Qualitative research in logistics and supply chain management 9781781904046, 9781781904039

Citation preview

ijpdlm cover (i).qxd

24/10/2012

13:21

Page 1

ISSN 0960-0035

Volume 42 Number 8/9 2012

International Journal of

Physical Distribution & Logistics Management Special issue on qualitative research in logistics and supply chain management Guest Editors: Britta Gammelgaard and Dan Flint Published in association with

www.emeraldinsight.com

International Journal of

ISSN 0960-0035

Physical Distribution & Logistics Management

Volume 42 Number 8/9 2012

Special issue on qualitative research in logistics and supply chain management Guest Editors Britta Gammelgaard and Dan Flint

Access this journal online __________________________ 719

CONTENTS

Editorial advisory board ___________________________ 720 Guest editorial ____________________________________ 721 Implementing mixed methods research in supply chain management Susan L. Golicic and Donna F. Davis _______________________________

726

Increasing the rigor of grounded theory research – a review of the SCM literature Nikola Denk, Lutz Kaufmann and Craig R. Carter ____________________

742

Moving beyond the systems approach in SCM and logistics research Fredrik Nilsson and Britta Gammelgaard ____________________________

764

A reviewer’s guide to the grounded theory methodology in logistics and supply chain management research Ila Manuj and Terrance L. Pohlen _________________________________

784

Rigor in qualitative supply chain management research: lessons from applying repertory grid technique Keith Goffin, Jawwad Z. Raja, Bjo¨rn Claes, Marek Szwejczewski and Veronica Martinez ______________________________________________

804

Using the “documentary method” to analyse qualitative data in logistics research Alexander Trautrims, David B. Grant, Ann L. Cunliffe and Chee Wong ____________________________________________________

828

This journal is a member of and subscribes to the principles of the Committee on Publication Ethics

CONTENTS continued

Towards a methodology for studying supply chain practice Benedikte Borgstro¨m ____________________________________________

843

Grounded theory: an inductive method for supply chain research Wesley S. Randall and John E. Mello________________________________

863

www.emeraldinsight.com/ijpdlm.htm As a subscriber to this journal, you can benefit from instant, electronic access to this title via Emerald Management eJournals. Your access includes a variety of features that increase the value of your journal subscription.

Structured abstracts Emerald structured abstracts provide consistent, clear and informative summaries of the content of the articles, allowing faster evaluation of papers.

How to access this journal electronically

Additional complementary services available

Our liberal institution-wide licence allows everyone within your institution to access your journal electronically, making your subscription more cost-effective. Our web site has been designed to provide you with a comprehensive, simple system that needs only minimum administration. Access is available via IP authentication or username and password.

When you register your journal subscription online you will gain access to additional resources for Authors and Librarians, offering key information and support to subscribers. In addition, our dedicated Research, Teaching and Learning Zones provide specialist ‘‘How to guides’’, case studies, book reviews, management interviews and key readings.

To benefit from electronic access to this journal, please contact [email protected] A set of login details will then be provided to you. Should you wish to access via IP, please provide these details in your e-mail. Once registration is completed, your institution will have instant access to all articles through the journal’s Table of Contents page at www.emeraldinsight.com/ 0960-0035.htm More information about the journal is also available at www.emeraldinsight.com/ijpdlm.htm

E-mail alert services These services allow you to be kept up to date with the latest additions to the journal via e-mail, as soon as new material enters the database. Further information about the services available can be found at www.emeraldinsight.com/alerts

Emerald online training services Visit www.emeraldinsight.com/help and take an Emerald online tour to help you get the most from your subscription.

Emerald Research Connections An online meeting place for the world-wide research community, offering an opportunity for researchers to present their own work and find others to participate in future projects, or simply share ideas. Register yourself or search our database of researchers at www.emeraldinsight.com/connections

Key features of Emerald electronic journals

Choice of access

Automatic permission to make up to 25 copies of individual articles This facility can be used for training purposes, course notes, seminars etc. within the institution only. This only applies to articles of which Emerald owns copyright. For further details visit www.emeraldinsight.com/copyright Online publishing and archiving As well as current volumes of the journal, you can also gain access to past volumes on the internet via Emerald Management eJournals. You can browse or search these databases for relevant articles. Key readings This feature provides abstracts of related articles chosen by the journal editor, selected to provide readers with current awareness of interesting articles from other publications in the field. Non-article content Material in our journals such as product information, industry trends, company news, conferences, etc. is available online and can be accessed by users. Reference linking Direct links from the journal article references to abstracts of the most influential articles cited. Where possible, this link is to the full text of the article. E-mail an article Allows users to e-mail links only to relevant and interesting articles to another computer for later use, reference or printing purposes.

Electronic access to this journal is available via a number of channels. Our web site www.emeraldinsight.com is the recommended means of electronic access, as it provides fully searchable and value added access to the complete content of the journal. However, you can also access and search the article content of this journal through the following journal delivery services: EBSCOHost Electronic Journals Service ejournals.ebsco.com Informatics J-Gate www.j-gate.informindia.co.in Ingenta www.ingenta.com Minerva Electronic Online Services www.minerva.at OCLC FirstSearch www.oclc.org/firstsearch SilverLinker www.ovid.com SwetsWise www.swetswise.com

Emerald Customer Support For customer support and technical help contact: E-mail [email protected] Web http://info.emeraldinsight.com/products/ subs/customercharter.htm Tel +44 (0) 1274 785278 Fax +44 (0) 1274 785201

IJPDLM 42,8/9

EDITORIAL ADVISORY BOARD

Professor Jan Albjørn University of Southern Denmark, Denmark Professor Chad W. Autry University of Tennessee, USA Professor Lauren Beitelspacher Portland State University, USA Professor Maria Bjo¨rklund Linko¨ping University, Sweden Professor James H. Bookbinder University of Waterloo, Canada Professor Christian Busse EBS Universita¨t fu¨r Wirtschaft und Recht, Germany Professor Craig R. Carter Arizona State University, USA Professor Joseph L. Cavinato Thunderbird – School of Global Management, USA Professor Karen Chapman Univeristy of Alabama, USA Professor Haozhe Chen East Carolina University, USA Professor Martin Christopher Cranfield School of Management, UK Professor David J. Closs Michigan State University, USA Professor Martha C. Cooper Ohio State University, USA Professor Thomas M. Corsi University of Maryland, USA Professor Paul D. Cousins Manchester Business School, UK Professor Michael R. Crum Iowa State University, USA Professor Marina Dabic University of Zagreb, Croatia Professor Patricia J. Daugherty Michigan State University, USA Professor Donna Davis Texas Tech University, USA Professor Jan de Vries University of Groningen, The Netherlands Professor George Deitz University of Memphis, USA Professor John Dinwoodie University of Plymouth Business School, UK Professor Lisa M. Ellram Miami University of Ohio, USA Professor Margaret A. Emmelhainz University of Georgia, USA Professor Philip T. Evers University of Maryland, USA Professor Nathalie Fabbe-Costes University of the Mediterranean, France Professor John Fernie Heriot-Watt University, UK Professor Ronald Fisher Griffith University, Australia Professor Helena Forslund International Journal of Physical Linnaeus University, Sweden Distribution & Logistics Professor Robert Frankel Management University of North Florida, USA Vol. 42 No. 8/9, 2012 p. 720 Professor Britta Gammelgaard # Emerald Group Publishing Limited Copenhagen Business School, Denmark

720

0960-0035

Professor Sebastian GarciaDastugue Universidad de San Andres, Argentina Professor Richard N. Germain University of Louisville, USA Professor Brian Gibson Auburn University, USA Professor Thomas J. Goldsby University of Kentucky, USA Professor Susan Golicic Colorado State University, USA Professor David B. Grant University of Hull, UK Professor Scott Grawe Iowa State University, USA Professor Stanley Griffis Michigan State University, USA ´ rni Halldo´rsson Professor A Chalmers University of Technology, Sweden Professor Trond Hammervol Harstad University College, Norway Professor Anthony F. Han National Chiao Tung University, Taiwan Professor Mike Harvey University of Mississippi, USA Professor Juliana Hsuan Copenhagen Business School, Denmark Professor Marianne Jahre BI Norwegian School and Lund University, Norway Professor Patrik Jonsson Chalmers University of Technology, Sweden Professor Lutz Kaufmann WHU-Otto Beisheim School of Management, Germany Professor Scott B. Keller The University of West Florida, USA Professor Timothy Kiessling Bond University, Australia Professor A. Michael Knemeyer The Ohio State University, USA Professor Chandra Lalwani The University of Hull, UK Professor Douglas M. Lambert The Ohio State University, USA Professor Paul Larson University of Manitoba, Canada Professor Alan C. McKinnon Ku¨hne Logistics University, Germany Professor John Mangan Newcastle University, UK Professor Gino Marchet Politecnico di Milano, Italy Professor David A. Menachof University of Hull, UK Professor Diane Mollenkopf University of Tennessee, USA Professor Paul R. Murphy John Carroll University, USA Professor Pieter J.A. Nagel Victoria University, Australia Professor Andreas Norrman Lund University, Sweden Professor John Ozment University of Arkansas, USA

Professor Photis Panayides Cyprus University of Technology, Cyprus Professor Stephen Pettit Cardiff University, UK Profesor Niall Piercy University of Bath, UK Professor Ram Mohan Pisharodi Oakland University, USA Professor Anthony Roath University of Bath, UK Professor Dale S. Rogers University of Nevada – Reno, USA Professor Ivan Russo University of Verona, Italy Professor Maria Jesus Saenz Zaragoza Logistics Centre, Spain Professor Katrina Savitskie University of Memphis, USA Professor Paola Signori University of Verona, Italy Professor Mohan Sodhi Cass Business School, UK Professor Amrik Sohal Monash University, Australia Professor Dong-Wook Song Heriot-Watt University, UK Professor Theodore Stank The University of Tennessee, USA Professor Gunnar Stefansson Chalmers University of Technology, Sweden Professor James R. Stock University of South Florida, USA Professor Frank Straube Berlin University of Technology, Germany Professor Peter Tatham Griffith University, Australia Professor Rodney Thomas Georgia Southern University, USA Professor Mert Tokman James Madison University, USA Professor Douglas Voss Central Arkansas University, USA Professor Stephan Wagner Swiss Federal Institute of Technology Zurich, Switzerland Professor Matt Waller University of Arkansas, USA Professor Judith Whipple Michigan State University, USA Professor Liu Xiaohung University of Erlangen-Nuremberg, Germany Professor Zhilin (Forrest) Yang City of Hong Kong University, People’s Republic of China

Guest editorial Qualitative research in logistics and supply chain management: beyond the justification for using qualitative methods This special issue on qualitative research in supply chain management (SCM) and logistics is of course about qualitative methods. But it is more than that; it is also about doing research that is qualitative in its essence. Qualitative research is not just about using for example interview data in a study, it is also about how one perceives qualitative data, the theoretical lens through which one conducts analyses, and the data analysis processes. Further it is about asking questions such as do the findings have truth value or do they provide insight based on interpretation in a subjective ontology? Such issues are seldom addressed in SCM and logistics journal papers even when qualitative research is presented. It is as if there is still a tacit agreement that there is one way to approach the discipline and that this way is assumed to still be a positivistic one as was stated about the logistics field 17 years ago by Mentzer and Kahn (1995). There are, however, more ways of conducting research and thereby extending and enriching the knowledge base of SCM. Qualitative research methods should be chosen when questions asked about a phenomenon require them. Thus, having the qualitative approaches in one’s toolbox enables far more research questions to be asked and helps to reveal more about the complexity of supply chain networks today. Sharing how some outstanding scholars approach qualitative research in SCM so that other researchers can leverage this knowledge is what this special issue is about. All eight papers in this special issue touch upon deeper issues than “just” discussing the legitimacy of using qualitative data and introducing new qualitative methods to the field without further reflection. With this special issue we are building on a tradition that in our perspective started with Ellram’s (1996) article, “The use of the case study method in logistics research.” Qualitative methods are still relatively sparse in the discipline even though an increasing number of papers adopting a qualitative approach can be observed throughout scientific journals across numerous fields (Sachan and Datta, 2005; update by Goffin et al. (2012) this special issue). Although still lagging behind quantitative approaches in our field, surveys in particular, qualitative methods are gaining ground and therefore deserve attention and greater discussion. From our perspective, this development is interesting not because we dislike surveys or math modeling, but because more approaches to conducting rigorous research enrich the discipline (Gammelgaard, 2004). Contributors to this special issue were asked to reflect on their own experiences in conducting qualitative research and share their insights rather than simply describe methodologies, something that can be done in other forums. Thereby, we wanted to enhance the collective field’s understanding of the use of qualitative approaches as well as qualitative aspects of SCM and logistics research. Such reflections are to be found especially in these articles: “A reviewer’s guide to the grounded theory methodology in logistics and supply chain management research” by Manuj and Pohlen (2012); “Rigor in qualitative supply chain management research: lessons from applying repertory grid technique” by Goffin, Raja, Claes, Szwejczewski and Martinez;

Guest editorial

721

International Journal of Physical Distribution & Logistics Management Vol. 42 No. 8/9, 2012 pp. 721-725 q Emerald Group Publishing Limited 0960-0035

IJPDLM 42,8/9

722

“Using the ‘documentary method’ to analyse qualitative data in logistics research” by Trautrims, Grant, Cunliffe and Wong; and “Towards a methodology for studying supply chain practice” by Borgstro¨m (2012). Some authors of articles here found additional ways of relating abstract methodological discussions to SCM and logistics research. For example, Denk, Kaufmann and Carter draw on a literature review of published studies using grounded theory (GT) in their article “Increasing the rigor of grounded theory research – a review of the SCM literature” while Golicic and Davis review studies of SCM and logistics using mixed methods (Denk et al., 2012). Nilsson and Gammelgaard as well as Borgstro¨m show ways of studying logistics and SCM issues from different and new angles. Nilsson and Gammelgaard (2012) point out in their article “Moving beyond the systems approach in SCM and logistics research” the importance of basic assumptions in the movement of the traditional systems approach towards higher and more contemporary levels of systems to depict the complexity of today’s global supply chains. Borgstro¨m (2012) proposes a practice approach to researching SCM in her article “Towards a methodology for studying supply chain practice.” By describing a constructivist/interpretivist approach, referred to as mystery methodology, Borgstro¨m contributes to a sparsely developed line of thought in SCM that emphasizes social action and organizational aspects of the field (Nilsson and Gammelgaard, 2012). It is worth noting that this special issue contains no less than three papers on GT following the Mello and Flint (2009) paper in JBL, “A refined view of grounded theory and its application to logistics research.” These three papers are Denk et al.’s (2012) “Increasing the rigor of grounded theory research – a review of the SCM literature,” Manuj and Pohlen’s (2012) “A reviewer’s guide to the grounded theory methodology in logistics and supply chain management research” and, Randall and Mello’s (2012) “Grounded theory: an inductive method for supply chain research.” With the extensive discussion of the use of GT in these three articles, the discipline has taken one more step in developing the qualitative method toolbox of the field. The special issue begins with Golicic and Davis (2012), “Implementing mixed methods research in supply chain management”. In this paper, they suggest that using both qualitative and quantitative methods within one study will advance the discipline by providing a richer understanding of phenomena and more robust explanations. The paper sets out by outlining the paradigmatic preconditions for using mixed methods and thereafter they present an overview of the use of mixed methods in the SCM and closely related fields. They show that using mixed methods is rare and further that often when using mixed methods this is not identified as such in the published research, creating problems when reviewing such contributions. By suggesting a balanced approach between qualitative and quantitative methods in a mixed method design, the authors claim that new fields of study, such as supply chain design, are much better understood with a mixed method study than with only a qualitative or qualitative, respectively. The relevance versus rigor discussion in SCM has gone on for some time (Mentzer, 2008). The question of rigor in GT studies in SCM is taken up by Denk et al. (2012) in the article “Increasing the rigor of grounded theory research – a review of the SCM literature.” After discussing the differences between the Glaserian and Straussian approach to grounded theories, they develop and discuss the quality criteria of each

approach and then use them to analyze published studies that have used this method. They conclude that there is still work to be done in making such studies more rigorous. But it is not only a question of mixing methods and being more rigorous in what is done that is needed to expand qualitative research in SCM according to Nilsson and Gammelgaard’s (2012) article “Moving beyond the systems approach in SCM and logistics research.” By disentangling the differences in basic assumptions about structure, human behavior and time in different systems approaches, they show an alternative way to study phenomena where human behavior in organizations is the focus of study. Quality and rigor is also the focus of Manuj and Pohlen’s (2012) article “A reviewer’s guide to the grounded theory methodology in logistics and supply chain management research.” In this article they propose recommendations for reviewing GT as this methodology is relatively new for the community of editors and reviewers in the field. By analyzing literature dealing with GT they present a reviewers’ guide to manuscript evaluation. The article by Goffin and colleagues is also aimed at increasing rigor in qualitative studies. In their article “Rigor in qualitative supply chain management research – lessons from applying repertory grid technique” these authors suggest a qualitative method that enables respondents to articulate views of complex issues where constructs are unclear at the outset. On the basis of coding and validating constructs from in-depth interviews, construct characteristics are counted based on frequency of appearance but at the same time given a weight of importance by a measure of average normalized variability (ANV). By this partially quantitative analysis of qualitative data, the repertory grid technique provides an insight into essential constructs and increased construct validity for subsequent quantitative studies. The documentary method proposed by Trautrims and colleagues, suggests using in-depth interviews as the background for getting more insight to business practice, with a focus on logistics micro processes, for example, in-store logistics for large retail outlets. In their article “Using the ‘documentary method’ to analyse qualitative data in logistics research” the authors emphasize the importance of rich descriptions of logistics micro processes by analyzing such descriptions through coding and conceptualizing the information towards higher levels of abstraction (Trautrims et al., 2012). The authors highlight authenticity as a major attribute of the method. By exploring logistics micro processes, the interview taps into tacit knowledge about how logisticians actually work on a particular process in a way where pure observation would not have been sufficient. Borgstro¨m points to connecting with actual practice as a very important line of inquiry in logistics and SCM. Richer insight through interpretation of supply chain and logistics processes provides the opportunity to explore unexpected outcomes in both theory and practice and to go beyond a normative approach common in SCM research. The “mystery methodology” analyzes field practices to explore the social construction of SCM and attempts to explain how participants interpret concepts and establish meaning. By doing this, researchers can understand better the dynamics and complexity of SCM. Capturing the complexity of supply chains – systems of supply chain systems – is a major argument for applying GT according to Randall and Mello’s (2012) article, “Grounded theory: an inductive method for supply chain research.” This final article in the special issue emphasizes GT as appropriate to relatively unexplored phenomena

Guest editorial

723

IJPDLM 42,8/9

724

within SCM, such as mindsets, culture, values and norms in supply chains. With a discussion of the paradigmatic standpoint of GT, the authors address GT research processes and consequently how GT research should be evaluated. Our conclusion from working with this special issue is that there is still a need for development of the methodological toolbox of the field. This special issue highlights tools related to “mixed methods”, “GT”, “repertory grid technique” and the “documentary method”, all of which should help researchers provide a deeper and richer understanding of logistics and SCM phenomena and at the same time deliver the much sought for rigor in qualitative studies. The articles in this special issue also suggest that logistics and SCM is a field continuously and constantly expanding, becoming more and more complex, and therefore needs a variety of methods to understand and explain it. “GT” is mentioned as one; but the “mystery methodology” is also an option. But departing from the normal approach to logistics and supply chain problems requires “rethinking basic assumptions” of theories and paradigms such as structure, behavior and relation to time. Finally, we would like to thank both authors and reviewers for embarking on this journey with us. We hope that both they and the readers will enjoy the final result. 18 papers were submitted on the final day of August 2010 and exactly one year later, eight of these papers were accepted for publication. Out of the ten papers that did not make it to the end, one was withdrawn in the process. B. Gammelgaard and D. Flint Guest Editors References Borgstro¨m, B. (2012), “Towards a methodology for studying supply chain practice”, International Journal of Physical Distribution & Logistics Management, Vol. 42 Nos 8/9 pp. 843-62. Denk, N., Kaufmann, L. and Carter, C.R. (2012), “Increasing the rigor of grounded theory research – a review of the SCM literature”, International Journal of Physical Distribution & Logistics Management, Vol. 42 Nos 8/9 pp. 742-63. Ellram, L.M. (1996), “The use of the case study method in logistics research”, Journal of Business Logistics, Vol. 17 No. 2, pp. 96-138. Gammelgaard, B. (2004), “Schools in logistics research? A methodology framework for analysis of the discipline”, International Journal of Physical Distribution & Logistics Management, Vol. 34 No. 6, pp. 479-91. Goffin, K., Raja, J.Z., Martinez, V., Claes, B. and Szwejczewski, M. (2012), “Rigor in qualitative supply chain management research – lessons from applying repertory grid technique”, International Journal of Physical Distribution & Logistics Management, Vol. 42 Nos 8/9 pp. 804-27. Golicic, S.L. and Davis, D.F. (2012), “Implementing mixed methods research in supply chain management”, International Journal of Physical Distribution & Logistics Management, Vol. 42 Nos 8/9 pp. 726-41. Manuj, I. and Pohlen, T. (2012), “A reviewer’s guide to the grounded theory methodology in logistics and supply chain management research”, International Journal of Physical Distribution & Logistics Management, Vol. 42 Nos 8/9 pp. 784-803. Mello, J. and Flint, D.J. (2009), “A refined view of grounded theory and its application to logistics research”, Journal of Business Logistics, Vol. 30 No. 1, pp. 107-25.

