on Minority Issues in Anthropology for the American Anthropological Association, he teaches courses in medical and research ethics, public health, health services, and organizational culture. Nancy López  

is an associate professor of sociology at the University of New Mexico. She is the author of Hopeful Girls, Troubled Boys: Race and Gender Disparity in Urban Education (2003) and many articles and book chapters. She teaches courses on “race,” the sociology of education, and the intersection of race, class, and gender.



Notes on Contributors 215

She is the co-­founder and director of the Institute for the Study of “Race” and Social Justice at the RWJF Center for Health Policy at UNM. Ethan H. Mereish  is a diversity fellow and doctoral candidate in counseling psychology at Boston College. He has coauthored articles regarding the effects of discrimination on mental and physical health for racial/ethnic and sexual minorities and the socializing factors of prejudice along with the processes that mitigate them. He teaches courses in psychology and counseling and leadership. Matthew Miller  is

an assistant professor at University of Maryland–­College Park. The majority of his publications have examined the role of cultural and racial factors in health and health disparities among Asian Americans and/or psychological measurement. He teaches courses on psychological assessment and multicultural counseling. Gabriel R. Sanchez  is

an associate professor of political science at the University of New Mexico. He is the author of Hispanics and the U.S. Political System (2008) as well as several journal articles and book chapters. He teaches courses on racial and ethnic politics, political behavior, and public opinion.

Aliya Saperstein  is

an assistant professor of sociology at Stanford University. Her research on the measurement of race and ethnicity and its implications for understanding social inequality has been published in the Proceedings of the National Academy of Sciences, PLoS One, Social Forces, Social Problems, and Ethnic and Racial Studies. She teaches courses on comparative racial formation, the social determinants of health, social demography, and research methods.

R. Burciaga Valdez  is executive director of the Robert Wood Johnson Foundation Center for Health Policy and the RWJF Professor in Family and Community Medicine and Economics at the University of New Mexico. He was the founding dean of the Drexel School of Public Health in Philadelphia and previously a professor of health services at the UCLA School of Public Health. Formerly he served in various capacities in Washington, DC, including deputy assistant secretary for Health in the Public Health Service, director of Interagency Health Policy in the U.S. Department of Health and Human Services, and White House senior advisor. His research and policy work include a focus in health services. Vickie D. Ybarra  is a doctoral fellow in the political science department at the University of New Mexico, where she holds a Robert Wood Johnson Foundation Pre-­doctoral Fellowship at the Center for Health Policy. Her research interests include health policy, the social determinants of health, immigration policy, and Latino politics. Prior to returning to school to pursue her doctoral degree, she enjoyed a twenty-­one-­year career in community health. She has been a leader in Washington State in health policy, education policy, and addressing racial/ethnic health disparities.

Index

Note: Page numbers in bold indicate a figure; italics indicate a table access: to health care, 4, 6, 27, 95, 140; and personal preferences, 35; to rights, 32 acculturation, 79–­80, 105, 108–­13, 126 Adarand v. Pena, 34 Add Health, 136, 138–­39 affirmative action, 32–­33 African Americans: and biological mortality, 41; and childhood depression, 49; and diabetes, 1–­2; and infant mortality, 39; and physiological stress activation, 165. See also Black age, 42–­43, 151–­52 Agency for Health Research and Quality (AHRQ), ix, 181 age-­specific mortality, 39, 40 AHANA, 159n1. See also African Americans; Asian Americans; Hispanic identity; Native Americans AHRQ. See Agency for Health Research and Quality AIMs. See ancestry informative markers ALANAs, 146, 148, 156, 159nn1–­2. See also African Americans; Asian Americans; Hispanic identity; Native Americans; people of color allostatic load, 165 Alternative Questionnaire Experiment Report (2010), 196 ALU insertion, 40 Alzheimer’s disease, 43 American Community Survey, 199 ancestry, 75, 199 ancestry informative markers (AIMs), 58–­ 59, 136, 189 angiotensin-­converting enzyme (ACE), 40 antagonistic pleiotropy, 43–­44 anthropology, 14 Arabic-­named women, 171, 189 ascertainment bias, 23–­24, 39–­52 ascribed racial status, 192–­94