Mentzer, J.T. (2008), “Rigor versus relevance: why would we choose only one?”, Journal of Supply Chain Management, Vol. 44 No. 2, pp. 72-7. Mentzer, J.T. and Kahn, K.B. (1995), “A framework of logistics research”, Journal of Business Logistics, Vol. 16 No. 1, pp. 231-50. Nilsson, F. and Gammelgaard, B. (2012), “Moving beyond the systems approach in SCM and logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 42 Nos 8/9, pp. 764-83. Randall, W.S. and Mello, J. (2012), “Grounded theory: an inductive method for supply chain research”, International Journal of Physical Distribution & Logistics Management, Vol. 42 Nos 8/9, pp. 863-80. Sachan, A. and Datta, S. (2005), “Review of supply chain management and logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 35 No. 9, pp. 664-705. Trautrims, A., Grant, D.B., Cunliffe, A.L. and Wong, C. (2012), “Using the ‘documentary method’ to analyse qualitative data in logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 42 Nos 8/9 pp. 822-42.

Guest editorial

725

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0960-0035.htm

IJPDLM 42,8/9

Implementing mixed methods research in supply chain management

726

Susan L. Golicic Department of Management, Colorado State University, Fort Collins, Colorado, USA, and

Received 28 November 2011 Accepted 16 December 2011

Donna F. Davis Department of Marketing, Texas Tech University, Lubbock, Texas, USA Abstract Purpose – The purpose of this paper is to describe how to implement mixed methods research in supply chain management. Design/methodology/approach – A review of past journal analyses on research methods used in supply chain management-related disciplines is conducted to demonstrate the low incidence of mixed methods research in supply chain management. Drawing from literature on multiple and mixed methods research, the paper provides guidelines for designing and reporting such studies. Findings – Knowledge development in logistics and supply chain management relies primarily on single-method quantitative research designs, while mixed methods approaches are rarely used. Thus, there is a significant opportunity to advance the discipline through the rigorous application of mixed methods research. Research limitations/implications – Supply chain management phenomena are complex and dynamic. Thus, the application of mixed methods research would serve the advancement of the discipline as these approaches provide richer understanding and more robust explanations of such phenomena. Practical implications – If supply chain research is to keep up with the dynamic business environment, research methods must be applied with the capability to fully explain supply chain phenomena. The application of a single-method research approach is not always adequate for this task. Originality/value – This paper is the first to draw on research from various disciplines to investigate the use of mixed methods in logistics and supply chain management research. It examines its prevalence in the discipline, provides examples of its application from the supply chain management literature, prescribes how to implement mixed methods research, and describes the benefits and limitations of such designs. Keywords Mixed methods research, Qualitative research, Research purpose, Research design, Research methods, Supply chain management Paper type Research paper

International Journal of Physical Distribution & Logistics Management Vol. 42 No. 8/9, 2012 pp. 726-741 q Emerald Group Publishing Limited 0960-0035 DOI 10.1108/09600031211269721

1. Introduction Successfully managing supply chains requires synchronizing numerous activities within the network. There is a growing awareness that issues tied to supply chain management are increasingly complex and that diversity in methods studying these issues leads to more robust results (Craighead et al., 2007). Moreover, diversity in research methods “is a healthy characteristic and suggests an intellectual vitality in a discipline” (Stewart, 2007, p. 2). Research in logistics and supply chain management has been criticized for its lack of methodological diversity and unwillingness to employ

additional methods that may be more appropriate for theory generation and investigation of dynamic, complex phenomena (Naslund, 2002). Traditionally, supply chain management research has relied heavily on designs using quantitative methods, such as surveys, experiments, and mathematical models (Boyer and Swink, 2008; Mentzer and Kahn, 1995; Sachan and Datta, 2005). Qualitative methods (e.g. case studies and action research) have been embraced for some time by European researchers and have received more attention recently (Craighead et al., 2007; Spens and Kovacs, 2006; Taylor and Taylor, 2009). However, the combination of qualitative and quantitative methods within a single supply chain study (i.e. mixed methods research) is rare. The dominance of single-method quantitative studies undermines the robustness of the body of supply chain research in a number of ways. First, relying on a few methods confines inquiry to only those research questions that can be answered by those methods. Authors agree that supply chain phenomena are complex and that seeking knowledge through more than one type of research approach is needed to fully understand the phenomena we are trying to research (Frankel et al., 2005; Mangan et al., 2004). Second, reliance on a relative handful of methods introduces “certain inherent methods biases” in the development of theory and jeopardizes the evolution of a discipline (Deshpande, 1983; Spens and Kovacs, 2006). Finally, all research methods have benefits and limitations (Boyer and Swink, 2008); hence, researchers are advised to “use multiple methods, selected from different classes of methods with different vulnerabilities” to assure the trustworthiness of their inferences (McGrath, 1982, p. 207). Mixed methods research design resolves these issues by advocating the application of qualitative and quantitative approaches within a single research project to generate multiple perspectives of the phenomenon of interest and to attenuate the considerable risk of method bias. Due to the logistics of conducting mixed methods research, the use of this design in business research has been rare (Tashakkori and Teddlie, 1998). Moreover, there have been no studies to date that specifically examine the application of mixed methods research in logistics and supply chain management. Thus, the purpose of this research is to prescribe how to appropriately and rigorously design mixed methods research in the study of supply chain phenomena. We begin with a discussion of the definition and philosophy of mixed methods research. Next, we present a review of past journal analyses on the prevalence of research methods in disciplines related to supply chain management (e.g. purchasing, logistics, operations, transportation, and marketing) in order to explore the current state of mixed methods research in supply chain management, confirming its infrequent use. We then describe the underlying logic of qualitative and quantitative research approaches that are integrated in mixed methods research. We synthesize literature on mixed and multiple methods to offer guidelines for designing and reporting a mixed methods study and provide examples from the supply chain literature. We conclude with a discussion of the benefits and limitations of this research design and the implications for its use in advancing theory and practice in supply chain management. 2. Definition and philosophy of mixed methods research Mixed methods research is defined as the type of research in which a researcher, or a team of researchers, integrates qualitative and quantitative research approaches within a single study or a set of closely related studies (Creswell and Plano Clark, 2007; Johnson et al., 2007). The roots of mixed methods research design can be traced

Implementing mixed methods research 727

IJPDLM 42,8/9

728

to the introduction of the multitrait-multimethod (MTMM) matrix, which was developed to assure that observed variance is actually attributable to the trait under study, rather than the method (Campbell and Fiske, 1959). Subsequently, the term “triangulation” was introduced to describe investigations that combine multiple data sources and multiple methods (Denzin, 1978). Jick (1979) further explicates the concept of triangulation by describing “within-method triangulation” (i.e. multiple quantitative or multiple qualitative methods) and “across-methods triangulation” (i.e. combining qualitative and quantitative methods). It is important to note that mixed methods research is more narrowly defined than multiple methods or triangulation studies. That is, mixed methods research requires integration across qualitative and quantitative approaches, whereas multiple methods and triangulation also include within-method research designs (e.g. observation and interviews; surveys and experiments). Research methods are grounded in philosophical traditions in a discipline that stem from the prevailing paradigm, defined as a “basic set of beliefs that guides action” (Guba, 1990, p. 17). The philosophical paradigm of the researcher heavily influences the methods chosen to address research questions (Frankel et al., 2005). Kuhn’s (1962) influential book, The Structure of Scientific Revolutions, ignited the paradigm wars of the late twentieth century in the social and behavioral sciences. The opposing camps of paradigm purists have been described as the positivists versus the constructivists, with the constructivist paradigm advocating qualitative methods and the positivist paradigm supporting quantitative research methods (Tashakkori and Teddlie, 1998). The constructivist paradigm rests on the philosophical assumption that “knowledge is in the meanings people make of it,” and “is gained through people talking about their meanings” (Creswell, 1998, p. 19). In contrast, the positivist paradigm assumes that the world is external and objective, the researcher should focus on facts and look for causality, and preferred research methods should reduce phenomena to measurable operationalizations (Mangan et al., 2004). In the paradigm debates, these two world views were described as incommensurable. Hudson and Ozanne (1988, p. 508) observed that “incommensurability does not mean that the two approaches cannot peacefully coexist or that other middle-ground approaches cannot or should not be developed”. Several philosophical approaches are proposed to resolve the incommensurability debate. For example, critical realism reconciles the independent, objective nature of the physical world (i.e. realism) with the sensory experiences whereby we know about that world (i.e. critical). That is, some sense-data do not rely on subjective judgment and, therefore, are accurate representations of external objects while other sense-data are socially constructed perceptions (Bhaskar, 1997). Similarly, the transformative paradigm holds that there is one reality about which there are multiple, socially constructed viewpoints (Mertens, 2010). The compatibility thesis has also been argued from the viewpoint of paradigm relativism, which has been called “pragmatic relativism” or “pragmatism” (Howe, 1988; Tashakkori and Teddlie, 1998). Many mixed methods scholars argue for some version of pragmatism as the philosophical foundation for mixed methods research. Pragmatists adopt “whatever philosophical and/or methodological approach works for the particular research problem under study” (Tashakkori and Teddlie, 1998, p. 5). The basic belief of the pragmatic paradigm is that the research problem dictates the research method. Hence, pragmatism rejects the incompatibility thesis and claims that research paradigms may either remain separate or be combined in new research approaches.

The disciplines that have participated in the development of supply chain theory – purchasing, logistics, operations, transportation, and marketing – all share a history of primarily following the positivist paradigm (Burgess et al., 2006; Davis et al., 2011; Mentzer and Kahn, 1995). Methods used to investigate supply chain phenomena are, therefore, typically quantitative where larger samples are obtained to test theories and construct law-like generalizations. As the discipline of supply chain management has grown, scholars have begun to examine the influence of methodology on the development of theory. The heavy reliance on quantitative methods has resulted in an increasing number of studies that examine research methods in these disciplines in order to support calls for a more balanced approach to research using qualitative (Ellram, 1996; Mentzer and Kahn, 1995; Naslund, 2002) and/or multiple methods (Boyer and Swink, 2008; Golicic et al., 2005; Spens and Kovacs, 2006; Taylor and Taylor, 2009). In the following section, we examine the incidence of mixed methods research in supply chain management. 3. Prevalence of mixed methods research in supply chain management We inspected all of the published research method examinations we could find in supply chain management-related disciplines to determine the extent to which mixed methods approaches are utilized in the field. We did not perform our own journal review as there are already several that have been published fairly recently. A summary of the methodology reviews is provided in Table I. We reference only those reviews that specifically present research methods or designs; there are other reviews that investigate phenomena studied, definitions, and citations, but these were outside the scope of our examination (Croom et al., 2000; Pilkington and Meredith, 2009). Some of the reviews report specific approaches (e.g. surveys, case studies), which we then classified as quantitative or qualitative. We focus our analysis only on reports of mixed methods designs to draw conclusions; triangulation and multiple methods studies could rely solely on multiple quantitative methods without using qualitative approaches which are required for classification as mixed methods. Indeed, Davis et al. (2011) find the use of multiple methods in 4 percent of empirical studies in marketing, but only 40 percent of these (1.6 percent in total) used mixed methods. Nearly all of the methodology reviews confirm that the majority of empirical studies are based on quantitative methods (ranging from 46 to 85 percent), with many of these reporting surveys as the most frequently used quantitative method. Surprisingly, the reported frequency of studies using qualitative methods varies a great deal; this is likely due to the different journals included in the different reviews. One review reports qualitative methods in as few as 10 percent of the articles, and one finds 51 percent of the studies they examined to use qualitative methods (this particular study included any journal in which “supply chain management” was a subject of the article). Six of the reviews highlight the use of multiple methods, and three of those specifically examine mixed methods. One of the three reviews that examine mixed methods covered nearly 20 years of five marketing journals, finding only 1.6 percent of studies using mixed methods. The remaining two reviews focused on logistics journals with one finding 7.4 percent in six years of JBL articles and the other 8.9 percent in six years of JBL, IJLM and IJPDLM (Frankel et al., 2005; Spens and Kovacs, 2006, respectively). The difference in the number of mixed methods studies can be attributed to different years and different journals in the studies. Overall, mixed methods research designs are infrequently used in supply chain management, accounting for a very small percent of the empirical articles over the

Implementing mixed methods research 729

IJPDLM 42,8/9

730

Methodology findingsa (as reported or calculated)

Review period

49% quantitative 51% qualitative

1985-2003 ABI/inform SCM peer reviewed search 1965-2001 JSCM

71% quantitative 29% qualitative 85% quantitative 10% qualitative 81% quantitative 15.5% qualitative 1.6% mixed methods 56% quantitative 37% qualitative 63.0% quantitative 12.0% qualitative 4.6% multiple methods 7.4% mixed methods 87% quantitative 13% qualitative

Journalsb reviewed

Source

Burgess et al. (2006) 100 (random from 614) 774 (all) Carter and Ellram (2003)

1993, IJLM, JBL, TJ 299 (all) 1998, 2003 1990-2008 JAMS, JCR JM, 3,289 (all) JMR, MKS

Craighead et al. (2007)

1986-1990 IJPDLM, JBL, JPMM, LTR, TJ 1999-2004 JBL

Not reported

Dunn et al. (2004)

108 (all)

Frankel et al. (2005)

405 (those mentioning SCM)

Guinipero et al. (2008)

71 (those mentioning SCM) 235 (all)

Halldorsson and Arlbjorn (2005)

1997-2006 DS, IJLM, IJOPM, IJPDLM, IMM, JBL, JOM, JSCM, MS 47.5% quantitative methods 1997-2003 IJLM, IJPDLM, JBL 35% qualitative methods 17.5% multiple methods 83% quantitative 1978-1993 JBL 17% qualitative 55% quantitative 1999-2003 IJPDLM, JBL, SCMIJ 44% qualitative 2% multiple methods 49.5% quantitative 1998-2002 HBR, IJLM, IJPDLM, JBL 12.4% qualitative 8.9% mixed methods 46% quantitative 2004-2009 IJOPM 38% qualitative 10.5% multiple methods

Table I. Summary of methodology reviews in supply chain management disciplines

Papers reviewed

Davis et al. (2011)

442 (all)

Mentzer and Kahn (1995) Sachan and Datta (2005)

378 (all)

Spens and Kovacs (2006)

310 (all)

Taylor and Taylor (2009)

Notes: aPercentages are reported based on the empirical papers and excluding conceptual papers and literature reviews, Qualitative is often reported as the number of interpretive, case study, interview and/or action research papers, Quantitative is often reported as the number of positivist, survey, experiment, modeling, archival data, and/or simulation papers, Triangulation is reported as using multiple methods, however, no information is provided regarding the rigor of the multiple methods used; bJournal Abbreviations: DS – Decision Sciences, HBR – Harvard Business Review, IJLM – International Journal of Logistics Management, IJOPM – International Journal of Production Management, IJPDLM – International Journal of Physical Distribution and Logistics Management, IJPR – International Journal of Production Research, JAMS – Journal of the Academy of Marketing Science, JBL – Journal of Business Logistics, JCR – Journal of Consumer Research, JM – Journal of Marketing, JMR – Journal of Marketing Research, JOM – Journal of Operations Management, JSCM – Journal of Supply Chain Management Formerly Journal of Purchasing and Materials Management, LTR – Logistics and Transportation Review, MS – Management Science, MKS – Marketing Science, POM – Production and Operations Management, SCMIJ – Supply Chain Management and International Journal, TJ – Transportation Journal

past 16-20 years. Six of the reviews did not report multiple or mixed methods as a method category, providing more evidence for the rarity of such studies. Only eight of the 12 reviews drew conclusions about methods use from their findings. Two of these simply called for broadening or expanding methods (beyond reliance on quantitative methods), one suggested triangulation for increased validity, and one noted a need for more relevance. One article did call for more abductive research which combines deductive and inductive methods; however, this does not necessarily equate to combining qualitative methods with quantitative. Two articles discussed incorporating qualitative and quantitative methods in a study to add richness and alternative viewpoints (Carter and Ellram, 2003; Taylor and Taylor, 2009). However, these discussions were presented as suggestions for future research without providing guidance for how one might accomplish this. The most recent article implored authors to increase the use of mixed methods designs and described the benefits that could be obtained from such studies (Davis et al., 2011), but this study was specific to marketing. According to the reviews, there are increasing numbers of articles using multiple or mixed methods in more recent years. This increase is probably due to the growth in the number of empirical articles and application of qualitative methods. The low incidence of mixed methods designs may demonstrate a lack of common understanding of what constitutes a mixed methods study. We next prescribe how to appropriately implement this research design in supply chain management, which should help to clarify its understanding.

Implementing mixed methods research 731

4. Designing mixed methods research At the heart of supply chain research is the researcher’s idea or question about a particular phenomenon. Once the phenomenon is identified, the researcher sharpens the focus of the proposed study by developing research questions. The choice of an initial research approach (i.e. qualitative or quantitative) should depend on what the researcher wants to know as determined by the nature of the phenomenon and the type of research question(s). The balanced approach to research presented by Golicic et al. (2005) provides a suitable framework for considering the choice of an initial research approach and describing the data collection and analysis steps on each path in a mixed methods study (Figure 1). Qualitative Approach

Quantitative Approach

Data Collection

Literature Review

Phenomenon

Description

Substantive Theory

Source: Golicic et al. (2005)

Formal Theory

Field Verification

Figure 1. Balanced approach model

IJPDLM 42,8/9

732

4.1 Choosing an initial research approach When the phenomenon of interest is new, dynamic, or complex, relevant variables are not easily identified and extant theories are not available to explain the phenomenon (Creswell, 1998). In this situation, a qualitative approach is often the preferred starting point in order to build an understanding grounded in a detailed description of the phenomenon generated by collecting field data. The qualitative approach provides researchers with access to deeper levels of understanding of new or complex phenomena by yielding a high level of detail. For example, the relatively new phenomenon of sustainability in supply chain management would be an appropriate topic for initiating research with the qualitative approach. Before we can accurately measure this phenomenon, we must first identify and understand the relevant variables. Context is intrinsic to the phenomenon in the qualitative approach; therefore, phenomena that involve the exploration of well-known concepts in new contexts (e.g. What is the meaning of “brands” in the supply chain context?) are also a good match for using qualitative methods to initiate research. Qualitative research questions often start with “what” or “how” indicating the researcher’s aim to describe a process (e.g. How is supply chain management implemented at different tiers in a supply chain?). On the other hand, phenomena that have been fully described and documented through previous research frequently lend themselves to beginning with a quantitative approach. In this case, the researcher can confidently turn to the literature to identify relevant variables, discover gaps in understanding that need further attention, and develop measures for research instruments. For example, business relationships and integration in the supply chain have received a great deal of attention; however, gaps in understanding still exist. Research questions aimed at explaining relationships among variables by examining variation are ideal for the quantitative approach (Creswell, 1998). These questions aim to determine the degree of variability in the phenomenon by asking “why” or “to what extent” (e.g. to what extent do suppliers’ reverse logistics practices impact cooperation and operational performance in the supply chain?). Before choosing an initial research approach, the researcher is advised to seriously consider the question, “How much do we (the discipline) know about this phenomenon?” If the answer implies the research focus is to develop an understanding of new or complex phenomena, then the qualitative approach is typically the best starting point. If the researcher aims to take a more general view in order to explain relationships or demonstrate cause-and-effect among well-researched concepts, then the broader view provided by the quantitative path is often more appropriate. An inductive approach is needed to begin to understand and generate substantive theory about new and/or complex phenomena while a deductive approach is better suited for developing and then testing formal theory. Mixed methods research design begins with the choice of an initial research approach and then progresses through the circles shown in Figure 1, tacking between qualitative and quantitative approaches. 4.2 The qualitative path When the qualitative approach is chosen as the starting point, the aim is commonly to understand the phenomenon in its context (Hirschman, 1986); thus, the first step on the qualitative path is data collection. Typically, researchers will observe phenomena in natural settings in order to “make sense of, or interpret, phenomena in terms of the

meanings people bring to them” (Denzin and Lincoln, 2005, p. 3). Qualitative methods are particularly effective for: [. . .] understanding the nature of personal experiences, providing insights that are difficult to obtain from quantitative methods, understanding underlying meanings in human interactions and relationships in organizational settings, and in researching areas where there is little previous knowledge (Mello and Flint, 2009, p. 108).

The literature is used for support and is embedded in various stages, depending on the qualitative research tradition. The second step in the qualitative path is to describe the phenomenon. Descriptions that explore the multiple dimensions and properties of the phenomenon are generated using qualitative methods such as case study, phenomenology, ethnography, and grounded theory. Data collection techniques include asking open-ended questions and examining multiple data sources, which can take the form of interviews, observations, documents, and audiovisual materials (Hirschman, 1986; Maxwell, 1996). When the qualitative path is chosen as a starting point, the aim is often to build a substantive theory from descriptive data. Inductive theory generation moves from the parts – grounded in field data – to the whole (Glaser and Strauss, 1967; Strauss and Corbin, 1998). “Qualitative data are analyzed working inductively from detailed parts to more general perspectives that may be called categories, themes, dimensions, or codes,” depending on the analytical method selected by the researcher (Golicic et al., 2005, p. 21). The analysis yields a substantive theory of the phenomenon describing relationships among variables with feedback loops that capture the dynamic nature of the phenomenon, bringing the researcher full-circle to a deeper understanding of the core phenomenon. 4.3 The quantitative path When quantitative methods are used as a starting point, the goal is most often to build and test formal theory that explains and predicts the phenomenon of interest. The first step in the quantitative path is to thoroughly review appropriate literature in order to develop a conceptual framework that specifies relevant variables and expected relationships among them (Bickman and Rog, 1998). The researcher may also enter the field to conduct interviews at this stage, but it is generally for the purpose of developing and refining measures or clarifying the variables and relationships among them. The next step is to build a formal theory grounded in the reviewed literature. The formal theory developed in the quantitative path should be capable of generating predictive hypotheses that can be tested with real-world data about the phenomenon (Hunt, 1991). In the third step, data is collected through measurement instruments administered in field surveys or experiments. The data are then used to evaluate the strength of the formal theory by testing the significance and strength of proposed relationships among the variables expressed in the hypotheses. The conclusion of a quantitative study brings the researcher full-circle to a greater ability to explain and predict the phenomenon. Mixed methods studies b egin with an initial research approach, progress through the stages of that approach, and then cross over to a different approach to gain new insights within a single research project. Thus, studies tack back and forth between qualitative and quantitative approaches, as shown in Figure 1, sometimes repeating a cycle with one approach before crossing over to a different approach. The next step in

Implementing mixed methods research 733

IJPDLM 42,8/9

734

designing this type of research is to consider the purpose for synthesizing the mixed methods, which impacts the timing and weight of the individual methods. 4.4 Integrating the qualitative and quantitative methods Tashakkori and Teddlie (1998) describe four basic research purposes for mixed methods research: . development, or use of one study to inform a subsequent study; . initiation, or the use of a preliminary study to launch the main study; . complementarity, or concurrent examination of various facets of a phenomenon through two or more studies; and . interpretation, or the concurrent use of a second study to explain or confirm the results of the main study. Creswell and Plano Clark (2007) propose the two fundamental design decisions are the “weight” or relative level of reliance assigned to each method (equal or unequal) and the “timing” or temporal order of the use of methods (sequential or concurrent). Figure 2 combines the two design decisions and four research purposes in a matrix that displays the basic mixed methods research designs. In the following sections, each of the designs is explained with an example of its application provided from the supply chain literature. 4.4.1 Development. When equally weighted methods are implemented sequentially, the second method is used for the development of findings from the first method. The researcher’s intent is to use the results obtained from the first method to inform a subsequent study, thereby expanding the insights generated about the research problem. Results are reported separately for each phase of the study, followed by a general discussion that ties them together by comparing and contrasting findings. WEIGHT

Equal

Results

METHOD 1

Unequal

Results

Method 1

Sequential Results

METHOD 2 Development

TIMING

METHOD 1

Results

METHOD 2

Results

Initiation

METHOD 2

METHOD 1

Results

Concurrent Method 2

Figure 2. Mixed methods research designs

Complementarity

Source: Davis et al. (2011)

Interpretation

For example, Gammelgaard and Larson (2001) employ a development design to examine skills and competencies for logistics managers. A survey is used to assess the importance of various “context-independent SCM skill areas” using “generally-accepted definitions”; case studies are then used to provide “a richer understanding of skills and context-dependent competencies” (p. 29). Findings reveal interesting similarities and differences across the two studies. Seven of the top ten skill areas on the survey were also mentioned during the case study interviews. However, the top-ranked survey item – “teamwork” – was not mentioned by any informant in the case study interviews. Furthermore, the only skill area that emerged in all case study interviews – “gathering and sharing information” – was not included on the survey. The authors conclude that the method of mixing a survey and case study permitted them to apply quantitative measures as well as probe more deeply into the context to develop a richer understanding of the phenomenon. 4.4.2 Initiation. Similar to the development design, the intent of the initiation design is to use results from an initial study to inform a second study that uses a different method. However, the initiation design differs from the development design in that the methods have unequal weights. The less heavily weighted method is employed to initiate the research and is secondary to the primary method that is used in the main study. Results are reported separately, but the focus of the discussion is on the main study. The initiation design is adopted when there is a need for preliminary exploration of the phenomenon to identify relevant variables or predict relationships among variables (Creswell and Plano Clark, 2007; Tashakkori and Teddlie, 1998). Pagell and Krause (1999) use a survey of US plant managers to initiate their study of the relationship between environmental uncertainty and manufacturing flexibility. The primary method, used in the second phase of the study, relies on depth interviews of informants at 30 plants in a four-state region. The authors report that they initiated the study with a mail survey for three reasons: (1) to provide a benchmark comparison with the non-randomized depth interviews in the primary study; (2) to provide statistical power for inferences drawn from the interviews; and (3) to allow refinement of the depth interview protocol. Results are reported separately, with the focus on the report of findings from the depth interviews. Findings provide evidence that, contrary to findings of many previous studies, environmental uncertainty does not always drive flexibility. 4.4.3 Complementarity. The purpose of the complementarity design is to examine different, but complementary, aspects of the same phenomenon to address the research question. In this design, the researcher employs equally weighted methods in a one-phase design; that is, the researcher analyzes and interprets the data from multiple studies concurrently, merging the findings in a single report of results. Unlike the development and initiation designs, the sequencing of methods is not relevant to answering the research questions; that is, results achieved using one method do not inform the design or implementation of subsequent methods. The data may be collected concurrently (e.g. experiments and interviews) or sequentially (i.e. case study followed by surveys). However, the data are analyzed and interpreted in a single report of results.

Implementing mixed methods research 735

IJPDLM 42,8/9

Fawcett and Magnan (2002) set out to obtain an accurate depiction of supply chain management as it is actually practiced. They gather data from key informants to examine two perspectives of supply chain management practice: . across functions within the firm; and . across positions in the supply chain.

736

A large-scale mail survey of purchasers, logisticians, and manufacturing managers is used to capture the functional perspective. 52 depth interviews are used to explore the perspectives of managers in firms at various positions in the supply chain (e.g. retailer, finished-goods assembler, first- and second-tier suppliers). The authors report that combining surveys with interviews in this complementary research design yields “a rich and robust view of modern supply chain integration practice” from different points of view (p. 343); the descriptive data collected in the survey are contextualized by the depth interviews in the report of findings. They conclude that actual practice does not resemble supply chain theory: “Even among the best supply chain companies, integrative practice tends to span only a triad of companies” (p. 358). Findings identify three levels of supply chain management implementation and limiting factors to successful implementation. 4.4.4 Interpretation. The interpretation design is used when the researcher anticipates the need to explain or confirm findings from the main study. Unlike the complementarity design, the methods are not equally weighted. It is a one-phase, concurrent design in which a secondary method is employed to interpret or support results obtained from the primary method. The data may be collected sequentially, but are used concurrently in the interpretation and report of findings. Burgess et al. (2005) combine a survey with interviews in their study of configuration management in the aerospace industry. The research asks a question about process in a specific context: “How is the discipline of configuration management (CM) practiced in the European aerospace industry?” The primary data are collected in a mail survey of managers at European aerospace companies. The questionnaire is designed to collect descriptive data about CM from managers in 30 first-tier companies (i.e. companies in direct contact with customers) and 180 second-tier companies (i.e. suppliers to first-tier firms). To aid in the interpretation of findings, they conduct interviews with managers at six companies that represent a cross-section of the aerospace companies. The interviews offset the “lack of depth and inflexibility” of the survey questionnaire and were designed to elicit critical process information that would “add value” to the data collected in the survey (p. 295). The report of results integrates the tabulated data from the survey with insights gained from the interviews. Findings demonstrate differences between first-tier firms and second-tier firms in the practice of configuration management. The four examples provided from the literature were chosen to highlight mixed methods research published in different supply chain management journals (two from logistics and two from operations). Interestingly, all four pair depth interviews (one within a case study analysis) with surveys for the research design. Depth interviews appear to be the qualitative data collection method of choice to develop, explain and/or interpret quantitative survey results in logistics and supply chain management. 4.5 Reporting mixed methods studies While it may sound obvious, authors should explicitly identify their research designs as mixed methods. Searching for examples of mixed methods research in supply chain

management for this study was more difficult than necessary because most authors do not use the term “mixed methods” to describe their methodology. Further, authors should state the research purpose of their mixed methods studies (i.e. development, initiation, complementarity, or interpretation) and provide a description of the design decisions (i.e. weight and timing of the studies), data sources, and specific qualitative and quantitative methods used. The previous discussion of research purposes provides guidelines for reporting results, that is, when study results should be reported separately and when they should be integrated in the report of findings. Using common terminology and a consistent reporting format makes the method clearer for the reader and provides good examples for other researchers who might consider implementing mixed methods research design. 5. Discussion and implications This study is the first to specifically review the prevalence of mixed methods research in logistics and supply chain management. The examination of recent reviews of methods used in supply chain management-related disciplines found that mixed methods research design is relatively rare, representing a very small percentage of published studies. Further, the examination revealed that there is ambiguity around the classification of mixed methods studies. This suggests there is a need in the discipline for greater clarity about the design and implementation of mixed methods research, which is the primary aim of this study. While other reviews have called for combining methods within a study, none have provided guidance to accomplish this. We therefore synthesize literature on multiple and mixed methods research to present a framework for integrating qualitative and quantitative approaches in mixed methods research and describe research designs for four mixed methods research purposes (Golicic et al., 2005; Tashakkori and Teddlie, 1998). As demonstrated in the published examples provided, mixed methods allow for a wide variety of supply chain research questions to be answered and provide strong, robust results (Boyer and Swink, 2008; Davis et al., 2011; Johnson and Onwuegbuzie, 2004). The implementation of mixed methods research in supply chain management would be advanced if researchers who employ such methods use consistent terminology and formats to report their studies, as described in this paper. Moreover, it is important that researchers refrain from identifying their work as mixed methods if they do not fully report multiple studies that use both qualitative and quantitative methods and rely on those studies to draw conclusions. A study cannot be classified as mixed methods when one method (such as depth interviews) is not described and neither the analysis nor results are provided. Indeed, both older and more recent reviews have decried a lack of rigor in applying and reporting qualitative methods based on information provided in published research (Dunn et al., 2004; Taylor and Taylor, 2009). We concur with Mello and Flint (2009) that there are too many studies that fail to fully and accurately report the qualitative methods in their research designs. Supply chain management is a relatively new as well as complex discipline, and thus many of the phenomena of interest are also new and complex. These are precisely the type of concepts that lend themselves to mixed methods research. For example, supply chain strategy has begun to receive increased attention. However, strategy is typically examined within a single firm. In order for supply chain strategy to be fully explained, it should be examined across multiple firms. Similar to Fawcett and Magnan’s (2002) study on the practice of supply chain management across firms, qualitative and

Implementing mixed methods research 737

IJPDLM 42,8/9

738

quantitative methods could be used to determine how supply chain strategy is implemented across firms linked in a supply chain. Sustainability, a salient phenomenon in current supply chain research, is another example of a fertile area that could benefit from a mixed methods approach. Much of this research is still trying to understand the phenomenon of sustainability in addition to being able to explain and predict its outcomes. While mixed methods research has several benefits, there are significant challenges to implementing such research. Mixed methods studies often require more time in data collection, analysis, and interpretation, which may be considered too great a cost for many researchers (Tashakkori and Teddlie, 1998). Unfortunately, tenure and promotion decisions often hinge on the quantity of publications, and the time needed to conduct rigorous mixed methods research may be prohibitive (Davis et al., 2011). This issue can be addressed by forming research teams with members who have expertise in different methodological approaches. Because supply chain management scholars are often trained in different disciplines (e.g. operations, logistics, marketing), they might be ideal candidates for method collaboration. Calls for studies using multiple and/or mixed methods have been increasing; thus, it is likely that journal editors will treat such manuscripts favorably. As with any research, this study has limitations that point to the need for further research. The current study examined the use of mixed methods research within a single research project. Hence, it offers no insights into the use of mixed methods in closely related studies across a research program. However, the relatively low incidence of qualitative studies (as compared to quantitative) in supply chain management found in the current investigation suggests that a mixed methods approach in a program of research is probably also relatively rare. Future research could investigate the extent to which supply chain researchers employ mixed methods in their research programs. Additionally, an in-depth examination of the phenomena studied and benefits and challenges of mixed methods research in various research streams might provide direction for the discipline on where and how to apply such research approaches. Ideally, the aim of rigorous, relevant research should be to use divergent means to produce convergent findings. Mixed methods research holds promise for advancing the supply chain management discipline and expanding knowledge beyond that obtained from relying on a limited set of quantitative methods. It is our hope that this study stimulates scholarly interest in increasing the implementation of mixed methods research in logistics and supply chain management. References Bhaskar, R. (1997), A Realist Theory of Science, 2nd ed., Verso, London. Bickman, L. and Rog, D.J. (1998), Handbook of Applied Social Research Methods, Sage, Thousand Oaks, CA. Boyer, K.K. and Swink, M.L. (2008), “Empirical elephants – why multiple methods are essential to quality research in operations and supply chain management”, Journal of Operations Management, Vol. 26 No. 3, pp. 337-48. Burgess, K., Singh, O.J. and Koroglu, R. (2006), “Supply chain management: a structured literature review and implications for future research”, International Journal of Operations & Production Management, Vol. 26 No. 7, pp. 703-29.

Burgess, T.F., McKee, D. and Kidd, C. (2005), “Configuration management in the aerospace industry: a review of industry practice”, International Journal of Operations & Production Management, Vol. 25 No. 3, pp. 290-301. Campbell, D. and Fiske, D.W. (1959), “Convergent and discriminant validation by the multitrait-multimethod matrix”, Psychological Bulletin, Vol. 54, pp. 297-312. Carter, C.R. and Ellram, L.M. (2003), “Thirty five years of the Journal of Supply Chain Management: where have we been and where are we going?”, Journal of Supply Chain Management, Vol. 38 No. 2, pp. 27-39. Craighead, C.W., Hanna, J.B., Gibson, B.J. and Meredith, J.R. (2007), “Research approaches in logistics: trends and alternative future directions”, International Journal of Logistics Management, Vol. 18 No. 1, pp. 22-40. Creswell, J.W. (1998), Qualitative Inquiry and Research Design: Choosing Among Five Research Traditions, Sage, Thousand Oaks, CA. Creswell, J.W. and Plano Clark, V.L. (2007), Designing and Conducting Mixed Methods Research, Sage, Thousand Oaks, CA. Croom, S., Romano, P. and Giannakis, M. (2000), “Supply chain management: an analytical framework for critical literature review”, European Journal of Purchasing & Supply Management, Vol. 6 No. 1, pp. 67-83. Davis, D.F., Golicic, S.L. and Boerstler, C.N. (2011), “Benefits and challenges of conducting multiple methods research in marketing”, Journal of the Academy of Marketing Science, Vol. 39 No. 3, pp. 467-79. Denzin, N.K. (1978), The Research Act, 2nd ed., McGraw-Hill, New York, NY. Denzin, N.K. and Lincoln, Y.S. (2005), The Sage Handbook of Qualitative Research, 3rd ed., Sage, Thousand Oaks, CA. Deshpande, R. (1983), “Paradigms lost: on theory and method in research in marketing”, Journal of Marketing, Vol. 47 No. 3, pp. 101-10. Dunn, S.C., Seaker, R.F. and Waller, M.A. (2004), “Latent variables in business logistics research: scale development and validation”, Journal of Business Logistics, Vol. 15 No. 2, pp. 145-72. Ellram, L.M. (1996), “The use of the case study method in logistics research”, Journal of Business Logistics, Vol. 17 No. 2, pp. 93-138. Fawcett, S.E. and Magnan, G.M. (2002), “The rhetoric and reality of supply chain integration”, International Journal of Physical Distribution & Logistics Management, Vol. 32 No. 5, pp. 339-61. Frankel, R., Naslund, D. and Bolumole, Y. (2005), “The ‘white space’ of logistics research: a look at the role of methods usage”, Journal of Business Logistics, Vol. 26 No. 2, pp. 185-209. Gammelgaard, B. and Larson, P.D. (2001), “Logistics skills and competencies for supply chain management”, Journal of Business Logistics, Vol. 22 No. 2, pp. 27-50. Glaser, B.G. and Strauss, A.L. (1967), The Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine De Gruyter, New York, NY. Golicic, S.L., Davis, D.F. and McCarthy, T.M. (2005), “A balanced approach to research in supply chain management”, in Kotzab, H., Seuring, S., Muller, M. and Reiner, G. (Eds), Research Methodologies in Supply Chain Management, Physica-Verlag, Heidelberg, pp. 15-30. Guba, E.G. (1990), The Paradigm Dialog, Sage, Newbury Park, CA. Guinipero, L.C., Hooker, R.E., Joseph-Matthews, S., Yoon, T.E. and Brudvig, S. (2008), “A decade of SCM literature: past, present and future implications”, Journal of Supply Chain Management, Vol. 44 No. 4, pp. 66-86.

Implementing mixed methods research 739

IJPDLM 42,8/9

740

Halldorsson, A. and Arlbjorn, J.S. (2005), “Research methodologies in supply chain management – what do we know?”, in Kotzab, H., Seuring, S., Muller, M. and Reiner, G. (Eds), Research Methodologies in Supply Chain Management, Physica-Verlag, Heidelberg, pp. 107-22. Hirschman, E. (1986), “Humanistic inquiry in marketing research: philosophy, method and criteria”, Journal of Marketing Research, Vol. 23 No. 3, pp. 237-49. Howe, K.R. (1988), “Against the quantitative-qualitative incompatibility thesis, or Dogmas die hard”, Educational Researcher, Vol. 17 No. 1, pp. 10-16. Hudson, L.A. and Ozanne, J.L. (1988), “Alternative ways of seeking knowledge in consumer research”, Journal of Consumer Research, Vol. 14 No. 1, pp. 508-21. Hunt, S.D. (1991), Modern Marketing Theory: Critical Issues in the Philosophy of Marketing Science, South-Western Publishing, Cincinnati, OH. Jick, T.D. (1979), “Mixing qualitative and quantitative methods: triangulation in action”, Administrative Science Quarterly, Vol. 24 No. 4, pp. 602-11. Johnson, R.B. and Onwuegbuzie, A.J. (2004), “Mixed methods research: a research paradigm whose time has come”, Educational Researcher, Vol. 33 No. 7, pp. 14-26. Johnson, R.B., Onwuegbuzie, A.J. and Turner, L.A. (2007), “Toward a definition of mixed methods research”, Journal of Mixed Methods Research, Vol. 1 No. 1, pp. 112-33. Kuhn, T.S. (1962), The Structure of Scientific Revolutions, University of Chicago Press, Chicago, IL. McGrath, J.E. (1982), “Dilemmatics: the study of research choices and dilemmas”, in McGrath, J.E. (Ed.), Judgment Calls in Research, Sage, Beverly Hills, CA, pp. 179-210. Mangan, J., Lalwani, C. and Gardner, B. (2004), “Combining quantitative and qualitative methodologies in logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 34 No. 7, pp. 565-78. Maxwell, J.A. (1996), Qualitative Research Design: An Interactive Approach, Sage, Thousand Oaks, CA. Mello, J. and Flint, D.J. (2009), “A refined view of grounded theory and its application to logistics research”, Journal of Business Logistics, Vol. 30 No. 1, pp. 107-26. Mentzer, J.T. and Kahn, K.B. (1995), “A framework of logistics research”, Journal of Business Logistics, Vol. 16 No. 1, pp. 231-50. Mertens, D.M. (2010), “Transformative mixed methods research”, Qualitative Inquiry, Vol. 16 No. 6, pp. 469-74. Naslund, D. (2002), “Logistics needs qualitative research: especially action research”, International Journal of Physical Distribution & Logistics Management, Vol. 32 No. 5, pp. 321-38. Pagell, M. and Krause, D.R. (1999), “A multiple-method study of environmental uncertainty and manufacturing flexibility”, Journal of Operations Management, Vol. 17 No. 3, pp. 307-25. Pilkington, A. and Meredith, J. (2009), “The evolution of the intellectual structure of operations management – 1980-2006: a citation/co-citation analysis”, Journal of Operations Management, Vol. 27 No. 2, pp. 185-202. Sachan, A. and Datta, S. (2005), “Review of supply chain management and logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 35 No. 9, pp. 664-705. Spens, K.M. and Kovacs, G. (2006), “A content analysis of research approaches in logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 36 No. 5, pp. 374-90.

Stewart, D.W. (2007), “New and improved! A look at the future”, Journal of the Academy of Marketing Science, Vol. 31 No. 1, pp. 1-4. Strauss, A.L. and Corbin, J.M. (1998), Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, Sage, Thousand Oaks, CA. Tashakkori, A. and Teddlie, C. (1998), Mixed Methodology: Combining Qualitative and Quantitative Approaches, Sage, Thousand Oaks, CA. Taylor, A. and Taylor, M. (2009), “Operations management research: contemporary themes, trends and potential future direction”, International Journal of Operations & Production Management, Vol. 29 No. 12, pp. 1316-40. About the authors Susan L. Golicic is an Associate Professor of Supply Chain Management and the Herbert Family Research Fellow at Colorado State University. Her research focuses on managing business relationships, strategy and sustainability in the supply chain. She has several years of professional experience in logistics and environmental engineering, has presented at many academic and practitioner conferences and has published in Journal of Operations Management, Journal of Business Logistics, International Journal of Physical Distribution & Logistics Management, Journal of the Academy of Marketing Science, Industrial Marketing Management, Sloan Management Review, Transportation Journal, The Wall Street Journal and Supply Chain Management Review. Susan L. Golicic is the corresponding author and can be contacted at: [email protected] Donna F. Davis is an Associate Professor of Marketing in the Rawls College of Business at Texas Tech University where she holds the Georgie G. Snyder Professorship. Her research interests are in supply chain relationships, brand management, and demand forecasting. Dr Davis has published research in the Journal of Supply Chain Management, Industrial Marketing Management, International Journal of Forecasting, International Journal of Physical Distribution & Logistics Management, and Journal of the Academy of Marketing Science. She has also presented her research at several national and international conferences.