Asian Americans: and generation, 120–­21; health status of, 114n7; mental health of, 117–­30; and racial identity development, 118–­20, 124–­27; and stereotype threat, 169 assessment: literature, 157–­58; of psychological well-­being, 121 assignment, racial, 9 attitudes, 89, 96 attitudinal variables, 126 autonomy, 156 baseline prevalence, 56 Bayesians, 63 behavior: of health care providers, 35; risky, 36 Behavioral Risk Factors Surveillance System (BRFSS), 97, 106–­7, 113n3, 138–­39, 192–­93 best practices, 17, 181 between-­group differences, 147–­48, 152 BiDil, 25, 27 bioethnic conscription, 190 biological difference, 28, 31 biology, 14–­15, 26, 42, 68 biomedical: model, 183; profession, 49; research, 15–­17, 24, 39–­52, 54–­58 biracial identity, 74–­76 birth: outcomes, 171, 189; parental place of, 199 Black: Hispanics, 180; mortality rates, 201; women, 4. See also African Americans; people of color blood pressure, 17–­18 bodies, racialized, 26, 36 breast cancer, 4–­6, 42 BRFSS. See Behavioral Risk Factors Surveillance System Brown, Henry, 30 Brown v. Board of Education, 31, 33–­34

217

218 Index 

California Endowment, ix cancer, 4–­6, 42, 87–­103, 99, 137 categories: creation of, 68; initial draft policy, 92, 92–­94; and regression models, 54–­55. See also racial categories categorization, 54–­57 “Caucasian,” 15–­16 causal inference, 58 causation, 36 causes: of health inequalities, 2, 25, 105; social justice view of, 19n3 Census, U.S.: and Hispanics, 196–97; and Native Americans, 195–­96, 205n13; and race-­specific data, 31; racial categories, 5, 70–­71; “some other race” category, 197; and statistics, 54 Center for Epidemiologic Studies Depression (CES-­D) scale, 147, 152–­53 Centers for Disease Control, 3, 106–­7, 139 CES-­D scale. See Center for Epidemiologic Studies Depression scale CFA. See confirmatory factor analysis Chain Reaction (Edsall and Edsall), 31 Chicago, 4–­6 Child Depression Inventory, 49 chronic health conditions, 127 City of Richmond v. J. A. Croson Company, 33–­34 Civil Rights Act (1964), 31 Clark, Kenneth, 31 clines, 45 clinical: care, 87–­103; context, 17; trial management system, 90 Collaborative Multi-­Racial Post Election Survey (2008), 86, 106–­8, 113nn2–­3 colorblind: ideology, 10, 14, 16, 190; racial attitudes, 119 colorism, 76 Commonwealth Fund, 87 confirmatory factor analysis (CFA), 121–­22, 122, 125, 153 conformity, 118, 122, 125–­26, 156, 157 constructionist view of race, 11–­15, 134, 135, 141–­43 constructs, race-­related, 148 continental origins, 58–­59 contingencies of social identity, 167–­70 control variable, 8–­9, 69, 142

coping strategies, 125, 165 Coronary Artery Risk Development in Young Adults, 143n1 critical race theory, 184 cultural: dynamics, 15; race, 192–­94; racism, 73, 154; socialization, 151 culture, 78 Current Population Survey, 71, 137, 199 Darwinian medicine, 41, 50 data: and clinical care, 87; and health disparities research, 181–­82; individual-­ level, 202; mortality, 19; multilevel race, 200; race-­specific, 31; reliability of, 85; validating hospital, 97–­98 data collection: and ascribed race, 193–­94; barriers to, 87–­88; best practices for, 98; and cancer patients, 88–­93, 99; context-­ specific, 181; federal, 19n1; innovations in, 138; and lived-­raced gender, 194–­ 95; and medical staff concerns, 96; and Native Americans, 195–­96; and race, 134; and social determinants of health, 181–­82; sociocultural obstacles to, 94; standardized, 82–­97; variations in, 88 death, 41–­42 decomposition of effects, 58 demographics: and depression, 150; and health status, 111; in New Mexico, 180–­ 81; of U.S. Latino population, 105 Department of Health and Human Services, 3, 26 dependent variable, 10 depression, 131–­32, 146–­62, 150 desegregation, 32–­33 determinants of health status, 111 diabetes, 1–­2, 15 diagnosis, 56; diagnostic inconsistencies, 148 dichotomous variable, 9 differential: item functioning, 153; treatment, 72 discordance, 44–­45 discrimination, 70; and health disparities, 25, 107–­10; and health status, 112, 126; and immigrants, 78; institutional, 142; perceived, 72, 127, 136, 137; privatization of, 34; state-­sanctioned, 31 disease, 40–­45