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Implementing mixed methods research 741

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0960-0035.htm

IJPDLM 42,8/9

Increasing the rigor of grounded theory research – a review of the SCM literature

742 Received 16 December 2011 Accepted 16 January 2012

Nikola Denk and Lutz Kaufmann WHU – Otto Beisheim School of Management, Koblenz, Germany, and

Craig R. Carter WHU – Otto Beisheim School of Management, Koblenz, Germany and Arizona State University, Tempe, Arizona, USA Abstract Purpose – This study aims to examine the quality of the extant supply chain management (SCM) research which has utilized a grounded theory (GT) approach. The purpose of this research is to better understand the current state of the field, by introducing and highlighting the distinctions between the Glaserian and Straussian schools of thought and examining the extent to which existing SCM research has either complied with or diverged from the six dimensions which distinguish the two schools of thought. By doing so, it aims to provide guidelines to both reviewers and researchers who might use GT in future studies, with the goal of improving the validity and rigor of GT research. Design/methodology/approach – The method underlying this paper followed the steps of a systematic literature review process. GT works within leading SCM journals were examined to determine the extent to which they complied with the methodological tenets of GT. Findings – The systematic literature review shows that, while the use of GT in the field of SCM appears to be increasing over time, over half of the investigated studies deviate from the chosen school of thought by not adhering to the six dimensions distinct to Glaser’s or Strauss’s approach to GT. Research limitations/implications – This study calls for researchers to revisit the methodological roots of GT in order to improve the validity of such studies and ultimately the acceptance of the GT methodology by the broader community of SCM researchers. Transparency must be increased with regard to the chosen school of thought and the research process itself. GT is an appropriate methodology for investigating behavioral and social aspects of organizations and inter-organizational relationships, and thus should be utilized more frequently in future SCM research. Originality/value – The paper aids researchers in understanding the methodological tenets of grounded theory and the divergence of schools of thought within this methodology. Keywords Qualitative research, Grounded theory, Literature review, Supply chain management, Research Paper type Literature review

International Journal of Physical Distribution & Logistics Management Vol. 42 No. 8/9, 2012 pp. 742-763 q Emerald Group Publishing Limited 0960-0035 DOI 10.1108/09600031211269730

1. Introduction Qualitative research methods have the potential to play a pivotal role as the field of supply chain management (SCM) continues to mature (Carter, 2011). These research approaches are often more appropriate than quantitative methods when researchers want to investigate phenomena with complex behavioral dimensions (Mello and Flint, 2009). However, the use of qualitative approaches is relatively underrepresented in the empirical SCM literature, compared to their quantitative counterparts (Carter and Ellram, 2003). One specific qualitative approach is the grounded theory (GT) methodology, which Glaser (1992, p. 12) defined as, “a general methodology of analysis linked with data

collection that uses a systematically applied set of methods to develop inductive theory about a substantive area.” GT is a powerful methodology for developing or extending theory from empirical data (Finch, 2002; Easterby-Smith et al., 2008; Wagner et al., 2010), and complements the conceptual theory development approach which uses reasoning, logic, and extant, published works to develop theory (Choi and Wacker, 2011; Ketchen and Hult, 2011). Here, we define theory as, “a set of interrelated constructs (concepts), definitions, and propositions that present a systematic view of phenomena by specifying relations among variables, with the purpose of explaining and predicting phenomena,” (Kerlinger, 1986, p. 9). In this study, we focus specifically on the GT methodology, and its application in SCM research. In particular, we assess the quality of SCM research which employs the GT methodology. We do so by investigating the extent to which these extant studies have followed the methodological tenets put forward by the two schools of thought developed by the founders of GT, Barney Glaser and Anselm Strauss. Our analysis suggests that GT research in the field of SCM may be at risk of drifting away from the integrity and disciplinary origins of GT, thus possibly hampering the development of theory in our field (Piekkari et al., 2010; Suddaby, 2006; Cova and Elliott, 2008). At the same time, our systematic review of the SCM literature finds that the GT approach is not widely utilized. Based on these findings, our paper: . contributes to a better understanding of recent GT practices in the SCM literature and how these practices deviate from prescribed methodological tenets; and . calls for researchers, reviewers, and editors to thoroughly revisit the GT literature to more rigorously utilize this approach. This allows us to help guide future researchers regarding how to improve the rigor of their studies. The remainder of this paper is organized as follows. First, we provide an overview of the different schools of thought concerning GT, and the consequences for researchers who employ this methodology (Section 2). Based on this overview, we systematically review GT studies in the extant SCM research in Section 3, and we discuss the results of our analysis in Section 4. Finally, we conclude by calling for researchers in the field of SCM: . to utilize interpretive approaches when appropriate; and . to better understand the two distinct schools of GT and their implications for the process of utilizing GT in their own research, as well as when reviewing and evaluating papers that employ a GT approach. 2. The Glaserian and Straussian approaches to GT Glaser and Strauss (1967) developed GT in response to criticism directed at the lack of rigor in interpretive research. Thus, GT came about as a systematic approach to investigate subjective perceptions of reality, by providing guidelines for systematic and reproducible data analyses, including rigorous analytic procedures (Goulding, 1998). However, over time, Glaser and Strauss developed differing methodological views which led to two separate schools of thought (Stern, 1994). The two schools (Glaserian versus Straussian) differ along six dimensions, the first five of which were identified by Jones and Noble (2007):

Increasing the rigor of GT research 743

IJPDLM 42,8/9

744

(1) (2) (3) (4) (5)

emergence and researcher distance; theory development; specific, non-optional procedures; the core category; and coding procedures.

Moreover, the two approaches differ along a sixth dimension, which we label evaluation criteria. The following paragraphs briefly outline the differences between schools along each dimension. 2.1 Emergence and researcher distance Regarding theory emergence, Glaser follows the original GT approach, in which, “researchers maintain distance and independence from the phenomena they study,” (Locke, 1996, p. 241). In contrast, Strauss (1987) encourages researchers to build their studies on existing knowledge gained from prior personal and professional experiences. Thus, researchers who follow the Glaserian approach to GT should enter the field without prior assumptions or theoretical knowledge. This allows the emergence of a dominant theme (labeled as “core category” by grounded theorists) that summarizes patterns of behavior in the participants’ social context. The core category is the: [. . .] theoretical formulation that represents the continually resolving of the main concerns of the participants. The core category is central, it accounts for the variation in the pattern of behavior, reoccurs frequently in the data, and relates meaningfully and easily to other categories ( Jones and Noble, 2007, p. 89).

For example, Mello et al. (2008, p. 9) examined logistics outsourcing strategy by entering the field without any knowledge due to “the lack of specific theory concerning the processes companies use to evaluate logistics outsourcing as a strategy.” We acknowledge here that entering the field without any prior knowledge about the phenomenon under investigation is hardly realistic. Rather, we suggest that – in cases where researchers follow Glaser’s approach – they use their prior knowledge to see the relevant data and to fragment collected data into significant categories: Because no researcher can possibly obliterate all the previously learned theories, the trick is to line up what one takes as theoretically plausible with what one finds in the substantive field via an emergent fit (Sousa and Hendriks, 2006, p. 323).

Conversely, grounded theorists who enter the field with an initial idea or pre-defined research questions should follow Strauss’s assumptions. For example, in investigating supply risk assessment techniques, Zsidisin et al. (2004) explicitly reviewed prior research to ensure that their research contributed to a gap in the literature. Accordingly, they aimed at a better understanding of how and why purchasing organizations assess supply risk. The core category is then, “inductively derived from the study of the phenomenon it represents,” (Strauss and Corbin, 1990, p. 23). In other words, Strauss aims at discovering a “main theme” (Strauss, 1987, p. 35) embedded in a pre-determined context. Accordingly, theory emerges only out of this context. Drawing on a hypothetical example of studying sustainability in supply chains (Carter and Rogers, 2008; Pagell et al., 2010; Tate et al., 2010; Paulraj, 2011), researchers following the Glaserian approach would enter the field not knowing what they will

find in the data. Tying in with our discussion above, researchers who have prior knowledge about supply chains rely on this knowledge to precisely analyze recurring themes that dominate individual behaviors in supply chains. As a result, they may find that the behavior of managers in supply chains is predominantly driven by sustainability aspects. Conversely, researchers following the Straussian approach would enter the field determined to investigate sustainability and reflect their findings to prior knowledge they have collected about supply chains. In both cases, when researchers have identified sustainability as the dominant theme in supply chains, they might have a general (a priori) idea about sustainability but no specific idea about the particular facets of sustainability efforts. 2.2 Core category The core category is defined as the “central category of the phenomenon about which the theory is concerned. [As such, it] explains the majority of the behavior in an [investigated] phenomenon” (Mello and Flint, 2009, p. 111). Glaser views the concept of the core category as a pattern of behavior that constantly concerns the participants. Thus, the subsequent GT analysis equates to an emergent theory about that one core category (Jones and Noble, 2007). In other words, the product of GT under the Glaserian approach is an integrative theory that pulls together all shades of data in one overall scheme that researchers must identify. Thus, Glaser allows only one core category to emerge. Similar to our hypothetical study of sustainability and following the Glaserian GT approach, Flint and Golicic (2009) investigated several supply chains. Directed by the dominant theme that concerned all participants in this research setting, they found the emergent core category, i.e. the main concern of all participants in the investigated supply chains, to be searching for advantage through sustainability. This core category was not determined beforehand but emerged as the authors better understood how managers approached sustainability in a competitive environment. Managers used sustainability as a strategic tool in the collaboration with supply chain partners to gain additional advantage. Advantages through sustainability included an enhanced brand value and stronger relationships to suppliers and customers (Reuter et al., 2010). Conversely, Strauss’s concept of a core category refers to the pattern of behavior underlying a pre-determined research question. Thus, the subsequent GT analysis results in an emergent theory about the description of the social interactions that determine that pattern of behavior. This pre-determination of the research result either allows categories and sub-categories to be connected to one emerging core category, or simply does not force the emergence of a core category at all. 2.3 Theory development From the perspective of the Glaserian school, the objective of GT is to, “generate a conceptual theory that accounts for patterns of behavior,” (Glaser, 2003, p. 3) whereas Strauss (1987), later in Strauss and Corbin (1990, 1998), also encourages the use of GT for purposes besides theory development. The latter two acknowledge that, “some will use our techniques to generate theory, others for the purpose of doing very useful description, or conceptual ordering,” (Strauss and Corbin, 1998, p. 9). Thus, if the purpose of a study is solely theory building or extending, researchers should follow Strauss’s GT approach, because unlike Glaser, Strauss allows for the use of GT for purposes other than theory development. Strauss and Corbin (1998, p. 15) state that:

Increasing the rigor of GT research 745

IJPDLM 42,8/9

746

[. . .] if one’s ultimate research goal is to arrive at a set of findings rather than theory development, then integration [of data to generate a formal theory] is not as relevant.

For example, Whipple et al. (2009) compared firms purchasing and/or selling food products internationally to firms with domestically-oriented food supply chains, and relied on GT to ensure a deep understanding of supply chain processes in this industry. Based on this understanding, they developed a quantitative survey. Nevertheless, in this case grounded theorists must also follow the second stage of data analysis, that is, axial coding procedures which aim at the contextual refinement of the identified data categories. 2.4 Specific, non-optional procedures Glaser (2003, p. 5) states that GT has, “clear, extensive procedures,” that need to be followed, “if the study is to be recognized as a product of the grounded theory methodology,” (Glaser, 2001, p. 225). Strauss and Corbin (1998, p. 8) are more flexible, and incite researchers to, “choose, reject, and ignore [single procedures] according to their own tastes.” Thus, the Glaserian school requires a strict adherence to analytical procedures that include coding, constant comparison, theoretical sampling, memoing, category building, property development, densification, core category identification, delimitation, saturation, sorting, and communicating of research results. Only then can the methodology be recognized as GT. If researchers intend to omit single steps, they should base their work on the Straussian approach. Strauss and Corbin (1998, p. 4) regard these procedures as guidelines and thus the procedures need not be followed, “dogmatically but rather [. . .] used creatively and flexibly,” as researchers deem suitable to investigating the phenomenon of interest. 2.5 Coding procedures The Glaserian approach relies on open, selective, and theoretical coding procedures. In the early stages of analysis, Glaser allows a great deal of freedom, so that the researcher can identify as many categories as possible during the open coding phase. Open coding ceases when a core category emerges. In the selective coding phase, the grounded theorist delimits the coding to only those variables that relate to the identified core category. In the final step of Glaser’s approach, all categories are interlaced during the theoretical coding phase to form a theory. Strauss and Corbin base their analyses on open, axial, and selective categorization. Open coding is performed by grouping similar conceptual labels under a common category. Through a number of techniques, such as the flip-flop technique or waving the red flag, theoretical sensitivity is enhanced by developing each category in terms of its sub-categories, properties and dimensions (Strauss and Corbin, 1998). During the axial coding phase, the researcher further specifies the context for each category by delineating the conditions, the strategies of how to handle a specific category, and the consequences of those strategies. Only in the selective coding phase are the various categories interwoven to “form a larger theoretical scheme” (Strauss and Corbin, 1998, p. 143). For example, in describing their selective coding phase, Davis-Sramek and Fugate (2007, p. 31) state: We chose themes and categories based on them being central, appearing frequently, ensuring the explanation evolved by relating the categories was logical and consistent, and making sure that they could explain variation as well as the main point made by the data.

2.6 Evaluation criteria Glaser (1978) identifies four characteristics that are indispensible in evaluating the efficacy of the GT approach: (1) fit of the emergent concept with the analyzed social patterns; (2) workability with regard to the emergent theory’s ability to explain the phenomenon under study; (3) relevance of the concept with regard to the main concerns of the participants; and (4) modifiability, which accounts for the fact that additional data might modify the emergent theory. Extending theory from the salesperson control literature to the motor carrier industry, Mello and Hunt (2009) explicitly address these four characteristics. For example, they ensure fit by involving seven participants in verifying that the derived theory fits, in the interviewees’ experience, with the predominant theme of the phenomenon under investigation. Strauss and Corbin’s tradition of evaluating GT studies differs. Here, researchers need to exhaustively apply different criteria: credibility, transferability, dependability, confirmability, integrity, fit, understanding, generality, and control (for a more detailed explanation of each criterion, see Flint and Mentzer (2000)). Moreover, Strauss and Corbin (1990) interpretively address the reliability and generalizabilty criteria. Reliability in GT is reached when subsequent studies apply the same data collection and analysis processes under equal or at least similar conditions. Only then are other researchers able to generate the same theoretical findings. Generalizability (Yin, 2003; Blair and Zinkhan, 2006) in GT means that a, “grounded theory is generalizable insofar as it specifies conditions that are linked through action/interaction with definite consequences,” (Strauss and Corbin, 1990, p. 15). Thus, Strauss and Corbin stress the necessity for assessing both the adequacy of the research process and the grounding of empirical results. 2.7 Summary of the Glaserian versus Straussian schools In summary, Glaser (1999) defends the original approach to GT by challenging scholars’ creative potential to let theory freely emerge from the data. In contrast, Strauss (1987), later in Strauss and Corbin (1990, 1998), focuses on the technical coding procedures in the development of GT. Despite the differences between these two schools, they are certainly not mutually exclusive. Constant comparison (investigation of similarities and differences within and across incidents found in the data); category (identification of abstract concepts under which other concepts are summarized due to a shared uniformity) and property development (identification of characteristics or attributes that add meaning to a specific category); and systematic coding (organization of data into categories) are foundational procedures in both approaches. As Walker and Myrick (2006, p. 558) point out, it is not so much the differences between these two schools, but rather: [. . .] the understanding of these differences and making of informed and knowledgeable choices about what one will do in their research. Perhaps it is more about the researchers and less about the method.

In other words, when conducting GT, researchers must carefully choose the underlying school of thought so that the research purpose and processes do not deviate from the methodological assumptions.

Increasing the rigor of GT research 747

IJPDLM 42,8/9

748

With this overview (also see the Appendix) and these differences in mind, we next examine GT studies in the extant SCM literature, with regard to their compliance with the methodological tenets of the GT approach. 3. Review methodology The methodology underlying this paper followed the systematic literature review process described by Denyer and Neely (2004). We investigated research which has employed GT since the recent contribution to the debate by Glaser regarding how to conduct GT. Specifically, Glaser (2001, 2003, 2005) published three books in a series to defend his GT approach. We investigated the period after the publication of his second book in 2003, in order to include a reasonable sized sample in our analysis. Given the potential time lag between submission and publication of a manuscript, we investigated GT works during the period of 2004-2009. We used the mainstream logistics, transportation, and SCM journals encompassed by the Carter et al. (2009) studies. In addition, we included the Journal of Operations Management, which is generally considered to be another top-tier outlet for empirical, SCM research (Theoharakis et al., 2007) and which is often encompassed in studies which examine the status and progression of the field of SCM (Giunipero et al., 2008). Thus, our search encompassed the following nine journals: International Journal of Logistics Management (IJLM), International Journal of Physical Distribution and Logistics Management (IJPDLM), Journal of Business Logistics (JBL), Journal of Operations Management (JOM), Journal of Supply Chain Management (JSCM), Journal of Transport Economics and Policy (JTEP), Journal of the Transportation Research Forum (JTRF), Transportation Journal (TJ), and Transportation Research Part E (TRE). In total, the keyword search for the 2004-2009 time period revealed 358 full-text articles across the nine journals. Next, we investigated abstracts and methodological sections in the respective studies with regard to their use of the GT approach. We excluded 242 papers because the authors used a methodology other than GT. Although some of those 242 papers followed an interpretive research approach, they needed to be eliminated because GT studies “are conducted differently and are analyzed differently” (Mello and Flint, 2009, p. 107) compared to for example, case studies, phenomenological studies, or ethnographies as more generically defined. In contrast to other qualitative methods, GT, and the Glaserian approach in particular, seeks to build theoretical concepts and is not bound to the description of phenomena or limited to a particular unit of analysis, place, or time. Moreover, we eliminated two contributions because they analyzed the use and trends of different research methods, amongst others GT methodology, in the SCM domain, and we removed 82 duplicates. Thus, the final sample comprised 32 papers in IJLM, IJPDLM, JBL, JOM, JSCM, TJ, and TRE. (No GT papers were found in either the Journal of Transport Economics and Policy or the Journal of the Transportation Research Forum). We more closely reviewed these 32 papers with regard to the two underlying schools of thought and the six differing dimensions described previously. Therefore, the authors separately allocated each of the 32 GT studies to either the Glaserian or Straussian school of thought. The reliability of the coding was tested by having two researchers separately code the field notes. Intercoder reliability, defined as the number of agreements divided by the number of combined agreements and disagreements, was equal to 0.94, which surpassed the 0.90 threshold recommended by Miles and Huberman (1994).

For the remaining articles, initial disagreements between coders were resolved through discussion and via the use of a third coder where necessary. 4. Findings 4.1 Developments over time and by journal In the first stage of analysis, we investigated the development of GT studies in SCM research with regard to the journals in which papers were published and their year of publication. Figure 1 summarizes the results. The number of studies using GT – as a sole approach or in a mixed methods approach – is negligible compared to the overall research output in the SCM research community. The nine investigated journals published more than 1,500 articles in the period from 2004 to 2009. Accordingly, the 32 GT studies account for only 2 percent of the total research output. Carter and Dresner (2001) also found that interpretive methods have “been quite underutilized, particularly with respect to beginning to develop theoretical frameworks and propositions that may help to guide future empirical research.” However, consistent with other studies (Ellram, 2006), we find that GT is becoming more accepted in the SCM domain “as a legitimate way to add insights [. . .] that traditional empirical and modeling approaches cannot provide” (p. 15). In other words, it appears that SCM researchers are increasingly relying on this method to extend the knowledge base in their fields of interest: more than twice as many GT studies were published from 2007 to 2009 than from 2004 to 2006. Leading journals in terms of publication frequency of GT works are IJPDLM (9), JBL (8), and JOM (6). As illustrated in Table I, our findings show that more than half of all reviewed GT works in the SCM domain (59 percent ¼ 19/32) include deviations from one or more of the six dimensions distinct to either Glaser’s or Strauss’s GT approach. In other words, our systematic review of the literature provides evidence that GT studies in the SCM domain reflect a prevailing laxity with regard to the adherence to the core tenets of this methodology. This finding suggests that authors, as well as editors and reviewers, must have adequate knowledge concerning the disciplinary origins out of which the differing approaches of conducting GT grew. Reviewers can play a key role by ensuring that GT

Increasing the rigor of GT research 749

12 10

10

IJLM 2 8

JOM 1 JBL 2

6

4

4 TRE 1

2

JOM 2

0

IJPDLM 1 2004

2 JOM 1 JBL 1 2005 ∑ =10/32 articles (31%)

4 IJLM 1 TJ 1 JOM 1 JBL 1 2006

JSCM 1

IJPDLM 4

2007

6 TJ 1 JOM 1

6 TJ 2

JBL 2

JBL 2

IJPDLM 2

IJPDLM 2

2008

2009

∑ =22/32 articles (69%)

Figure 1. Development of GT studies in SCM research based on articles published from 2004 to 2009

IJPDLM 42,8/9

studies demonstrate adequate methodological validity. Editors can play an equally important role by ensuring the inclusion of methodological sections that demonstrate the transparency and traceability of GT research processes (perhaps as an appendix, if necessary), rather than eliminating these sections during the review process due to length restrictions of the publishing journal.

750

4.2 Deviations from core GT tenets In the second step of our analysis, we investigated the deviations from the six dimensions of GT in greater detail across each of the 32 articles. Table II displays our findings for each of the 32 papers, with deviations from the indicated school of thought noted by a delta (D). In line with the findings displayed in Table I, Table II illustrates that a large proportion of the 32 articles failed to adhere to one or more of the six key dimensions of GT. To facilitate our analysis, we categorized all contributions that relied on the works of Glaser and Strauss (1967) or Glaser (1978, 1992, 2001, 2003) as following the Glaserian approach and all GT works that based their contributions on the ideas of Strauss (1987) or Strauss and Corbin (1990, 1992, 1994, 1996, 1998) as following Strauss’s research tradition. In the few cases where papers did not clearly outline the underlying school of thought, we more closely investigated the research procedures. Here, the dimensions emergence and researcher distance and coding procedures served as strong indicators for whether the authors followed the Glaserian or the Straussian school of thought. For example, Vivek et al. (2008, p. 180) entered the research field with “a priori themes from the literature.” Manuj and Mentzer (2008, p. 195) indicated that “open, axial, [and] selective coding was followed.” Thus, despite the missing delineation of the underlying school of thought, both cases clearly reflected the adherence to the Straussian approach to GT and were allocated accordingly. We then analyzed the overall deviations for each school of thought and with regard to their development over time. Figure 2 summarizes the results. In contrast to the extant findings (Jones and Noble, 2007), we find that both schools of thought are popular in the SCM domain: 44 percent ( ¼ 14/32) of all investigated papers based their analyses on Glaser’s school of thought, whereas 56 percent ( ¼ 18/32) relied on Strauss’s tradition of conducting GT. However, grounded theorists in the field of SCM need to further increase the transparency regarding the disciplinary school that they are following. Authors often state that they have relied on GT guidelines and then simultaneously cite for example Glaser (2001, 2003), Glaser and Strauss (1967), Strauss (1987) and Strauss and Corbin (1990). In these cases, the underlying school of thought is only apparent when the reader is familiar with the methodology, and when the research

Table I. Evaluation of published GT works per journal from 2004 to 2009

Journal

Correctly applied

Correctly applied (%)

Incorrectly applied

Incorrectly applied (%)

Total

IJLM IJPDLM JBL JOM JSCM TJ TRE Total

1 5 5 0 0 1 1 13

33 56 63 0 0 25 100 41

2 4 3 6 1 3 0 19

67 44 38 100 100 75 0 59

3 9 8 6 1 4 1 32

TRE JOM JOM IJPDLM JBL JOM JBL IJLM JOM TJ JBL

Carter et al. (2004) Pagell (2004) Tucker (2004) Zsidisin et al. (2004) Flint et al. (2 005) Wu and Choi (2005) Fugate et al. (2006) Nilsson (2006) Sawhney (2006) Shao and Ji (2006) Davis-Sramek and Fugate (2007) Davis-Sramek et al. (2007) Esper et al. (2007) Haytko et al. (2007) Loesch and Lambert (2007) Matos and Hall (2007) Mollenkopf et al. (2007) Perry (2007) Whipple and Russell (2007) Yeung et al. (2007) Fugate et al. (2008) Manuj and Mentzer (2008) Mello et al. (2008)

Strauss Glaser Strauss Strauss Glaser Glaser Strauss Strauss Glaser Glaser Strauss

School of thought

Glaser

JOM

TJ

Glaser

IJPDLM Strauss JBL Strauss IJPDLM Strauss

IJPDLM Strauss IJLM Glaser

IJPDLM Strauss

Strauss Strauss Glaser

JBL IJLM JSCM

IJPDLM Strauss

Journal

Author(s) (years)

D

D

D

D

D

D

D

D

D

D

D D D

D D

D

D

D

D D

4) Specific, nonoptional procedures

D D

D

3) Theory development

D

2) Core category

D

1) Emergence and researcher distance

5) Coding procedures

(continued)

D

D D D

6) Evaluation criteria

Increasing the rigor of GT research 751

Table II. Evaluation of individual GT studies in SCM research from 2004 to 2009

Note: D refers to a discrepancy for a particular study in a specific dimension

D D

TJ Strauss JBL Glaser IJPDLM Glaser

D D

D

6) Evaluation criteria

Glaser

5) Coding procedures

TJ

D

4) Specific, nonoptional procedures

D

D

3) Theory development

Glaser D

2) Core category

JBL

D D

1) Emergence and researcher distance

D

Strauss

School of thought

IJPDLM Strauss JOM Strauss IJPDLM Glaser

JBL

Stefansson and Russel (2008) Rodney (2008) Vivek et al. (2008) Flint and Golicic (2009) Kaufmann et al. (2009) Mello and Hunt (2009) Russel et al. (2009) Voss et al. (2009) Whipple et al. (2009)

Table II.

Journal

752

Author(s) (years)

IJPDLM 42,8/9

Increasing the rigor of GT research 753

Notes: *Includes Glaser & Strauss and Glaser; **includes Strauss and Strauss & Corbin; ***a sum > 14 (>18) for papers following the Glaserian (Straussian) school of thought is possible because some papers deviate in more than one of the six dimensions

process is described in greater detail. Thus, SCM researchers must be aware of the differences between the two schools of thought and must make informed decisions about which school to choose (Mello and Flint, 2009). In addition, reviewers need to call for transparency in cases where no clear indication is given. Moreover, we do not find a single study which relied on Glaser’s approach, and which adhered to all tenets of this school of thought. In total, 85 percent ( ¼ 35/41) of all deviations in the examined period are identified in GT works following Glaser’s tradition. This frequency of occurrence and the discrepancies to methodological claims might be due to the more rigorous and inflexible research approach demanded by Glaser, combined with a lack of knowledge about the requirements of this approach. A closer examination of Figure 2 shows that the GT studies following Glaser accounted for the predominant number of deviations (96 percent ¼ 27/28) identified in the dimensions emergence and researcher distance, theory development, and core category. Further, we find a large proportion of deviations in the specific, non-optional procedures and evaluation criteria dimensions, independent of the underlying school of thought. To better understand the occurrences of these methodological discrepancies, we next analyzed each of the six dimensions in greater depth. 4.2.1 Deviations by dimension. First, deviations with regard to the dimension emergence and researcher distance occur throughout the entire investigated period for

Figure 2. Analysis of deviations from underlying schools of thought in GT works from 2004 to 2009

IJPDLM 42,8/9

studies following Glaser’s approach. Grounded theorists basing their work on this school of thought should not bring a priori knowledge to the research field: We do not know what we are looking for when we start [. . .] we simply cannot say prior to the collection and analysis of data what our study will look like (Glaser, 2001, p. 176).

754

Thus, the phenomenon of interest will gradually emerge from the data as the main concern of the participants under study. In other words, the emergence of theory is a key tenet of the Glaserian school of thought that cannot be forced or preconceived by the researcher. In contrast, studies deviating from this methodological tenet were mostly embedded in an extant research base that aimed at a contribution to the literature where only little was known prior to their GT analyses. Hence, grounded theorists following Glaser should not enter the field with pre-defined research questions, and should not have a detailed understanding of the literature in the investigated field (Glaser, 1992). Reviewers must also be aware of the differences between the two schools of thought to be able to judge the contexts under which a particular study took place. Otherwise, they will not be able to assess the true emergence of the theory. Second, deviations in the dimension core category refer to the deduction of no core category or of more than one. As described earlier, either of these research results violates the rigorous tenets of the Glaserian approach. Although Strauss’s less stringent dictum allows the use of GT for a non-theory purpose, similar to Glaser, Strauss advises the researcher to concentrate only on one core category when generating theory. In four cases, we find the emergence of more than one core category, which contradicts the emergence of one single dominant theme that concerns the participants involved in the study. Thus, grounded theorists extending or developing theory need to be aware of the abstract levels of their research findings, and may need to continue their analysis until they identify a single, dominant, core category. Here again, we challenge reviewers to push grounded theorists to the final level of abstraction when necessary. Third, while not as numerous, we also find deviations along the second dimension, theory development. The raison d’eˆtre of the Glaserian school is to develop a formal theory explaining the dominant theme that drives the phenomenon under study (Glaser, 2003). Put differently, if the researcher’s “ultimate research goal is to arrive at a set of findings rather than theory development,” (Strauss and Corbin, 1998, p. 155) she or he should base their research on the Straussian school of thought rather than follow Glaser. Strauss and Corbin (1998) not only allow GT in developing theory but also allow researchers to justify their research findings as being developed by a GT approach if they do not develop a formal theory. Reviewers unaware of these differences are unable to correctly judge whether the methodology served the purposes underlying a reviewed study. Nonetheless, considering the encouraged flexibility of choosing, rejecting, and ignoring single GT procedures, it is surprising to find any deviation with regard to this dimension in studies following the Straussian school of thought. In seven cases, we find that although the authors claimed to apply GT, their data analysis processes did not necessarily follow any GT procedure at all. In other words, grounded theorists in the SCM domain partly use GT as “encompassing any qualitative approach in which an inductive analysis is grounded in the data” ( Jones and Noble, 2007, p. 100). Mello and Flint (2009, p. 107) also found that:

[. . .] the vast majority of qualitative research in logistics relying on interviews, case studies and to some extent, observation, fails to articulate a specific tradition and philosophical perspective within which the studies were conducted.

They elaborate further: Even when grounded theorist methods are cited and followed, researchers tend to pick and choose what they want and simply assume that because they used interviews their research is grounded and as such constitutes grounded theory research (Mello and Flint, 2009, p. 109).

In essence, what we find are mixed methods approaches of interpretive research methodologies. In these cases, authors claimed to have employed a GT development approach suggested by Glaser and Strauss, but the data were collected, coded, and analyzed in a different fashion – for example concentrating on within-case descriptions and cross-case comparisons. Case descriptions seek to uncover the question of what happened, whereas GT aims at the understanding of complex, social processes that are relevant for the individuals involved (Gephart, 2004), and which are understandable to all relevant stakeholders (Glaser and Strauss, 1967; Mello and Flint, 2009). Moving forward, it is important for our field to avoid the impression that GT is being used “as rhetorical sleight of hand by authors who are unfamiliar with qualitative research” (Suddaby, 2006, p. 633). We therefore challenge grounded theorists in the SCM domain to adhere to previously described research procedures underlying the GT methodology. Researchers need to consciously choose the interpretive methodology with which to conduct their studies. And, in the event that a GT approach is appropriate, researchers must in turn follow GT research procedures and refrain from other interpretive analysis tools. Moreover, these findings suppose that even some reviewers might be unaware of the different methodological claims of different interpretive research methods. The same holds true for reviewers as they ensure the quality of published articles. Thus, we also call for editors and reviewers to push researchers to more adequately describe their methods and analysis. Fourth, we could not identify any deviation from the methodological tenets underlying the dimension coding procedures. Nonetheless, this seemingly perfect result might not necessarily be due to informed decision-making by the authors but rather by a lack of transparency. In 66 percent ( ¼ 21/32) of all investigated GT studies, the authors failed to describe coding procedures used in their data analysis. This could leave the reader with the impression that analytical procedures were not appropriately and rigorously conducted or that “authors [might] hold some serious misconceptions about grounded theory” (Suddaby, 2006, p. 633). A lack of transparency with regard to how research findings were derived can cater to the misconception that qualitative research methodologies are less valid and less rigorous compared to quantitative research traditions (Stuart et al., 2002). Grounded theorists need to outline how theory emerged from the evidentiary base of primary and secondary data. Otherwise, the knowledgeable editor, reviewer, or reader may assume that the term “grounded theory was interpreted to mean anything goes. Data may have been collected randomly, coded by forced application of preexisting conceptual categories” (Suddaby, 2006, p. 640). In other words, non-transparent research processes might be a result of having ignored or deliberately violated the core tenets and research procedures of GT (Locke, 1996). Thus, we call for grounded theorists in the field of SCM to invest more effort in

Increasing the rigor of GT research 755

IJPDLM 42,8/9

756

disclosing how their findings were derived. Additionally, we call for reviewers to demand the necessary transparency from GT studies. Fifth, we find the same – though less numerous – lack of transparency when analyzing the dimension evaluation criteria: 47 percent ( ¼ 15/32) of all reviewed GT articles did not outline how the quality of the research findings was assessed. In the SCM domain, we need to refrain from what Piekkari et al. (2010) identified to be common practice in the interpretive research community: “It is good enough to get published rather than good practice.” This again assumes that reviewers are adequately versed in the GT methodology. Editors need to ensure that reviewers of GT studies are well trained in this methodology and have the expertise to judge what constitutes good practice. Additionally, as described previously, different positions have emerged regarding the evaluative criteria for GT research. Thus, grounded theorists need to adhere to the quality criteria proposed by the school of thought they choose to follow. In summary, in more than half of all reviewed GT studies (59 percent ¼ 19/32) in the SCM domain, we find that foundational procedures and tenets of GT appear to be arbitrarily used without justification. While far from an excuse, this is consistent with findings in other disciplines (Jones and Noble, 2007). Regardless, our findings highlight the need for grounded theorists in the SCM domain to revisit the methodological tenets characterizing the differing schools of thought which comprise GT. 5. Conclusion and future directions As in all reviews, the findings and recommendations for future research are limited to the reviewed papers, which were published in a selection of top academic journals. We carefully selected our journals, relying on a commonly accepted selection of mainstream logistics, transportation, and SCM journals ranked by the Carter et al. (2009) studies. In addition, we included the Journal of Operations Management. This selection might limit the generalizability of our review findings. Thus, we call future researchers to include a broader selection of academic journals – for example including marketing and operations management journals – when reviewing researchers’ adherence to GT tenets. Based on the findings of the systematic review conducted in this paper, we present the following three prescriptions to grounded theorists in the SCM domain: first, we call for researchers to (re)visit the methodological roots of GT, to further increase trustworthiness, by clearly outlining and adhering to these epistemological origins. Our findings suggest that GT researchers often violate one or more of the analytic tenets of the methodology. We offer three explanations for this development: (1) Many researchers lack in-depth knowledge about the methodological origins out of which the different schools of thought have developed. (2) GT studies in the SCM literature are characterized by an inconsistent following of these schools of thought. This “may lie not so much with researcher ignorance but more with a state of confusion” ( Jones and Noble, 2007, p. 99). In either case, grounded theorists in the SCM domain need to go back to the methodological origins and the distinctive schools developed by the originators of GT. As GT is rooted in the sociological literature, we advise researchers in the SCM domain to better use inter-disciplinary sources to increase their methodological knowledge base.

(3) Apparently, reviewers also need to increase their knowledge about differing GT traditions. Editors must ensure that reviewers of GT studies are able to restore the integrity of this methodology and thereby eliminate the laxity and inconsistencies that currently appear to exist in the SCM domain. Grounded theorists need to remain true to the procedures of the chosen school of thought throughout the entire research process. Following one school of thought and communicating research findings accordingly will still permit grounded theorists enough room for determining how the analytical procedures proposed by their originators are performed. Second, we call grounded theorists in the field of SCM to increase transparency when describing their research. Researchers need to address all six of the previously discussed dimensions to reflect a deep knowledge about the applied methodology. This would help to decrease the perception by the research community that GT studies lack validity and objectivity (George and Bennett, 2004; Easton, 2010). A failure to present all phases of the research process leaves the reader with the impression that the differences between the two schools of thought are not thoroughly understood. Thus, reviewers must ensure that grounded theorists adequately describe their research purposes and processes. This will in turn uncover any methodological misconceptions held by the GT researcher. Finally, the GT approach should be utilized more in future SCM research, particularly given the preponderance of more positivist studies in the SCM discipline (Carter and Ellram, 2003; Choi and Wacker, 2011) and the dominance of a deductive theory paradigm in the broader social sciences (Ketchen and Hult, 2011; Rindova, 2011). Reviews of the growth of the SCM discipline (Kent and Flint, 1997; Frankel et al., 2005) have recognized the necessity of investigating the behavioral and social aspects of our field, and the need to develop our own theory as an academic discipline. However, as noted by Carter (2011, p. 3), while, “Theory defines a scientific discipline [. . .] the supply chain management discipline has largely failed to develop its own theoretical bases,” despite a gap of almost 15 years since Mello and Flint’s call for such theory development. This disparity offers unprecedented opportunities for GT application, along with conceptual theory development (Rindova, 2011; Skilton, 2011) as a separate but complementary technique. We hope that researchers, editors, and reviewers alike find our analyses helpful in better understanding the methodological tenets of GT and the divergence of schools of thought within this methodology. References Blair, E. and Zinkhan, G.M. (2006), “Nonresponse and generalizability in academic research”, Journal of the Academy of Marketing Science, Vol. 34 No. 1, pp. 4-7. Carter, C.R. (2011), “A call for theory: the maturation of the supply chain management discipline”, Journal of Supply Chain Management, Vol. 47 No. 2, pp. 3-7. Carter, C.R. and Dresner, M. (2001), “Purchasing’s role in environmental management: cross-functional development of grounded theory”, Supply Chain Management, Vol. 37 No. 3, pp. 12-27. Carter, C.R. and Ellram, L.M. (2003), “Thirty-five years of the Journal of Supply Chain Management: where have we been and where are we going?”, Journal of Supply Chain Management, Vol. 39 No. 1, pp. 38-50.

Increasing the rigor of GT research 757

IJPDLM 42,8/9

758

Carter, C.R. and Rogers, D.S. (2008), “Sustainable supply chain management: toward new theory in logistics management”, International Journal of Physical Distribution & Logistics Management, Vol. 38 No. 5, pp. 360-87. Carter, C.R., Easton, L.P., Vellenga, D.B. and Allen, B.J. (2009), “Affiliation of authors in transportation and logistics academic journals: a reevaluation”, Transportation Journal, Vol. 48 No. 1, pp. 42-52. Carter, C.R., Kaufmann, L., Beall, S., Carter, P.L., Hendrick, T.E. and Petersen, K.J. (2004), “Reverse auctions-grounded theory from the buyer and supplier perspective”, Transportation Research Part E: Logistics and Transportation Review, Vol. 40 No. 3, pp. 229-54. Choi, T.Y. and Wacker, J.G. (2011), “Theory building in the OM/SCM field: pointing to the future by looking at the past”, Journal of Supply Chain Management, Vol. 47 No. 2, pp. 8-11. Cova, B. and Elliott, R. (2008), “Everything you always wanted to know about interpretive consumer research but were afraid to ask”, Qualitative Market Research: An International Journal, Vol. 11 No. 2, pp. 121-9. Davis-Sramek, B. and Fugate, B.S. (2007), “State of logistics: a visionary perspective”, Journal of Business Logistics, Vol. 28 No. 2, pp. 1-34. Davis-Sramek, B., Fugate, B.S. and Omar, A. (2007), “Functional/dysfunctional supply chain exchanges”, International Journal of Physical Distribution & Logistics Management, Vol. 37 No. 1, pp. 43-63. Denyer, D. and Neely, A. (2004), “Introduction to special issue: innovation and productivity performance in the UK”, International Journal of Management Reviews, Vol. 5/6 Nos. 3/4, pp. 131-5. Easterby-Smith, M., Golden-Biddle, K. and Locke, K. (2008), “Working with pluralism: determining quality in qualitative research”, Organizational Research Methods, Vol. 11 No. 3, pp. 419-29. Easton, G. (2010), “Critical realism in case study research”, Industrial Marketing Management, Vol. 39 No. 1, pp. 118-28. Ellram, L.M. (2006), “The implementation of target costing in the United States: theory versus practice”, Journal of Supply Chain Management, Vol. 42 No. 1, pp. 13-27. Esper, T.L., Fugate, B.S. and Davis-Sramek, B. (2007), “Logistics learning capability: sustaining the competitive advantage gained through logistics leverage”, Journal of Business Logistics, Vol. 28 No. 2, pp. 57-81. Finch, J.H. (2002), “The role of grounded theory in developing economic theory”, Journal of Economic Methodology, Vol. 9 No. 2, pp. 213-34. Flint, D.J. and Golicic, S.L. (2009), “Searching for competitive advantage through sustainability”, International Journal of Physical Distribution & Logistics Management, Vol. 39 No. 10, pp. 841-60. Flint, D.J. and Mentzer, J.T. (2000), “Logiticians as marketers: their role when customers’ desired value changes”, Journal of Business Logistics, Vol. 21 No. 2, pp. 19-45. Flint, D.J., Larsson, E., Gammelgaard, B. and Mentzer, J.T. (2005), “Logistics innovation: a customer value-oriented social process”, Journal of Business Logistics, Vol. 26 No. 1, pp. 113-47. Frankel, R., Naslund, D. and Bolumole, Y. (2005), “The ‘white space’ of logistics research: a look at the role of methods usage”, Journal of Business Logistics, Vol. 26 No. 2, pp. 185-208. Fugate, B.S., Mentzer, J.T. and Flint, D.J. (2008), “The role of logistics in market orientation”, Journal of Business Logistics, Vol. 29 No. 2, pp. 1-26. Fugate, B.S., Sahin, F. and Mentzer, J.T. (2006), “Supply chain management coordination mechanisms”, Journal of Business Logistics, Vol. 27 No. 2, pp. 129-61.