Index 219

disintegration, 156 disparities. See health disparities dissimilation, 74 dissonance, 118, 119, 122, 125, 156 distal risk factors, 150–­51 diversity within minority groups, 112 domination theory, 185–­86 Douglas, J. Allen, 30 Du Bois, W.E.B., 29–­31 Du Bois Review, 104 dummy variables, 10 dynamism, 68–­69

factor analysis, 120–­21, 129n1, 152–­53 federal funding, 94–­95 focus groups, 89 food excess, 46 Fourteenth Amendment, 30 framing, 23, 25–­38 frequentists, 63 Frist, Bill, 36–­37

Edsall, Mary, 31, 32 Edsall, Thomas, 31 electronic medical record (EMR), 89–­92, 94, 99, 100 embodiment: and health disparities, 132; life course, 202; pathways of, 176, 185–­ 86, 199–­202 emersion/immersion, 118–­19, 122, 125–­26, 156 empirical settings, 85–­86 employment policies, 32–­33 EMR. See electronic medical record environment: and depression, 150; differences in, 48; novel, 45–­46; racist, 73 Epic EMR, 90–­92 epidemiological: research, 126; standardization, 54 epigenetics, 164, 175n1 Epstein, Richard, 35 equal opportunity, 32 Equal Protection Clause, 30 essentialist: characteristics, 163; views, 19n6 ethics, 24, 53–­66, 85 ethnicity: and Asian Americans, 117; and cancer research, 89–­92; and data collection, 197; and health disparities research, 180; Hispanic, 96; measurement of, 104; as multidimensional, 191, 198, 198–­99; and race, 71, 77–­78, 179–­80, 196–­97 ethnography, 195 eugenic theories, 28–­29 evolutionary: biology, 42; legacies, 47; medicine, 41, 50 externalities, 69

Garcia, John A., 24, 67–­83, 213 gender, 151–­52, 169, 184–­85. See also lived race-­gender gendered: racism, 194; social order, 185 gene-­environment interactions, 164 General Social Survey (GSS), 140, 142 generation, 120–­21, 128, 198, 199 genetic: ancestry, 205n12; drift, 41, 43; markers, 154; predisposition, 164; reductionism, x, 23–­24; testing, 58–­59; variation, 23, 39–­40 genetics: and health outcomes, 1; molecular, 47–­48; and race, 14–­15; and surveys, 136 geography, 23, 39–­40, 154 Geronimus, Arline T., 132, 163–­78, 182, 195, 213 Glazer, Nathan, 61 Gómez, Laura E., xii–­xv, 1–­22, 193, 213 Graves, Joseph L. Jr., 15–­17, 23, 39–­52, 213–­14 Gravlee, Clarence, 9, 17–­18 Griggs v. Duke Power, 32–­33 group-­based racial remedies, 34 group identity, 105 Grutter v. Bollinger, 34 GSS. See General Social Survey Harlan, John, 30, 31 Hartigan, John Jr., 11 health: and contingencies of social identity, 169; and discrimination, 126; inequalities, 2, 8, 73; of Mexican immigrants, 170–­71; psychosocial factors and, 46–­47; social determinants of, 7, 36–­37, 179–­211 health care: access to, 4; as consumer good, 27, 28, 37; providers, 35, 80, 97, 203; and residential segregation, 5–­6 Healthcare System Distrust Scale, 98