George, A.L. and Bennett, A. (2004), Case Studies and Theory Development in the Social Sciences, MIT Press, Cambridge, MA. Gephart, R.P. (2004), “Qualitative research and the Academy of Management Journal”, Academy of Management Journal, Vol. 47 No. 4, pp. 454-62. Giunipero, L.C., Hooker, R.E., Joseph-Matthews, S., Yoon, T.E. and Brudvig, S. (2008), “A decade of SCM literature: past, present and future implications”, Journal of Supply Chain Management, Vol. 44 No. 4, pp. 66-86. Glaser, B.G. (1978), Theoretical Sensitivity, Sociology Press, Mill Valley, CA. Glaser, B.G. (1992), Basics of Grounded Theory Analysis: Emergence vs. Forcing, Sociology Press, Mill Valley, CA. Glaser, B.G. (1999), “The future of grounded theory”, Qualitative Health Research, Vol. 9 No. 6, pp. 836-45. Glaser, B.G. (2001), The Grounded Theory Perspective: Conceptualization Contrasted with Description, Sociology Press, Mill Valley, CA. Glaser, B.G. (2003), The Grounded Theory Perspective II: Description’s Remodeling of Grounded Theory Methodology, Sociology Press, Mill Valley, CA. Glaser, B.G. (2005), The Grounded Theory Perspective III: Theoretical Coding, Sociology Press, Mill Valley, CA. Glaser, B.G. and Strauss, A.L. (1967), The Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine de Gruyter, Hawthorne, NY. Goulding, C. (1998), “Grounded theory: the missing methodology on the interpretivist agenda”, Qualitative Market Research, Vol. 1 No. 1, pp. 50-7. Haytko, D.L., Kent, J.L. and Hausman, A. (2007), “Mexican maquiladoras: helping or hurting the US/Mexico cross-border supply chain?”, The International Journal of Logistics Management, Vol. 18 No. 3, pp. 347-63. Jones, R. and Noble, G. (2007), “Grounded theory and management research: a lack of integrity?”, Qualitative Research in Organizations and Management: An International Journal, Vol. 2 No. 2, pp. 84-103. Kaufmann, L., Michel, A. and Carter, C.R. (2009), “Debiasing strategies in supply management decision-making”, Journal of Business Logistics, Vol. 30 No. 1, pp. 85-106. Kent, J.L. and Flint, D.J. (1997), “Perspectives on the evolution of logistics thought”, Journal of Business Logistics, Vol. 18 No. 2, pp. 15-29. Kerlinger, F.M. (1986), Foundations of Behavioral Research, Holt, Rinehart and Winston, New York, NY. Ketchen, D.J. and Hult, G.T.M. (2011), “Building theory about supply chain management: some tools from the organizational sciences”, Journal of Supply Chain Management, Vol. 47 No. 2, pp. 12-18. Locke, K. (1996), “Rewriting the discovery of grounded theory after 25 years?”, Journal of Management Inquiry, Vol. 5 No. 3, pp. 239-45. Loesch, A. and Lambert, J. (2007), “E-reverse auctions revisited: an analysis of context, buyer-supplier relations and information behavior”, Journal of Supply Chain Management, Vol. 43 No. 4, p. 47. Manuj, I. and Mentzer, J.T. (2008), “Global supply chain risk management strategies”, International Journal of Physical Distribution & Logistics Management, Vol. 38 No. 3, pp. 192-223. Matos, S. and Hall, J. (2007), “Integrating sustainable development in the supply chain: the case of life cycle assessment in oil and gas and agricultural biotechnology”, Journal of Operations Management, Vol. 25 No. 6, pp. 1083-102.

Increasing the rigor of GT research 759

IJPDLM 42,8/9

760

Mello, J.E. and Flint, D.J. (2009), “A refined view of grounded theory and its application to logistics research”, Journal of Business Logistics, Vol. 30 No. 1, pp. 107-25. Mello, J.E. and Hunt, C.S. (2009), “Developing a theoretical framework for research into driver control practices in the trucking industry”, Transportation Journal, Vol. 48 No. 4, pp. 20-39. Mello, J.E., Stank, T.P. and Esper, T.L. (2008), “A model of logistics outsourcing strategy”, Transportation Journal, Vol. 47 No. 4, pp. 5-25. Miles, M.B. and Huberman, A.M. (1994), Qualitative Aata Analysis: An Expanded Sourcebook, Sage, Thousand Oaks, CA. Mollenkopf, D., Russo, I. and Frankel, R. (2007), “The returns management process in supply chain strategy”, International Journal of Physical Distribution & Logistics Management, Vol. 37 No. 7, pp. 568-92. Nilsson, F. (2006), “Logistics management in practice – towards theories of complex logistics”, The International Journal of Logistics Management, Vol. 17 No. 1, pp. 38-54. Pagell, M. (2004), “Understanding the factors that enable and inhibit the integration of operations, purchasing and logistics”, Journal of Operations Management, Vol. 22 No. 5, pp. 459-87. Pagell, M., Wu, Z. and Wasserman, M.E. (2010), “Thinking differently about purchasing portfolios: an assessment of sustainable sourcing”, Journal of Supply Chain Management, Vol. 46 No. 1, pp. 57-73. Paulraj, A. (2011), “Understanding the relationships between internal resources and capabilities, sustainable supply management and organizational sustainability”, Journal of Supply Chain Management, Vol. 47 No. 1, pp. 19-37. Perry, M. (2007), “Natural disaster management planning: a study of logistics managers responding to the tsunami”, International Journal of Physical Distribution & Logistics Management, Vol. 37 No. 5, pp. 409-33. Piekkari, R., Plakoyiannaki, E. and Welch, C. (2010), “‘Good’ case research in industrial marketing: insights from research practice”, Industrial Marketing Management, Vol. 39 No. 1, pp. 109-17. Reuter, C., Foerstl, K., Hartmann, E. and Blome, C. (2010), “Sustainable global supplier management: the role of dynamic capabilities in achieving competitive advantage”, Journal of Supply Chain Management, Vol. 46 No. 2, pp. 45-63. Rindova, V.P. (2011), “Moving from ideas to a theoretical contribution: comments on the process of developing theory in organizational research”, Journal of Supply Chain Management, Vol. 47 No. 2, pp. 19-21. Rodney, T. (2008), “Exploring relational aspects of time-based competition”, International Journal of Physical Distribution & Logistics Management, Vol. 38 No. 7, pp. 540-50. Russell, D.M., Ruamsook, K. and Thomchick, E.A. (2009), “Ethanol and the petroleum supply chain of the future: five strategic priorities of integration”, Transportation Journal, Vol. 48 No. 1, pp. 5-22. Sawhney, R. (2006), “Interplay between uncertainty and flexibility across the value-chain: towards a transformation model of manufacturing flexibility”, Journal of Operations Management, Vol. 24 No. 5, pp. 476-93. Shao, X. and Ji, J. (2006), “Reconfiguration of pharmaceutical logistics operations in China: an empirical study”, Transportation Journal, Vol. 45 No. 4, pp. 52-66. Skilton, P.F. (2011), “Getting the reader to ‘I get it’: clarification, differentiation and illustration”, Journal of Supply Chain Management, Vol. 47 No. 2, pp. 22-8. Sousa, C.A.A. and Hendriks, P.H.J. (2006), “The diving bell and the butterfly: the need for grounded theory in developing a knowledge-based view of organizations”, Organizational Research Methods, Vol. 9 No. 3, pp. 315-38.

Stefansson, G. and Russell, D.M. (2008), “Supply chain interfaces: defining attributes and attribute values for collaborative logistics management”, Journal of Business Logistics, Vol. 29 No. 1, pp. 347-60. Stern, P.N. (1994), “Eroding grounded theory”, in Morse, J.M. (Ed.), Critical Issues in Qualitative Research Methods, Sage, Thousand Oaks, CA, pp. 212-23. Strauss, A.L. (1987), Qualitative Analysis for Social Scientists, Cambridge University Press, Cambridge. Strauss, A.L. and Corbin, J. (1990), Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, Sage, Thousand Oaks, CA. Strauss, A.L. and Corbin, J. (1992), Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, Sage, Thousand Oaks, CA. Strauss, A.L. and Corbin, J. (1994), “Grounded theory methodology: an overview”, in Denzin, N.K. and Lincoln, Y.S. (Eds), Handbook of Qualitative Research, Sage, Thousand Oaks, CA, pp. 273-85. Strauss, A.L. and Corbin, J. (1996), Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, Sage, Thousand Oaks, CA. Strauss, A.L. and Corbin, J. (1998), Basics of Qualitative Research, Sage, Thousand Oaks, CA. Stuart, I., McCutcheon, D., Handfield, R., McLachlin, R. and Samson, D. (2002), “Effective case research in operations management: a process perspective”, Journal of Operations Management, Vol. 20 No. 5, pp. 419-33. Suddaby, R. (2006), “From the editors: what grounded theory is not”, Academy of Management Journal, Vol. 49 No. 4, pp. 633-42. Tate, W.L., Ellram, L.M. and Kirchoff, J.F. (2010), “Corporate social responsibility reports: a thematic analysis related to supply chain management”, Journal of Supply Chain Management, Vol. 46 No. 1, pp. 19-44. Theoharakis, V., Voss, C., Hadjinicola, G.C. and Soteriou, A.C. (2007), “Insights into factors affecting Production and Operations Management (POM) journal evaluation”, Journal of Operations Management, Vol. 25 No. 4, pp. 932-55. Tucker, A.L. (2004), “The impact of operational failures on hospital nurses and their patients”, Journal of Operations Management, Vol. 22 No. 2, pp. 151-69. Vivek, S.D., Banwet, D.K. and Shankar, R. (2008), “Analysis of interactions among core, transaction and relationship-specific investments: the case of offshoring”, Journal of Operations Management, Vol. 26 No. 2, pp. 180-97. Voss, M.D., Closs, D.J., Calantone, R.J., Helferich, O.K. and Speier, C. (2009), “The role of security in the food supplier selection decision”, Journal of Business Logistics, Vol. 30 No. 1, pp. 127-55. Wagner, S.M., Lukassen, P. and Mahlendorf, M. (2010), “Misused and missed use – grounded theory and objective hermeneutics as methods for research in industrial marketing”, Industrial Marketing Management, Vol. 39, pp. 5-15. Walker, D. and Myrick, F. (2006), “Grounded theory: an exploration of process and procedure”, Qualitative Health Research, Vol. 16 No. 4, pp. 547-59. Whipple, J.M. and Russell, D.M. (2007), “Building supply chain collaboration: a typology of collaborative approaches”, The International Journal of Logistics Management, Vol. 18 No. 2, pp. 174-96. Whipple, J.M., Voss, M.D. and Closs, D.J. (2009), “Supply chain security practices in the food industry”, International Journal of Physical Distribution & Logistics Management, Vol. 39 No. 7, pp. 574-94. Wu, Z. and Choi, T.Y. (2005), “Supplier-supplier relationships in the buyer-supplier triad: building theories from eight case studies”, Journal of Operations Management, Vol. 24 No. 1, pp. 27-52.

Increasing the rigor of GT research 761

IJPDLM 42,8/9

762

Yeung, J.H.Y., Selen, W., Deming, Z. and Min, Z. (2007), “Postponement strategy from a supply chain perspective: cases from China”, International Journal of Physical Distribution & Logistics Management, Vol. 37 No. 4, pp. 331-56. Yin, R.K. (2003), Case Study Research: Design and Methods, Sage, Thousand Oaks, CA. Zsidisin, G.A., Ellram, L.M., Carter, J.R. and Cavinato, J.L. (2004), “An analysis of supply risk assessment techniques”, International Journal of Physical Distribution & Logistics Management, Vol. 34 No. 5, pp. 397-413. Further reading Carter, C.R., Leuschner, R. and Rogers, D.S. (2007), “A social network analysis of the Journal of Supply Chain Management: knowledge generation, knowledge diffusion and thought leadership”, Journal of Supply Chain Management, Vol. 43 No. 2, pp. 15-28. Kaufmann, L. and Denk, N. (2011), “How to demonstrate rigor when presenting grounded theory research in the supply chain management literature”, Journal of Supply Chain Management, Vol. 47 No. 4, pp. 64-72. Parry, J. (2003), “Making sense of executive sensemaking: a phenomenological case study with methodological criticism”, Journal of Health Organization and Management, Vol. 17 No. 4, pp. 240-63. Schreiber, R.S. and Stern, P.N. (2001), Using Grounded Theory in Nursing, Springer, New York, NY. About the authors Nikola Denk is a Research Associate at WHU – Otto Beisheim School of Management in Vallendar, Germany. Her primary research streams focus on qualitative research methods and the internationalization of emerging market firms. Lutz Kaufmann is a Professor and holds the Chair of International Business and Supply Management at WHU – Otto Beisheim School of Management in Vallendar, Germany. He is also an Associate Fellow of Said Business School of Oxford University, UK. His research focuses on supply management and international strategy. He is the European Editor of the Journal of Supply Chain Management. His work has appeared, amongst others, in the Journal of Operations Management, International Journal of Physical Distribution & Logistics Management, Journal of Business Ethics, Journal of World Business, Journal of International Marketing, Management International Review, Journal of Business Logistics, Journal of Supply Chain Management, and The Journal of Purchasing & Supply Management. Lutz Kaufmann is the corresponding author and can be contacted at: [email protected] Craig R. Carter is a Professor of Supply Chain Management at Arizona State University. His primary research stream focuses on the sustainable management of the supply chain. This research stream encompasses ethical issues in buyer-supplier relationships, environmental supply management, diversity sourcing, perceptions of opportunism surrounding electronic reverse auctions, and the broader, integrative concepts of social responsibility and sustainability. A secondary and often intersecting area of research examines international and cross-cultural supply chain management issues. He is a member of the editorial review boards of several journals, and the Co-Editor-in-Chief of the Journal of Supply Chain Management. His research has appeared in numerous supply chain management journals, including Decision Sciences, International Journal of Physical Distribution & Logistics Management, Journal of Business Logistics, Journal of Operations Management, Journal of Supply Chain Management, Transportation Journal, and Transportation Research E. To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Evaluation criteria

Coding

Specific, non-optional procedures

Theory development

Core category

Emergence and researcher distance

Researchers proactively use their knowledge and experience as a basis for their analyses Known theories are compared to the collected data and emerging theoretical insights

Straussian school

Data analyses must not necessarily result in a core category (especially if the research objective is not theory development) No more than one core category should be derived The research objective is either theory development or other purposes, such as conceptual ordering or detailed descriptions of a phenomenon Researchers must follow all procedures rigidly if the Researchers might chose, reject, and ignore single research findings should be recognized as GT procedures depending on what best fits to the research objective Researchers need to follow open, selective, and theoretical Researchers need to follow open, axial, and selective coding procedures coding procedures Researchers must adhere to the quality criteria of: Researchers must adhere to the quality criteria of: fit credibility workability transferability relevance dependability modifiability confirmability integrity fit understanding generality control

Researchers are distant as they approach the data. Either they have: (a) no a priori knowledge when investigating a phenomenon, or (b) they use a priori knowledge only to better understand collected data Data analyses must result in a core category that summarizes main concerns and thus, the behavior of participants No more than one core category should be derived The research objective is to develop theory

Glaserian school

Appendix

Increasing the rigor of GT research 763

Table AI. Differences between the Glaserian and Straussian schools of thought (based on Jones and Noble, 2007, p. 89)

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0960-0035.htm

IJPDLM 42,8/9

Moving beyond the systems approach in SCM and logistics research

764 Received 4 December 2011 Accepted 15 December 2012

Fredrik Nilsson Department of Design Sciences, Division of Packaging Logistics, Lund University, Lund, Sweden, and

Britta Gammelgaard Department of Operations Management, Copenhagen Business School, Frederiksberg, Denmark Abstract Purpose – The purpose of this paper is to provide a paradigmatic reflection on theoretical approaches recently identified in logistics and supply chain management (SCM); namely complex adaptive systems and complexity thinking, and to compare it to the dominant approach in logistics and SCM research, namely the systems approach. By analyzing the basic assumptions of the three approaches, SCM and logistics researchers are guided in their choice of research approaches which increases their awareness of the consequences different approaches have on theory and practice. Design/methodology/approach – The point of departure for the research presented is conceptualization based on literature reviews. Furthermore, years of observations, discussions and empirical studies of logistics operations and management have also influenced the design of this research. Findings – With a discourse set in relation to the dominant approach in SCM and logistics research, the systems approach, it is concluded that the underlying assumptions of complex adaptive systems and complexity thinking are more appropriate than systems approach for contemporary challenges of organizational complexity in SCM and logistics. It is found that the two complexity-based approaches can advance SCM and logistics research and practice especially when focusing on innovation, learning and sense-making. Research limitations/implications – Reflections of underlying assumptions when considering and selecting methodological approaches have implications for research results. This paper provides both a framework for and an analysis of such reflection which contributes to the further development of SCM and logistics research. Future research is needed to empirically provide insights on how complexity approaches can advance the area of SCM and logistics. Practical implications – For logistics researchers and practitioners dealing with creativity, innovation, learning and sense-making and other human-related aspects, the complexity approaches, with underlying assumptions, presented will provide reflection, inspiration and guidance for further development. Originality/value – This paper contributes to the further development of SCM and logistics research and practice by providing a reflective analysis and discussion of established and new research approaches with potential benefits for the SCM and logistics community.

International Journal of Physical Distribution & Logistics Management Vol. 42 No. 8/9, 2012 pp. 764-783 q Emerald Group Publishing Limited 0960-0035 DOI 10.1108/09600031211269749

Keywords Complexity theory, Basic assumptions, Paradigm, Logistics, Supply chain management, Systems approach, Distribution management Paper type Research paper

Introduction Supply chain management (SCM) and logistics are receiving increased attention in companies and in research. SCM is gaining strategic importance as supply chains are

increasingly perceived as a source of competitive advantage (Skjøtt-Larsen et al., 2007). While historically a sub-discipline in marketing or operations management (Kent and Flint, 1997) with its main focus on physical handling and transportation operations, the SCM discipline today encompasses the collaboration and integration of interorganizational processes, creation of customer value and innovation – all this in a global context where cultural differences are a fact. Given the trends of shortened product life cycles and decreasing time to market, together with challenges such as innovation, sustainability, and other societal changes, the complexity confronted by both managers and researchers will increase. One emerging theme to handle these challenges, and one that has been highlighted in SCM and logistics research, is the application of models and frameworks derived from complexity theory. For example, Ellram et al. (2007) use complexity theory in their study of concurrent development of product, process, and supply chain. Choi et al. (2001, p. 352) conclude that it is “not enough to recognize a supply network as simply a system – a supply network is a complex adaptive system (CAS),” something also Wysick et al. (2008) and Surana et al. (2005) conclude. The works by Johannessen (2003), Johannessen and Solem (2002), Nilsson and Darley (2006), Nilsson (2006), Holweg and Pil (2008) and Svensson (2010), are also examples of complexity theory applied to SCM and logistics. According to Simon (1996), complexity theory represents a third phase in the study of complex systems. The initial phase ended up in holism and gestalt, while the second established cybernetics and general systems theory (Simon, 1996). In regard to the third phase, MacIntosh and MacLean (2001, p. 1345) state that “the development of complexity theory [. . .] is regarded by some as signaling the arrival of a new scientific paradigm in the Kuhnian sense.” Consequently, there are a growing number of researchers applying complexity theories and approaches with the conclusion that they create increased understanding of the complex phenomena that supply chains or networks represent. However, for the future development of the SCM and logistics discipline, reflections of ontological, epistemological, and teleological assumptions ought to be made. This is especially true since SCM and logistics research has a history of being strongly influenced by positivism (Mentzer and Kahn, 1995; Gammelgaard, 2004; Nilsson, 2003; Golicic et al., 2005), i.e. influenced by assumptions such as rationality, stability, objectivity, linearity, determinism, value-freeness, designability, and controllability. Lewis and Suchan (2003, p. 312) state that: [. . .] examining the language in the published research [. . .], it becomes clear that the dominant or root metaphor of logistics researchers is the organization as a machine and its members as rational actors.

Arlbjørn and Halldorsson (2002, p. 22) question the dominant positivistic paradigm in logistics research by stating: “if we take this view for granted, we may produce a unilateral view of logistics knowledge that only focuses on objective and observable phenomena.” Trim and Lee (2004, p. 473) state that “management researchers need to have the confidence to challenge basic assumptions relating to interpreting research outcomes, and what constitutes appropriate research.” The question is then what set of assumptions are appropriate when approaching the complex and challenging problems and issues in today’s SCM and logistics? Furthermore, what are the consequences on research processes and outcomes of applying methodological approaches based on complexity theory compared to the dominant approaches in SCM and logistics today?