220 Index 

health disparities: and acculturation, 79–­80; and biological difference, 28; and cancer, 88–­92; in cancer survival, 4–­6; cause of, 2, 25, 105; and contextualizing lived race-­gender, 202–­3; and discrimination, 25, 107–­10; fatalism about, 58; framing of, 25–­38; genetic reductionist interpretations of, x; and immigrants, 47, 199; and infectious disease, 41; literature on, 3, 7; in Native American communities, 195–­96; and racial socialization experiences, 158; research, 1–­22, 78–­80, 180; response to, 57; and skin color, 104–­5; and socioeconomic status, 190–­91; source of, 23, 36 health outcomes: genetic basis for, 1; and group identity, 105; in-­group differences in, 112–­13; and racial identity, 127, 157; and skin color, 110, 112 health status: determinants of, 111; and political participation, 86, 106–­8; and race, 108–­12; and socially defined race, 193 Helms, Janet E., 86, 118, 131–­32, 146–­62, 214 help-­seeking behaviors, 128 Hispanic identity, 93–­96; and U.S. Census, 196–­97. See also Latinos Hoffman, Frederick, 28–­29 hospitals, 4–­5, 87–­103, 181 human: stress genomics, 164, 175n1; variability, 54 Human Genome Project, 184, 190 hypertension, 17–­18 identity: multiracial, 74–­76; safety, 173–­74; as situational, 95. See also racial identity ideology, racial, 13 illness prevalence rates, 126 immersion/emersion, 118–­19, 122, 125–­26, 156 immigrants: Asian Americans as, 120–­21; and health disparities, 47, 111, 199; health status of, 112, 170–­71; racialization of, 78 independent variables, 147, 149–­51 indicators: and depression research, 150, 150–­51; of holistic race, 69–­70, 70 individual: preferences, 14; racism, 73, 154; rights, 33

inequalities: as cause, 19n3; health, 2, 8, 73 infant mortality, 39 infections, 41–­42 inference, 59 inferiority, 29 in-­group analysis, 112–­13 insertion allele, 40 Institute for the Study of “Race” and Social Justice, x, xii Institute of Medicine (IOM), 25–­26, 87, 100 institutional: barriers, 150, 150–­51; racism, 72, 73, 119, 142, 154 Institutional Review Board (IRB), 89–­90 insurance, 28 integrative awareness, 156 interaction effects, 142 interdisciplinary research, ix intergenerational drag, 73 internalization, 118, 122, 125, 156 internalized: cultural learning, 151; racism, 155, 156, 157 internal variation, 105–­9 interpersonal anxiety, 92 intersectionality, 141–­42, 181, 185, 200, 203 Inter-­University Consortium for Political and Social Research, 67 interviewer-­classified race, 138–­41, 143n2 intra-­minority group homogeneity, 86 IOM. See Institute of Medicine isolation by distance, 39–­40 Iwamoto, Derek Kenji, 86, 117–­30, 214 James-­Stein estimator, 60 Jewish Americans, 171 Jones, Camara, 18 Kahn, Jonathan, 23, 25–­38, 214 Kaufman, Jay S., 14, 24, 53–­66, 214 Kelly’s Paradox, 60 Kindaichi, Mai M., 86, 117–­30, 214 King, Martin Luther Jr., 32 Korean Americans, 127 language, 80, 108, 113n2, 128, 198 Latinas, 108 Latinos: and data collection, 93–­96; health status of, 111; in-­group analysis of, 112; internal variation within, 105–­8;



Index 221

measuring differences among, 86; and race, 71, 179–­80; self-­defined health status of, 108–­12; and “some other race” category, 196–­97 LaVeist, Thomas, 1–­2, 7 Lee, Catherine, 19n4 Lee, Simon J. Craddock, 85, 87–­103, 214 Lee, Taeku, 9–­10 Levine, Alan, 27 life course embodiment, 202 life expectancy, 3 lived experience, 132 lived race-­gender, 182–­83, 187–­91, 194–­95, 204n9 López, Nancy, xiii–­xv, 12–­13, 179–­211, 214–­15 macro-­level analysis, 12, 13, 185, 200 malaria, 43–­44 marginalization, 164 market-­based approach, 26–­27, 34–­37 Massachusetts Hospital Association, 96 measurement: error, 170; invariance, 152–­53; multidimentional, 79; of racial identity, 68; of racial self-­identification, 12; of racial subordination, 73; of social disadvantage, 100; validity of, 72 Measures of Racism Working Group, 139 Medical College Admissions Test (MCAT), 183 medicine: evolutionary, 41, 50; racialized, 25 men of color, 13 mental health, 117–­30, 139, 146–­62 Mereish, Ethan H., 131–­32, 146–­62, 215 meso-­level analysis, 12, 200, 202 methodology, 134–­38 Mexican Americans, 13–­14, 109–­12, 111, 193 Mexican immigrants, 170–­71 microaggressions, 13, 18, 132, 188, 194 micro-­level analysis, 11–­12, 192, 200 middle-­class privilege, 188 Miller, Matthew, 86, 117–­30, 215 minority groups, 112 Minority Identity Development Model, 121 misclassification, 19n1, 55–­56 mixed ancestry, 75 model specification, 58 molecular genetics, 47–­48

“Mongoloid,” 15–­16 Montoya, Michael, 2, 15 mood disorders, 146 Morning, Ann, 10, 19n6 Morris, Edward, 9 mortality: age-­specific, 39, 40; Black, 201; causes of, 41–­42; data, 19n1; and self-­ reported health status, 114n5 Moynihan Report on “The Negro Family” (1965), 36 Multi-­City Study of Urban Inequality, 143n1 multidimensional: indicators, 68, 79; race data, 192 multiracial: identity, 74–­76; status, 70, 140 multivariate statistical models, 10 mutation, 40, 43–­44 Myrdal, Gunnar, 31 National Center on Minority Health and Health Disparities (NCMHHD), xii National Health Interview Survey, 180 National Institutes of Health, xiii, 49–­50 National Latino and Asian American Survey (NLAAS), 127–­28 National Longitudinal Mortality Study, 137 National Longitudinal Study of Adolescent Health, 136, 138–­39 National Longitudinal Survey of Youth (NLSY), 140–­41 National Survey of Family Growth (NSFG), 140, 142, 143n2 Native Americans, 195–­96, 205n13 nativity, 108, 128 natural selection, 41–­43 “Negroid,” 15–­16 neighborhoods, 5 neoconservative race theorists, 34 NEO Five-­Factor Inventory, 157 neoliberal frame, 23 New Immigrant Survey, 136 New Mexico, 13–­14, 204n3, 204n10 New Mexico Hospitals Association Race and Ethnicity Advisory Committee, 181–­82 NLAAS. See National Latino and Asian American Survey NLSY. See National Longitudinal Survey of Youth

222 Index 

non-­random sampling, 23–­24 novel environment, 45–­46 NSFG. See National Survey of Family Growth O’Connor, Sandra Day, 33 Office of Management and Budget (OMB) racial categories, 154, 204n5; and cancer research, 90, 97; and Hispanic identity, 197; and Latinos, 71, 96 Office of Research on Minority Health, 3 off-­White: racial status, 193; wedge group, 13–­14 OMB Directive 15. See Office of Management and Budget racial categories oncology services, 97 one-­stage protocol, 72 outcomes, 32, 95. See also health outcomes oxidative stress, 175–­76n2 parental place of birth, 199 Parents Involved in Community Schools v. Seattle School District No. 1, 33–­34 participant observation, 195 patient: care, 98; self-­report, 90, 95, 96 Patient Protection and Affordability Act (2010), 5 pattern of aging, 42–­43 Penner, Andrew, 12 people of color, 156, 159n2, 201–­2 People of Color Racial Identity Attitudes Scale (PRIAS), 118, 120–­27, 122, 123, 155–­57; 12-­question version of, 122, 122–­27, 123; 15-­question version of, 122, 122–­27, 123; 32-­question version of, 121 perceived: discrimination, 72, 136–­37; race, 12, 140; racism, 119; stress, 165–­66 peripheral blood mononuclear cells (PBMC), 164 personal responsibility, 25 pharmacogenomic practices, 35 phenotype, 8, 14, 69, 76–­77 physical features, 69, 76–­77, 136, 154 physicians, 91, 98 physiological: differences, 187; reactions, 73, 170; stress activation, 165, 169–­71, 174 physiology, stress, 164–­67

Plessy v. Ferguson, 29–­30, 34 political: context, 105–­6; factors, 106–­8, 110–­13, 111; participation, 86, 108, 110; science, 104–­16 politicization, 55 population health, 164 positionality, 182, 195, 204n6 poverty, 6, 165–­66 prediction, 60–­62, 62 predictor variable, 147, 149 pregnancy, 188–­89 prevalence, 56 prevention, 36 PRIAS. See People of Color Racial Identity Attitudes Scale primary racial identity, 75 profiling, 188 progeria, 44 protocol, survey, 72 proximal risk factors, 151 proxy: risk factors, 150, 150–­51; variables, 158 pseudo-­independence, 156 psychocultural risk factors, 151 psychological: characteristics, 149; inclusion criteria, 153; research, 152; well-­ being, 86, 119, 123, 123, 125, 126 psychometrics, 53 psychoracial risk factors, 151 psychosocial factors, 46–­47 public health, 7, 163–­78 qualitative research, 89 quantitative research, 9, 53, 133–­45 race: and acculturation, 105; analysis of, 11–­12; and biology, 5, 14–­15, 26, 29, 68; as category, 147; category creation, 68; in clinical context, 17, 87; conceptualization of, 3–­4, 10, 182; constructionist view of, 11–­15; constructivist theory of, 134, 135, 141–­43; as control variable, 7, 69; and culture, 78; definition of, 7, 134; dimensions of, 24, 135, 191–­96; dynamism of, 68–­69; essentialist view of, 19n6; and ethnicity, 71, 77–­78, 179–­80, 196–­97; and group-­based remedies, 34; and health, 167; holistic concept of, 67, 69–­70, 70;