Beyond the systems approach 765

IJPDLM 42,8/9

766

The aim of this paper is to provide a paradigmatic reflection on two approaches identified in complexity theory – CASs and complexity thinking (CT) – and on how they relate to the dominant approach to SCM and logistics research, namely, the systems approach (SA). By analyzing the basic assumptions of these three approaches, SCM and logistics researchers are guided in their choice of research approaches, and increase their awareness of the consequences different approaches have on theory and practice. Consequently, the purpose of such a discussion is: . to provide the SCM and logistics research society with a paradigmatic reflection on newly applied research approaches in order to address and solve new types of questions and old questions with new approaches; and . to contribute to the paradigmatic discourse in logistics research and practice, thereby strengthening the SCM and logistics discipline. The reasons for choosing the SA as the basis of this discourse are twofold. First of all, the SA is fundamental for both researchers and practitioners in the field of SCM and logistics. As many researchers argue, to understand SCM and logistics one needs to understand the concept of systems in order to get a holistic perspective, to value total cost, and avoid suboptimizing (Grant et al., 2005; Bechtel and Jayaram, 1997). However, the influence of positivism when adopting a SA to SCM and logistics research is prevalent. Second, since complexity theory is regarded by several researchers as a development of systems theory it might also become the “upgraded” version of systems theory in SCM and logistics. The remaining paper is organized as follows. Second section will present and discuss the use of SA in SCM and logistics. In third section, two approaches originating from complexity theory will be introduced: CASs and CT. Fourth section provides a comparative analysis of the underlying assumptions related to the three approaches, starting off with a framework of assumptions. Finally, a conclusion discusses implications for SCM and logistics research and practice. The SA to SCM and logistics For both researchers and practitioners in the field of SCM and logistics the SA is fundamental (Lambert et al., 1998b; Bowersox and Closs, 1996; Stock et al., 1999). Nonetheless, literature that either explicitly states that a system approach is used or uses the system concept to illustrate research areas or components, seldom reports on what grounds, theories, and definitions the system approach is derived from. And literature that provides such grounds (Andersson, 1997; Ballou, 2004; Coyle et al., 2003; Holmberg, 2000; Lambert et al., 1998a, b; Mears-Young and Jackson, 1997; Saghir, 2004; Vafidis, 2007; Waidringer, 2001) are most often referred back to early works in systems theory. Most publications refer to general systems theory (von Bertalanffy, 1969), cybernetics (Ashby, 1956; Beer, 1959), and system dynamics (Forrester, 1968), often referred to as the hard SA. Several publications also refer to Checkland (1993), often referred to as the soft SA, and to Senge (1990) as well as Arbnor and Bjerke (1997). Furthermore, while the assumptions differ ontologically, epistemologically and teleologically among systems approaches, when applied to logistics the treatment of SA often falls back to positivistically influenced assumptions (Vafidis, 2007). Gubi et al. (2003) in their review of Scandinavian doctoral dissertations 1990 to 2001 concluded that as many as 45 percent have not explicitly incorporated methods or theories originating from the philosophy or theory of science. Mears-Young and Jackson (1997), conclude that it might be

useful and beneficial for logistics researchers to be more reflective about the foundations for their methods and solutions. Consequently, underlying assumptions are seldom reflected on when the SA is used in logistics research, education, or practice. A reason for this is suggested by Kuhn (1996, p. 46):

Beyond the systems approach

Scientists work from models acquired through education and through subsequent exposure to the literature often without quite knowing or needing to know what characteristics have given these models the status of community paradigm.

767

Lewis and Suchan (2003, p. 312) state that: [. . .] logistics research may need a growth-inducing language about logistics managerial behavior, a new set of theoretical constructs that will help researchers reframe and thus “resee” their thinking and research designs in ways that help them break from variance theories and the larger positivist paradigm.

The consequences of the positivist assumptions underlying the SA on research design and results in SCM and logistics are several. First of all, the assumptions breed mechanical metaphors and views of logistics phenomena in which people involved are assumed to behave rationally. The systems studied should be as efficient – or optimized – as possible. Mentzer et al. (2004, p. 616) propose, for instance, that “logistics information management capabilities meet the supply chain operational and strategic information needs [. . .], which leads to optimization of system-wide capital investment.” This optimized situation is reached when all uncertainty is eliminated and the systems’ constituent parts are simplified. Lambert and Cooper (2000, p. 72) state: “controlling uncertainty in customer demand, manufacturing processes, and supplier performance are critical to effective SCM.” Reality is objective and research is value-free and context-independent (Sachan and Datta, 2005), i.e. normative models created in Western cultures should work anywhere (Monczka et al., 1998; Hunt, 1981). Despite the holistic and integrative assumptions in systems theory, logistics phenomena under study can be fragmentized into units and then treated as separate entities, i.e. reductionism (Nilsson, 2005). Consequently, the researcher or manager has the ability to define a system by establishing its borders and determining its goals. Bowersox and Closs (1996, p. 459), embody this goal-setting ability by arguing that use of the systems concept in logistics stresses “total integrated effort toward the accomplishment of predetermined objectives.” Finally, normative models and sequential process frameworks are often desirable results of research activities Childerhouse and Towill, 2003; Towill, 1999). However, as Powell (2003, p. 286) argues, the avoidance of paradigmatic reflections might be perilous; “for any empirical discipline, epistemological beliefs have theoretical and methodological consequences, and habitual beliefs can lead to dogmatism, illusion, or despair.” Consequently, despite the overwhelming amount of SCM and logistics research based on the SA, due to underlying assumptions there are some aspects which may be reconsidered when research is conducted. This is especially the case if the research object includes or is influenced by people, something often excluded in logistics research (Tokar, 2010). Such aspects would, for example, concern collaboration, power, conflicts, creativity, novelty, culture, innovation, and sustainability. Rigby et al. (2000) state that: [. . .] the usual focus for improvement in the supply chain has been the optimization of a particular company’s inventory or scheduling protocols: [. . .] These systems analysis and redesign methods are excellent at describing and modeling the physical flow of materials,

IJPDLM 42,8/9

inventory data and demand patterns. However, the human-to-human interaction in network forms of organization has a much higher degree of complexity (Rigby et al. (2000, p. 181).

Nilsson (2006) concludes that based on the challenges logisticians are facing in their daily work, models and theories that can handle complex situations and contexts are needed. He states:

768

These challenges are characterized by novelty (the type of problems are contemporary), and paradoxes which are of an “unsolvable character” and can only be handled by balancing efforts each and every day (Nilsson, 2006, p. 51).

Hopper and Powell (1985, p. 429) state: “there is no such thing as a totally objective and value free investigation.” Furthermore, a number of researchers report that culture has impact on values and norms (Hofstede, 1991; Schwartz, 1994), e.g. “individualist and collectivist cultures produce distinctly different normative orientations toward establishing and maintaining relationships” (Cannon et al., 2010, p. 508). Finally, Gammelgaard (1997, p. 17) declares that: [. . .] an objective world view will not be able to bring forward the subjective pictures of the world as those, e.g. caused by difference in power between the actors in the system. Consequently, the perception of reality becomes insufficient if systems theory is the only approach in logistics.

However, before any deeper analysis of SA in comparison with the complexity approaches, it is necessary to provide a description of CASs and CT. Complexity approaches The emergence of the complexity theory is – as explained by most researchers in the field – first and foremost an attempt to move science away from the strong emphasis on reductionism and positivism that characterize the majority of scientific disciplines today. The argument is that reductionist and positivist assumptions restrain further progress and cannot explain empirical phenomena easily found in nature and social life. Nonetheless, within the complexity theory there are branches that diverge into different directions as a result of presumed assumptions. One branch, often referred to as CASs, can be regarded as a further development of SA on account of similar terminology and shared assumptions, including the notion of systems, boundaries, and formative teleology. However, as Phelan (1999, p. 237) states: [. . .] a common terminology suggests a high degree of commensurability between [systems theory and complexity theory]. However, on closer examination, although they share a common worldview, the two theories differ markedly in their research agenda and methodologies.

Other branches of the science of complexity, which do not share the assumptions of the first type, clearly set out to present other perspectives. Richardson and Cilliers (2001) indicate that there are different types of complexity theories and Stacey et al. (2000) highlight the different use of complexity theory in literature and practice. Stacey et al. (2000) even question whether the complexity movement is a fad or a radical challenge to the SA. For the purpose of this paper, the two branches that will be described and discussed in the context of logistics are CASs and CT. Examples of logistics-related publications reporting on the usage of CAS are Wysick et al. (2008), Ellram et al. (2007), Choi et al. (2001, p. 352), Nilsson (2003), Surana et al. (2005), and Nilsson and Darley (2006).

The works of Johannessen and Solem (2002), Johannessen (2003, 2005) and Nilsson (2007) are examples of CT in logistics. The distinction between CAS and CT is important because the paradigms that guide researchers are different. It is important to mention at this stage that this discussion does not aim to find a true or best approach for logistics research and practice. Instead, the proposed value lies in gaining understanding of what approach provides appropriate guidance, for the phenomenon to be studied. This is the case since many of the issues raised in logistics practice, concerning innovation, creativity, collaboration, and other directly human-related aspects, will not be addressed or handled in the best way possible without appropriate assumptions for such phenomena. As will be further elaborated, it is the authors’ belief that a multi-paradigmatic approach is the best way to approach any research problem. Consequently, the dilemma is to know when, and under what circumstances, to use what approach. Complex adaptive systems CAS can be described as: [. . .] a special kind of complex systems since they have the property of adaptation. [. . .], adaptation means that the agents or elements in the system are responsive, flexible, reactive and often proactive regarding inputs from other agents or elements that affect them (Nilsson, 2003).

CAS consist of several parts which are commonly referred to as agents, and which act in correlation and interdependence to each other (Kauffman, 1995). In the context of logistics these agents could be processes and activities, but on a lower level they could also be the machine operators, the truck drivers, i.e. the people within organizations, and even artifacts like machines and packages. This means that some agents might have greater influence on the system, and some less, but no one controls the whole system. Compared to the brain, there is no master neuron controlling what we think. The complexity arises in the adaptive, self-organizing processes among the agents from which perpetual novelty emerges. Attempts to reduce organizational complexity in order to control are often counterproductive (Colbert, 2004). CAS act most creatively in states far from equilibrium, often referred as to the edge of chaos, for at the other extreme “equilibrium is a precursor to death” (Pascale, 1999, p. 85). In CASs assumptions of determinism can still be revealed since there is an underlying belief that identifiable rules of cause and effect can be found in the system, i.e. a few simple rules that can be decided and elaborated on in order to realize predetermined outcomes for a whole system. This notion of designable and controllable systems through the use of simple rules has led to several normative management writings implicitly suggesting that the science of complexity opens new doors for control and prediction (Beinhocker, 1999; Kupers, 2001). Complexity thinking CT differs ontologically, epistemologically, and teleologically from both SA and CAS. Ontologically, the underlying belief is that of unorder (Snowden, 2002) and subjectivity; epistemologically, of heuristics or anti-positivism; and teleologically, of a transformative nature. CT is, in the authors’ point of view, the collection of the works by Stacey (2001) and Stacey et al. (2000) on complex responsive processes, Snowden (2002) and Snowden and Stanbridge (2004) Cynefin framework, and the works by Richardson et al. (2001). A major differentiator of CT in comparison to SA and CAS is that CT is developed for human- and human organizational phenomena, i.e. there are no extrapolations from

Beyond the systems approach 769

IJPDLM 42,8/9

770

physical, chemical or biological sciences. This is in line with the consideration Boulding (1956) makes in his hierarchy of complexity on nine levels for theoretical discourse, where on the seventh and eight levels, human and human organizations and societies are placed. The complex responsive process perspective sets out to explain transformational change as emerging from self-organizing processes which are born out of human interaction which cannot be observed objectively (Stacey, 2001). Based on transformational teleology the future in CT is regarded as mainly unknown, or, as Prigogine (1997) states, under “perpetual construction.” It follows from this that epistemological assumptions are in line with the limitations of handling or even understanding perceived reality. Focus is on exploratory analysis of the phenomena being studied and these complex phenomena are incompressible in CT (Richardson et al., 2001). As Kurtz and Snowden (2003, p. 480) point out: “Conceivability is not the point: preparation for the unexpected is.” Paradoxes are seen as natural in CT and follow Hegel’s logic in that they possess different characteristics and explanations at the same time, i.e. phenomena being predictably unpredictable (Stacey, 2002), or knowledge characterized as both a substance and a flow, simultaneously (Snowden, 2002). By the acceptance of paradoxes in organizational contexts, the CT approach involves considerations of central aspects in human life such as conflicts, power, creativity, novelty, joy, and love, where these aspects are central to the way human beings relate to each other and these qualities help create meaning in most situations. A comparative analysis of assumptions Ontological, epistemological, and teleological assumptions are prerequisites for the methodological assumptions and choices made in any discipline (Burrel and Morgan, 1979; Morgan, 1983; Snowden, 2002). For example, Guba and Lincoln (1998, p. 195) state: “questions of methods are secondary to questions of paradigm.” Consequently, starting out from a paradigmatic discussion about assumptions has direct implications on the approaches, methods, and tools to be used when dealing with SCM and logistics phenomena. Examples of assumptions associated with positivism (i.e. most systems approaches) are simplicity, order, objective reality, reductionism, deliberate design, rationality, and determinism. Those associated with complexity perspectives (i.e. CAS and CT) are complexity, unorder, intersubjective/subjective reality, emergence, self-organization, coevolution, bounded rationality, and indeterminism (Nilsson, 2003; Dent, 1999). It is important to point out that a complexity perspective considers all assumptions including the positivist. Hence, out of a complexity perspective it is not a question of duality where, for example, simplicity and complexity are extremes on an axis, but instead one based on an underlying assumption that both states can exist simultaneously. The underlying nature is dialectic when a complexity perspective is adopted. In this regard, Richardson and Cilliers (2001, p. 11) state that: [. . .] insights from the human sciences on the one hand, and natural science on the other, should not be set against each other, nor should they be assimilated too easily. They should be used to challenge each other.

To provide a fruitful analysis of assumptions and their treatment in SA, CAS, and CT, a categorization of assumptions will be provided as a starting point. The categories are: . structural; . behavioral; and . time-related.

The structural assumptions, i.e. assumptions regarding how structural aspects of logistics phenomena are formed and connected, are simplicity/complexity, order/unorder, reductionism/emergence, and objectivity/subjectivity. Assumptions categorized as behavioral are deliberate design/self-organization, coevolution, and rationality/bounded rationality, since these all relate to the creatures involved, both individually and collectively, and to how they interact. Finally, assumptions regarding determinism/indeterminism will be seen as time-related assumptions since they relate to future states or conditions. It is important to point out that this categorization is not a claim to be complete or inclusive since some of the assumptions are related to both time, behavior, and structure, e.g. coevolution. Instead the purpose is, from a logistics point of view, to bundle them together in such a way that could benefit further development of the SCM and logistics discipline. In other words, the reader may, from his/her subjective view, regard this characterization differently, and is encouraged to. As a result, the important message is for the logistician to rethink the assumptions that guide him or her. This categorization is beneficial to that process.

Beyond the systems approach 771

Structural assumptions One apparent similarity between SA and CAS is the assumption of an objectively accessible reality and that a chosen system can be separated from its environment (Table I). However, how this separation is done differs between the two. In SA, Assumption

SA

CAS

CT

Complexity is the result of Complexity and simplicity interacting agents or parts coexist, and relate to the perceptions of the actors that follow simple rules involved. It is in the relating processes that complexity appears as well as disappears Order is the natural result Paradoxical situations Order/unorder Order is hidden in every situation. Order is achieved of non-linear interactions. involving both order and Behavior creates structure disorder, simultaneously, by holistic control and creating unorder. Behavior and structure is an deliberate design so that “structure drives behavior” emergent outcome of self- creates the perception of structure, and perceived organizing processes structure affects the perceptions of both structures and behaviors. Cultural context matters Through emergent Reductionism/ Systems are mostly treated Systems are regarded as dynamic processes formed outcomes novelty is emergence as static descriptions, created which cannot be from the bottom-up, i.e. focus on “being”, and i.e. focusing on “becoming” taken apart or controlled, often separated into only post-rationalized solvable parts Reality is a result of Things can be assessed Objectivity/ Systems are observable subjective experiences and objectively but are very subjectivity from the outside where views that are created in the limited due to the both researchers and process of human relating, assumption of their managers can position causing sometimes dynamic nature themselves intersubjective views

Simplicity/ complexity

Complexity is a cause of how many parts and interactions are present, i.e. it can, and should, be simplified in order to be dealt with

Table I. A comparison of structural-related assumptions in SA, CAS, and CT

IJPDLM 42,8/9

772

systems are observable from the outside where both researchers and managers can position themselves. In CAS systems can be assessed objectively but are very limited due to the assumption of their dynamic nature, and it is especially problematic if the objectivity is to be used for intervention. The situation is different in CT. In CT the assumption is of a more interpretive character, where meaning and value are created in the interactions among people in their daily activities. Some common view of some characteristics of life is thus contextually created and – formally or informally – agreed on by those involved, creating some kind of intersubjectivity. However, due to people’s multiple identities, paradoxical and time-related aspects might change such agreements, as contexts perpetually change. This is due to the inherent complexity perceived in making commitments to several characteristics and situations simultaneously. In this regard Stacey et al. (2000, p. 61) state that “systems work, to the extent that they do, because of the informal, freely chosen, ordinary, day-to-day cooperative interactions of an organization’s members, and this cannot be controlled.”As reported in Nilsson (2006) the challenges logistics managers are concerned about are related to the sense making of situations in their organizations which calls for intersubjective perspectives and not objective realities. A central emphasis in SA is the focus on structure, and on how essential the structure is in controlling and directing the dynamics in a system. This point is especially stressed by Sterman (2000) and Senge (1990), who argue that “structure drives behaviour.” In this regard Lee et al. (1997, p. 548) conclude that companies that aim to control the bullwhip effect should attack “the institutional and inter-organizational infrastructure and related processes.” While, Disney et al. (1997, p. 176) declare that poorly designed order handling systems may cause amplifying behavior through the supply chain and thus, “it is essential to select the appropriate structure for the production ordering system, and then to set the system parameters at their ‘best’ value.” Hence, a hidden order or structure exists, which can be selected, designed and controlled by someone in a holistic, outside position such as top management, for instance, to direct behavior in lower hierarchies. Furthermore, as Johannessen (2003, p. 15) states; “it is assumed that the system is stable for such a time period that analysis can be made [. . .] a system should be able to display predictable and stable behavior.” In both CAS and CT one would agree with the SA assumption that “structure drives behavior” (CAS) or at least influences it (CT). However, the complexity researcher would also argue that “behavior creates structure” or at least, “behavior creates the perception of structure” and emphasizes the creation even more. The “structure” or order in CAS and CT is an emergent outcome of self-organizing behaviors of the constituent elements, where the elements are heterogeneous in both substance and action, resulting in a natural variety in all phenomena. This variety is the true source for both novelty (Stacey, 2002) and innovativeness (Allen, 2000) as well as survival (Kauffman, 1995). Cross-cultural research emphasizes the apparent heterogeneity of assumptions in supply chains. For example, the way trust is created in an individualized versus a collectivized culture (Cannon et al., 2010; Zhao et al., 2007) has a direct impact on the perceived order that emerges in a relationship. Kauffman (1995) calls this self-organizing outcome “order for free” as temporal stability emerges as a natural result of the non-linear interactions among the elements. Consequently, from a CAS and CT perspective, structure cannot be deliberately created, only more or less stimulated to reach temporarily stable conditions. Furthermore, what CT shares with CAS is the notion of unorder, i.e. paradoxical situations involving both order and disorder simultaneously, and with the soft SA, the uniqueness of humans and organizations (Snowden and Stanbridge, 2004).

The perception of supply networks and logistics as being complex is emphasized by several authors (Bowersox et al., 2002; Christopher, 1998; Lambert et al., 1998a). However, there is a difference in how complexity is understood and how it can be simplified. In SA the complexity appears in systems as a cause of how many parts and interactions are present. For example, Milgate (2001, p. 107) observes that “complexity should be viewed as a deterministic component more related to the numerousness and variety in the system.” Furthermore, using a system dynamics approach Childerhouse and Towill (2003, p. 18) use the concept of seamless supply chains that is guided by twelve simplifying rules such as rule seven; “elimination of all uncertainties in all processes”, and rule twelve; “all players should think and act as one” (Towill, 1999, p. 11), i.e. a concept of simplifying, unifying, and eliminating assumptions. In contrast to SA, the overall belief in CAS is that complexity is the result of interacting and self-organizing agents which follow simple rules. On higher levels of description they create patterns of coherence or chaos. This means that the heterogeneity of the agents and the variety among them are taken into consideration and not regarded or treated as an average in which variety can be reduced or eliminated. CT shares most of the CAS considerations but emphasizes the subjectivity and the paradoxical nature of people. Consequently, in CT treating logistics as complex means considering human involvement and the paradoxes created in human interactions. For example, in global supply chains the incorporation of cultural differences influences the way trust is created (Zhao et al., 2007). Consequently, it also means considering the concrete, actual work being done and the mental models created by the humans involved. Behavioral assumptions The phenomenon of self-organization can be found in all approaches; however, it differs in its meaning and in how it is defined (Table II). In SA some elements of a system’s behavior are sometimes referred to as being formed by the feedback loops that are prevalent in logistics systems. At the same time there is a great belief in the engineering capability of the observer (researcher or manager) as being able to control the behavior and direct it toward some predetermined goal. A holistic picture of the situation provides the observer with the chance to see “all” feedback loops, and from that he/she rejects those which are deemed less efficient or effective for the overall goal. Van Ackere et al. (1993, p. 413) exemplify this: we are all used to the idea that automobiles, ships, aircrafts, office buildings, and bridges need careful design to achieve their purpose. But there is much less awareness that business organizations too are “designable”. Stacey et al. (2000, p. 120) state that: [. . .] in systems thinking, causality is primarily of the formative type taking a linear form in which the feedback process of the system causes its patterns of behavior, usually in a predictable way.

In CAS the discussion about self-organization is less determinable since there is a belief that it is far more spontaneous, i.e. less engineerable. Kauffman (1995, p. 185) states that “self-organization may be the precondition of evolvability itself. Only those systems that are able to organize themselves spontaneously may be able to evolve further.” In addition, the notion of system designer in CAS is less prevalent. In CT this notion of designability is limited, unpredictable, and paradoxical since the firm belief is that a process or phenomenon “can self-organize into disintegration just as it can into a rigid, repetitive pattern” (Stacey et al., 2000, p. 147). Furthermore, in CT a transformative teleology is assumed, meaning that changes do appear spontaneous and unpredictable and, as paradoxical it may seem, emergent continuity is kept simultaneously.