Index 223

and Human Genome Project, 184; and hypertension, 17–­18; as individual characteristic, 7, 16; interviewer-­classified, 138–­41, 143n2; and Latinos, 71, 179–­80; as lived experience, 132, 163; measurement of, 8, 86, 104, 135, 139–­40; and medicine, 47–­50; multidimensional, 68, 109, 134, 191, 192, 198; as multilevel, 200; outside the U.S., 71; perceived, 12, 141; and perceived discrimination, 137; and phenotype, 76–­77; and physician notes, 91; proxies for, 150; in quotation marks, 203n1; and research, xii–­xiii, 134–­ 38; self-­identified, 5; simplistic measures of, 9; as situational, 71, 95, 167; as social construction, x, 1–­22, 69, 183, 199; socially defined, 192–­94; social meaning of, 6, 7; “some other” category, 196–­99; and statistical analysis, 153; and survey research, 67–­83; theory, 30–­31, 34; thinking, 18; variability of, 136–­37 race-­based pharmacogenomic practices, 35 race-­conscious law and policy, 10 race-­gender profiling, 188 race-­of-­interviewer effect, 77 race-­related: behavioral criteria, 153; variables, 147 racial: ambiguity, 19n1; assignment, 9; classification, 135–­36; difference, 31, 137; discrimination, 33, 34; formation theory, 184; grouping, 58–­59, 149; groups, 105, 147; hierarchy, 13–­14; identification, 70, 74, 75; identity data, 87–­103; ideology, 13; inequality, 28, 134–­35; perceptions, 12, 141; profiling, 16; remedies, 34; segregation, 5; socialization, 131, 151, 155; status, 14; stereotyping, 16, 149; subordination, 73; theoretical constructs, 158–­59 racial categories, 2; and biomedical research, 17; and Census Bureau, 70–­71; dynamic nature of, 13; researcher-­ imposed, 148–­49; social context of, 7 racial identity: and Asian Americans, 117–­30; consciousness of, 167; and depression, 146–­62; development, 86, 117–­20, 124–­28, 147; expressed, 135–­36; four-­factor model of, 120; measure of,

68; playing down one’s, 107–­8; primary, 75; psychological attributes of, 155; and psychological well-­being, 86; and racism, 72; schemas, 155, 156; status, 118, 121, 127; theory, 131–­32, 146–­62; and visual assessment, 94; within-­group studies of, 158 racialization: of immigrants, 78; process, 2, 68 racialized bodies, 36 racialized-­gendered pathways of embodiment, 186 racialized-­gendered social determinants of health, 182–­87, 200, 200–­201, 204n8 racial self-­identification, 8, 16, 107, 118, 191–­92; and Add Health, 138; and externalities, 69; and interviewer classification, 142; measure of, 12; situational, 74; variables in, 10 racism: environmental, 73; forms of, 73; gendered, 194; and health outcomes, 18; and health status, 153–­54; institutional, 72, 119; internalized, 72, 155; scientific, 11; social meaning of, 7; structural, 18, 70; systemic, 199; types of, 154 racist environments, 73 radical contextualization, 186, 202 Reactions to Race survey module, 97, 192–­93 Reagan, Ronald, 32–­33 reductionism, 48 reflected appraisal, 131, 136, 139 reflectometers, 107 regression: analysis, 123, 158; to the mean, 60–­61; methods, 8, 54–­55 reliability: of data, 85; of self-­reported health status, 107 research: biomedical, 15–­17, 24, 39–­52, 58; and cancer disparities, 88–­92; and categorization, 54; depression, 146–­62; design, 134, 141–­43; epidemiological, 126; health disparities, 1–­22, 78–­80, 180; inclusivity policies of, 88; interdisciplinary, ix; mental health, 147; and misclassification, 55–­56; and policy changes, 3; and private industry, 50; public health, 163–­78; and race, xii–­xiii, 134–­38; subjects, 93; translational, 172–­73

224 Index 

researchers, 49–­50, 195 residential segregation, 5–­6, 201 residual: method, 142; race effects, 170 restrictive inclusion criteria, 121 risk factors: and depression, 150, 150–­51; for stress-­related diseases, 166; types of, 151, 158 Roberts, Dorothy, 4, 35 Roberts, John G. Jr., 33–­34 Robert Wood Johnson Foundation (RWJF), ix; Center for Health Policy, x, xiv Rose, Michael, 17 sampling, 23–­24 Sanchez, Gabriel R., 86, 104–­16, 215 Sankar, Pamela, 7 Saperstein, Aliya, 12, 131, 133–­45, 197, 215 Satel, Sally, 16, 35, 57 Scales of Psychological Well-­Being (SPWB), 121 Scalia, Antonin, 34 schools, 12–­13, 32–­34, 173 Science of Eliminating Health Disparities Summit (2008), ix scientific racism, 11 SEER cancer registry, 137 segregation, 4–­6, 12–­13, 30–­34, 201 self-­affirming narratives, 174 self-­identification: and external perceptions, 137; as multiracial, 75; and phenotype, 77; and racial categories, 70; and reflected appraisal, 139; of research subjects, 93; single measure of, 135; static, 18. See also racial self-­identification self-­reported health status, 107–­13, 114n5 Separate Accommodations Act (1890), 29–­30 SES. See socioeconomic status sexual orientation, 128 sickle cell anemia, 43–­44 situational: cues, 172; racial status as, 14, 74; salience, 166–­67 skin color, 44–­45, 70; ascribed, 107; and colorism, 76; data on, 143n1; and health disparities, 104, 111; and health status, 109–­12; interviewer-­classified, 138; and racial categories, 76–­77, 136

SOAP (Subjective-­Objective Assessment Plan) note, 91, 100n3 social: context, 7, 137; determinants of health, 7, 36–­37, 179–­211; disadvantage, 98, 100; discrimination, 25; dynamics, 26; groupings, 55; hierarchy, 49–­50; identity, 74, 163–­78; inequality, 46–­47; patterning, 164; responsibility, 35; sciences, 13, 67; segregation, 73; status, 137; stratification, ix social construction: of biology, 15; contingent, 182; of race, x, 1–­22 socialization, 151 socially: assigned multiracial status, 75; defined race, 192–­94 social-­psychological processes, 166 sociocultural risk factors, 151 socioeconomic status (SES), 204n2; and cancer patient research, 90; and health disparities, 79, 190–­91; and health research, 183; and health status, 110, 111 sociological vantage point, 11–­12 sociology, 14 “some other race” U.S. Census category, 196–­99 Southwest, U.S., 13–­14 standardization, 54 statistical: analysis, 62, 153, 158; inference, 59; practice, 53–­66 status gains, 139 Steele, Claude, 167–­70 stereotypes, 16, 149 stereotype threat, 167–­69, 172–­73 stigma, 30, 58, 108, 132 stomach cancer, 42 stress: activation process, 172, 174; and Asian Americans, 119; hormones, 165; and Latinas, 108; and microaggressions, 13; physiology, 164–­67; and racism, 73; and stereotype threat, 169–­70 structural racism, 70, 73, 199 Supreme Court opinions, U.S., 28–­35 survey: data sets, 131, 138–­41; research, 67–­83, 133–­45 survival disparity, 4–­6 Sweet, Elizabeth, 9 symbolic interactionism, 95 systemic racism, 199



Index 225

telomere length, 165–­66, 175–­76n2 testing, 61–­62, 62 translational research, 172–­73 transparency, 195, 204n6 treatment, 5, 16 tribal: affiliation data, 181; status, 195–­96 Unequal Treatment: Confronting Racial and Ethnic Disparities in Healthcare (IOM), 26, 35 University of New Mexico, ix–­x, xiv University of Texas Southwestern Medical Center, 91 U.S. Census. See Census, U.S. Utah Population Database, 137 Valdez, R. Burciaga, ix–­xi, 215 validity: of data, 97–­98; of racial measurement, 72 variability: human, 54; of race, 136–­37

variable: categorized, 54–­55; control, 8–­9, 69, 142; dependent, 10; independent, 147; life-­history, 42 Velos Research Patient Database, 90–­91, 99 visual assessment of race, 93–­95 voting, 108 Voting Rights Act (1965), 31 weathering, 164 White: Hispanics, 180; Mexican Americans as, 13–­14; privilege, 184; racial identity, 156, 157; women, 4 White Racial Identity Attitude Scale (WRIAS), 157 Williams, David, 3, 6 within-­group: analyses, 152, 158; comparisons, 148; differences, 117, 128 women of color, 194 Ybarra, Vickie D., 86, 104–­16, 215

Available titles in the Critical Issues in Health and Medicine series: Emily K. Abel, Suffering in the Land of Sunshine: A Los Angeles Illness Narrative Emily K. Abel, Tuberculosis and the Politics of Exclusion: A History of Public Health and Migration to Los Angeles Marilyn Aguirre-­Molina, Luisa N. Borrell, and William Vega, eds. Health Issues in Latino Males: A Social and Structural Approach Susan M. Chambré, Fighting for Our Lives: New York’s AIDS Community and the Politics of Disease James Colgrove, Gerald Markowitz, and David Rosner, eds., The Contested Boundaries of American Public Health Cynthia A. Connolly, Saving Sickly Children: The Tuberculosis Preventorium in American Life, 1909–­1970 Tasha N. Dubriwny, The Vulnerable Empowered Woman: Feminism, Postfeminism, and Women’s Health Edward J. Eckenfels, Doctors Serving People: Restoring Humanism to Medicine through Student Community Service Julie Fairman, Making Room in the Clinic: Nurse Practitioners and the Evolution of Modern Health Care Jill A. Fisher, Medical Research for Hire: The Political Economy of Pharmaceutical Clinical Trials Alyshia Gálvez, Patient Citizens, Immigrant Mothers: Mexican Women, Public Prenatal Care and the Birth Weight Paradox Gerald N. Grob and Howard H. Goldman, The Dilemma of Federal Mental Health Policy: Radical Reform or Incremental Change? Gerald N. Grob and Allan V. Horwitz, Diagnosis, Therapy, and Evidence: Conundrums in Modern American Medicine Rachel Grob, Testing Baby: The Transformation of Newborn Screening, Parenting, and Policymaking Mark A. Hall and Sara Rosenbaum, eds., The Health Care “Safety Net” in a Post-­Reform World Laura D. Hirshbein, American Melancholy: Constructions of Depression in the Twentieth Century Timothy Hoff, Practice under Pressure: Primary Care Physicians and Their Medicine in the Twenty-­first Century Beatrix Hoffman, Nancy Tomes, Rachel N. Grob, and Mark Schlesinger, eds., Patients as Policy Actors Ruth Horowitz, Deciding the Public Interest: Medical Licensing and Discipline Rebecca M. Kluchin, Fit to Be Tied: Sterilization and Reproductive Rights in America, 1950–­1980 Jennifer Lisa Koslow, Cultivating Health: Los Angeles Women and Public Health Reform Bonnie Lefkowitz, Community Health Centers: A Movement and the People Who Made It Happen Ellen Leopold, Under the Radar: Cancer and the Cold War Barbara L. Ley, From Pink to Green: Disease Prevention and the Environmental Breast Cancer Movement Sonja Mackenzie, Structural Intimacies: Sexual Stories in the Black AIDS Epidemic David Mechanic, The Truth about Health Care: Why Reform Is Not Working in America

Alyssa Picard, Making the American Mouth: Dentists and Public Health in the Twentieth Century Heather Munro Prescott, The Morning After: A History of Emergency Contraception in the United States David G. Schuster, Neurasthenic Nation: America’s Search for Health, Happiness, and Comfort, 1869–­1920 Karen Seccombe and Kim A. Hoffman, Just Don’t Get Sick: Access to Health Care in the Aftermath of Welfare Reform Leo B. Slater, War and Disease: Biomedical Research on Malaria in the Twentieth Century Paige Hall Smith, Bernice L. Hausman, and Miriam Labbok, Beyond Health, Beyond Choice: Breastfeeding Constraints and Realities Matthew Smith, An Alternative History of Hyperactivity: Food Additives and the Feingold Diet Rosemary A. Stevens, Charles E. Rosenberg, and Lawton R. Burns, eds., History and Health Policy in the United States: Putting the Past Back In Barbra Mann Wall, American Catholic Hospitals: A Century of Changing Markets and Missions Frances Ward, The Door of Last Resort: Memoirs of a Nurse Practitioner