Beyond the systems approach 773

IJPDLM 42,8/9

Assumption

SA

CAS

CT

Deliberate design/selforganization

Designed and planned change through rational decisions and actions. Self-organization is mostly regarded a source for uncertainty and should be reduced to a controllable state Is not explicitly dealt with. It does not fit any deterministic assumptions or linear causality, and it is difficult to include in defined systems

Self-organization is regarded as a natural process and often as a source of novelty, creativity, and innovativeness

Diverse microinteractions of a paradoxical nature sustain identity and potentially transform it

774 Coevolution

Rationality/ bounded rationality Table II. A comparison of behavioral-related assumptions in SA, CAS, and CT

Due to interdependence coevolution is created through local active and lively behavior. Global characteristics emerge which then alter the way agents on lower levels interact and adapt Bounded and accepted as Rationality is bounded but desirable to achieve – such. Aspects such as conflict, power, etc. are with increased regarded as natural and information and knowledge rationality can essential for the be more or less attained. advancement, evolution, and survival of any Conflict, power, etc. are disregarded and cannot be system satisfactorily explained when the SA is used

Provides meaning to the paradox of simultaneous competition and cooperation

Bounded rationality is a prime assumption; subjectivity and narratives are central

The transformative teleology is derived from Hegel’s dialectical notion, i.e. an individual is both constructing and being influenced by the society at the same time (Johannessen, 2003), and there are no notions of system levels or hierarchies. Transformational processes can be understood by “paying attention to phenomena like self-organization, emergence, paradox and unpredictability” (Johannessen, 2003, p. 42). Conclusively, as stated by Nilsson (2003, p. 33): From a positivistic perspective, self-organization causes uncertainty and since it cannot be effectively controlled, planned or designed it should be reduced, or even eliminated. However, this process of self-organization is in several cases the source for novelty, creativity and innovativeness.

Another assumption, that of coevolution, can be defined as “a process of coupled, deforming landscapes where the adaptive moves of each entity alter the landscapes of its neighbors in the ecology or technological economy” (Kauffman and Macready, 1995, p. 27). Products and services “live” in niches created by other products and services, i.e. there is an interdependence that makes the coevolution occur. Power et al. (2010) provide an example from a cross-cultural study of operational investments in Asian versus Western countries. They (Power et al., 2010, p. 219) conclude that “the implication for plants in the industrialized economies is that they must find alternate combinations of resources to stay competitive, and that paradoxically their best ally may be the very individualism that in this study is found to be acting to constrain investment in team-based manufacturing methods.” Coevolutionary processes could be seen as combinations of traditional evolutionary

thinking and self-organization; it is a case of survival of the fittest but in both cooperative and competitive ways. In CT the assumption of coevolution provides meaning to the paradox of simultaneous competition and cooperation, causing human freedom and constraint to arise in the spontaneity and diversity of microinteractions in daily life. It is in these processes that true novelty emerges. CAS explains self-organization as arising when local, active, and lively behaviour results in global characteristics that alter the way agents on lower levels interact (Anderson, 1999). This in turn leads to changes in objectives, structures, and motivations, i.e. in coevolution of organizational functions, company networks and of the economy. In SA, again, there is a wish and striving to design change and that this can be done in some kind of isolation, i.e. by setting the boundaries correctly, and can be achieved by rational decisions and actions. Consequently, the assumption of coevolution is problematic since it challenges the belief in designability. The assumption of rationality is another assumption that differs widely among the three approaches. Rationality implies that every constituent part of a system being planned operates rationally, i.e. they all have access to perfect information, have the same background, similar beliefs and assumptions, and work toward the same goal (known and designed by someone outside the system). Furthermore, the environment in which the company or department works is stable both before and after the decision has been taken (Allen, 2000). While (perfect) rationality is more or less rejected by all the approaches, it is treated differently by them. In this regard Rigby et al. (2000, p. 181) state that: [. . .] in practice, open systems theory gives a central role to “management” to maximize bounded rationality. This “gatekeeper” role requires management to predict and design appropriate structures and responses and to manipulate resources and connected actors in what is perceived to be a desirable manner.

In other words, based on the assumption of objectivity and the notion of a formative and rational teleology, highly empirically observable aspects such as conflict, power, and creativity, are disregarded and cannot be satisfactorily explained when a SA is used. Gammelgaard (1997, p. 17) concludes, for example, that “conflicts are simply not on the research agenda in that part of reality which systems theory uncovers.” To a great extent this is also the case in CAS; however, instead of the great emphasis on the avoidance of, and reductions in, such uncomfortable aspects (e.g. conflict), these are seen as natural and essential for advancement, evolution, and survival. Since CT is developed for social complexity, bounded rationality is the rule, and considerations of subjective pictures of situations, narratives, and contexts are greatly emphasized. It is in acts of freedom that creativity and innovations are to be best produced, i.e. where the rationale is not determined by objective functions or rules set by organizations. Time-related assumptions The last category deals with the assumptions of determinism (Table III). As Phelan (1999) explains, the SA focuses on problem solving, and on action to improve some identified or conceptualized system in order to reach an optimal state, i.e. a predetermined objective that represents a stable future situation. Bowersox and Closs (1996, p. 459), exemplify this determinable ability by arguing that use of the SA in logistics stresses “total integrated effort toward the accomplishment of predetermined objectives.” Stacey et al. (2000, p. 59) provides a reflection on such assumptions. They state that:

Beyond the systems approach 775

IJPDLM 42,8/9

776

Table III. A comparison of time-related assumptions in SA, CAS, and CT

Assumption

SA

CAS

CT

Determinism/ indeterminism

Deterministic assumptions underlie the great emphasis for reducing complexity and the focus on actions to improve some identified or conceptualized system in order to reach an optimal state

The future emerges in the interaction among individuals. Patterns may be predictable since the future is regarded as nonrandom, if there are underlying rules that can be assessed

Indeterminable since the future is under perpetual construction and is created in the relational processes among the agents. By focusing on exploratory analysis aimed at understanding a certain phenomenon, the individuals involved are helped to live with complexity instead of trying to remove it

[. . .] systems thinking provide powerful ways of thinking about, and designing means of securing, organizational stability and continuity and, in the case of systems thinking, unfolding potential change already enfolded in the system.

This means that a formative and rationalist view is taken for granted and the assumption is that systems are deterministic. This notion is less emphasized in CAS, as Choi et al. (2001, p. 356) observe: “in a complex system, it is often true that the only way to predict how the system will behave in the future is to wait literally for the future to unfold.” Changes in CAS are seen as being non-linear and toward a non-random future, meaning that the future is not totally random; however, neither can it be predicted in a deterministic manner. In addition, in CAS the role of path dependency is also emphasized since projections from the past may influence both its present actions and its anticipations and expectations of the future. In the process of anticipating the future, agents search for patterns in the past and, as Beinhocker (1999, p. 97) declares, “our drive to see patterns and trends is so strong that we will even see them in perfectly random data.” This makes the connection with a more transformative view for the CAS approach explicit: due to the inherent indeterminism, the ability to make predictions is heavily reduced. However, in the “rule”-based CAS community the wish for prediction is still apparent, often through computer simulations, but it is often on the level of emerging patterns that result from these models. The focus in CT is, in contrast, on exploratory analysis where understanding helps the researcher and practitioner to live with the complexity and the indeterminism it brings instead of trying to remove it. The emphasis is even more on the living present from which we change both the future and the past in the activities we perform and in how we relate to other people. Concluding discussion This paper has taken up the paradigmatic discourse initiated before in SCM and logistics in order to further develop the discipline. By providing a reflective analysis and discussion of established and new research approaches, further advancements – of both theoretical and practical nature – can be gained for the SCM and logistics community. Several authors have asked for the introduction of new paradigms and new approaches in logistics research and practice (Arlbjørn and Halldorsson, 2002; Ellram et al., 2007; Gammelgaard, 2004; Sachan and Datta, 2005). This paper analyzes approaches from

complexity theory that have been of growing interest in literature. With a discourse in SCM and logistics research set on the SA; it is concluded that the underlying assumptions of CAS and CT are more appropriate than those of SA for contemporary challenges of organizational complexity in SCM and logistics, e.g. challenges of sustainability, innovation, collaboration, and sense making. The three approaches presented and discussed in this paper (SA, CAS, and CT) all represent different perspectives on reality, based on different assumptions, and are thus applicable to different contexts, situations, and problems. Knowing the limitations of each approach and its underlying assumptions is valuable when using these approaches in any research endeavor. In SA there are certain assumptions which are less beneficial in explaining logistics activities on a daily basis or in making sense for the people involved. Dealing with issues of sense making, CT is more appropriate. SAs provide objective, illustrative descriptions of technical parts of logistics “systems”; however, they place less emphasis on the subjective perspectives logistics people have in “reality”. The problem with CAS applied to human phenomena is similar to that of SA, in the sense that CAS has been developed and derives from studies of mathematics, physics, chemistry, and biology, and not for human or social phenomena. As a final conclusion, in order for the SCM and logistics discipline to develop strategic dimensions and increase the strategic value for firms, aspects such as creativity, novelty, and innovation based on human involvement will be necessary for approaches that – based on appropriate assumptions – are developed to handle such aspects. Both CAS and CT provide alternatives in exploring these aspects in SCM and logistics. However, in line with the notion by Richardson (2003, p. 8) – “models are tools that can be used and abused – the best models are worthless in linear hands” – methodological approaches can be taken from many disciplines and used in various ways. Without a process reflecting the underlying assumptions, the potential of such effort wills be hampered. Further directions for research There are several beneficial aspects of approaching SCM and logistics using different paradigmatic grounds. Examples of new approaches are exemplified by Lewis and Suchan (2003) who propose structuration theory, and Na¨slund (2002) who suggests action research. We argue that for further developments of SCM and logistics, approaching SCM and logistics phenomena from several perspectives and theories may achieve even better results. The efforts of unifying SCM and logistics research may provide value, since in some situations common, defined terminology may facilitate communication and consensus. However, based on the assumption of an ever-changing reality filled with transformations of identities, an approach with multiple perspectives on SCM and logistics may be even more valuable. As MacIntosh and MacLean (1999) state: [. . .] if one accepts the notion that systems are not only complex and adaptive, but that their complexity and adaptiveness can itself change, then one can see different implications for the evolution of organizations.

This is especially challenging in creating effective logistics (Kohn and McGinnis, 1997), i.e. in making strategic decisions concerning “the right things to do” in a world where possibilities continually increase. With a set of assumptions that focus on the effectiveness of supply chain processes instead of on today’s efforts on efficiency, the dominant cost-reducing focus can be complemented with both revenue-enhancing activities and with higher leverage from improvement efforts. Today, SCM and logistics

Beyond the systems approach 777

IJPDLM 42,8/9

778

effectiveness efforts are restrained by the mechanical beliefs inherent in the SA applied since these need a different type of logic. This logic covers human-related aspects such as creativity and innovativeness, and from this novel processes and activities can thus be created, i.e. different ways of fulfilling customer requirements, and ensuring the satisfaction of other stakeholders involved. Consequently, future research should encompass further exploration of perspectives and how different perspectives can contribute to increased understanding of SCM and logistics strategies and processes, i.e. effectiveness. This exploration should be both conceptually and empirically based, and CAS and CT should provide theoretical grounds for development. Avoidance of such exploration will only provide a self-reinforcing belief that SCM and logistics is about running the machine faster. Furthermore, application of CT is suitable for cross-culture research in supply chains since the underlying assumptions are appropriate when addressing human-related aspects in logistics and SCM. Finally, comparative studies of SA, CAS, and CT in a global supply chain context can contribute to developing theories that can capture the supply chain complexity most organizations operate with today. References Allen, P.M. (2000), “Knowledge, ignorance and learning”, Emergence, Vol. 2 No. 4, pp. 78-103. Anderson, P. (1999), “Complexity theory and organization science”, Organization Science, Vol. 10 No. 3, pp. 216-32. Andersson, D. (1997), Third Party Logistics – Outsourcing Logistics in Partnerships, Department of Management and Economics, Linko¨ping Institute of Technology, Linko¨ping. Arbnor, I. and Bjerke, B. (1997), Methodology for Creating Business Knowledge, 2nd ed., Sage, Thousand Oaks, CA. Arlbjørn, J.S. and Halldorsson, A. (2002), “Logistics knowledge creation: reflections on content, context and processes”, International Journal of Physical Distribution & Logistics Management, Vol. 32 No. 1, pp. 22-40. Ashby, W.R. (1956), An Introduction to Cybernetics, 1st ed., Chapman & Hall, London. Ballou, R.H. (2004), Business Logistics/Supply Chain Management: Planning, Organizing, and Controlling the Supply Chain, 5th ed., Pearson Prentice-Hall, Upper Saddle River, NJ. Bechtel, C. and Jayaram, J. (1997), “Supply chain management: a strategic perspective”, International Journal of Logistics Management, Vol. 8 No. 1, pp. 15-34. Beer, S. (1959), Cybernetics and Management, 1st ed., The English Universities Press Ltd, London. Beinhocker, E.D. (1999), “Robust adaptive strategies”, Sloan Management Review, Vol. 40 No. 3, pp. 95-107. Boulding, K.E. (1956), “General systems theory – the skeleton of science”, Management Science, Vol. 2 No. 3, pp. 197-208. Bowersox, D.J. and Closs, D.J. (1996), Logistical Management: The Integrated Supply Chain Process, International edition, McGraw-Hill, New York, NY. Bowersox, D.J., Closs, D.J. and Cooper, M.B. (2002), Supply Chain Logistics Management, First International edition, McGraw-Hill/Irwin, New York, NY. Burrel, G. and Morgan, G. (1979), Sociological Paradigms and Organizational Analysis; Elements of the Sociology of Corporate Life, 1st ed., Heinemann Educational Books Ltd, London.

Cannon, J.P., Doney, P.M., Mullen, M.R. and Petersen, K.J. (2010), “Building long-term orientation in buyer-supplier relationships: the moderating role of culture”, Journal of Operations Management, Vol. 28 No. 3, pp. 506-21. Checkland, P. (1993), Systems Thinking, Systems Practice, Wiley, Chichester. Childerhouse, P. and Towill, D.R. (2003), “Simplified material flow holds the key to supply chain integration”, Omega, Vol. 31 No. 1, pp. 17-27. Choi, T.Y., Dooley, K.J. and Rungtusanatham, M. (2001), “Supply networks and complex adaptive systems: control versus emergence”, Journal of Operations Management, Vol. 19 No. 3, pp. 351-66. Christopher, M. (1998), Logistics & Supply Chain Management, 2nd ed., Pearson Education, London. Colbert, B.A. (2004), “The complex resource-based view: implications for theory and practice in strategic human resource management”, The Academy of Management Review, Vol. 29 No. 3, pp. 341-58. Coyle, J.J., Bardi, E.J. and Langley, C.J. (2003), The Management of Business Logistics: A Supply Chain Perspective, 7th ed., South-Western, Mason/OH. Dent, E.B. (1999), “Complexity science: a worldview shift”, Emergence, Vol. 1 No. 4, pp. 5-19. Disney, S.M., Naim, M.M. and Towill, D.R. (1997), “Dynamic simulation modelling for lean logistics”, International Journal of Physical Distribution & Logistics Management, Vol. 27 No. 3, pp. 174-96. Ellram, L.M., Tate, W. and Carter, C. (2007), “Product-process-supply chain: an integrative approach to three-dimensional concurrent engineering”, International Journal of Physical Distribution & Logistics Management, Vol. 37 No. 4, pp. 305-30. Forrester, J.W. (1968), “Industrial dynamics – after the first decade”, Management Science, Vol. 14 No. 7, pp. 398-415. Gammelgaard, B. (1997), “The systems approach in logistics”, Proceedings from the 8th Nordic Logistics Conference, Copenhagen Business School, Institute for Logistics and Transport. Gammelgaard, B. (2004), “Schools in logistics research? A methodological framework for analysis of the discipline”, International Journal of Physical Distribution & Logistics Management, Vol. 34 No. 6, pp. 479-91. Golicic, S.L., Davis, D.F. and McCarthy, T.M. (2005), “A balanced approach to research in supply chain management”, in Kotzab, H., Seuring, S., Mu¨ller, M. and Reiner, G. (Eds), Research Methodologies in Supply Chain Management, Physica, Heidelberg. Grant, D.B., Lambert, D.M., Stock, J.R. and Ellram, L.M. (2005), Fundamentals of Logistics Management, European Edition, McGraw-Hill, Boston, MA. Guba, E.G. and Lincoln, Y.S. (1998), “Competing paradigms in qualitative research”, in Denzin, N.K. and Lincoln, Y.S. (Eds), The Landscape of Qualitative Research, Sage, Thousand Oaks, CA, pp. 195-220. Gubi, E., Arlbjørn, J.S. and Johansen, J. (2003), “Doctoral dissertations in logistics and supply chain management: a review of Scandinavian contributions from 1990 to 2001”, International Journal of Physical Distribution & Logistics Management, Vol. 33 No. 10, pp. 854-85. Hofstede, G. (1991), Cultures and Organizations, McGraw-Hill, London. Holmberg, S. (2000), Supply Chain Integration Through Performance Measurement, Department of Industrial Management and Logistics, Division of Engineering Logistics, Lund University, Lund.

Beyond the systems approach 779

IJPDLM 42,8/9

780

Holweg, M. and Pil, F.K. (2008), “Theoretical perspectives on the coordination of supply chains”, Journal of Operations Management, Vol. 26 No. 3, pp. 389-407. Hopper, T. and Powell, A. (1985), “Making sense of research into the organizational and social aspects management accounting: a review of its underlying assumptions”, Journal of Management Studies, Vol. 22 No. 5, pp. 429-65. Hunt, J.W. (1981), “Applying American behavioural science: some cross-cultural problems”, Organizational Dynamics, Summer. Johannessen, S. (2003), An Explorative Study of Complexity, Strategy and Change in Logistics Organizations, Department of Industrial Economics and Technology Management, Trondheim, Norway. Johannessen, S. (2005), “Supply chain management and the challenge of organizational complexity – methodological considerations”, in Kotzab, H., Seuring, S., Mu¨ller, M. and Reiner, G. (Eds), Research Methodologies in Supply Chain Management, Physica, Heidelberg. Johannessen, S. and Solem, O. (2002), “Logistics organizations: ideologies principles and practice”, International Journal of Logistics Management, Vol. 13 No. 1, pp. 31-42. Kauffman, S. (1995), At Home in the Universe: The Search for Laws of Self-organization and Complexity, Oxford University Press, New York, NY. Kauffman, S. and Macready, W. (1995), “Technological evolution and adaptive organizations”, Complexity, Vol. 1 No. 1, pp. 26-43. Kent, J.L. Jr and Flint, D. (1997), “Perspectives on the evolution of logistics thought”, Journal of Business Logistics, Vol. 18 No. 2, pp. 15-30. Kohn, J.W. and McGinnis, M.A. (1997), “Advanced logistics organization structures: revisited”, Journal of Business Logistics, Vol. 18 No. 2, pp. 147-63. Kuhn, T.S. (1996), The Structure of Scientific Revolutions, 3rd ed., The University of Chicago Press, Chicago, IL. Kupers, R. (2001), “What organizational leaders should know about the new science of complexity”, Complexity, Vol. 6 No. 1, pp. 14-19. Kurtz, C.F. and Snowden, D.J. (2003), “The new dynamics of strategy: sense-making in a complex and complicated world”, IBM Systems Journal, Vol. 42 No. 3, pp. 462-82. Lambert, D.M. and Cooper, M.C. (2000), “Issues in supply chain management”, Industrial Marketing Management, Vol. 29 No. 1, pp. 65-83. Lambert, D.M., Cooper, M.C. and Pagh, J.D. (1998a), “Supply chain management: implementation issues and research opportunities”, The International Journal of Logistics Management, Vol. 9 No. 2, pp. 1-18. Lambert, D.M., Stock, J.R. and Ellram, L.M. (1998b), Fundamentals of Logistics Management, International edition, McGraw-Hill Higher Education, London. Lee, H.L., Padmanabhan, V. and Whang, S. (1997), “Information distortion in a supply chain: the bullwhip effect”, Management Science, Vol. 43 No. 4, pp. 546-58. Lewis, I. and Suchan, J. (2003), “Structuration theory: its potential impact on logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 33 No. 4, pp. 296-315. MacIntosh, R. and MacLean, D. (1999), “Conditioned emergence: a dissipative structures approach to transformation”, Strategic Management Journal, Vol. 20 No. 4, pp. 297-316. MacIntosh, R. and MacLean, D. (2001), “Conditioned emergence: researching change and changing research”, International Journal of Operations & Production Management, Vol. 21 No. 10, pp. 1343-57.

Mears-Young, B. and Jackson, M.C. (1997), “Integrated logistics – call in the revolutionaries!”, Omega, Vol. 25 No. 6, pp. 605-18. Mentzer, J.T. and Kahn, K.B. (1995), “A framework of logistics research”, Journal of Business Logistics, Vol. 16 No. 1, pp. 231-50. Mentzer, J.T., Min, S. and Bobbitt, L.M. (2004), “Toward a unified theory of logistics”, International Journal of Physical Distribution & Logistics Management, Vol. 34 No. 8, pp. 606-27. Milgate, M. (2001), “Supply chain complexity and delivery performance: an international exploratory study”, Supply Chain Management: An International Journal, Vol. 6 No. 3, pp. 106-18. Monczka, R.M., Petersen, K.J., Handfield, R.B. and Ragatz, G.L. (1998), “Success factors in strategic supplier alliances; the buying company perspective”, Decision Sciences, Vol. 29 No. 3, pp. 553-77. Morgan, G. (1983), Beyond Method – Strategies for Social Research, 1st ed., Sage, Newbury Park, CA. Na¨slund, D. (2002), “Logistics needs qualitative research – especially action research”, International Journal of Physical Distribution & Logistics Management, Vol. 32 No. 5, pp. 321-38. Nilsson, F. (2003), A Complex Adaptive Systems Approach on Logistics, Division of Packaging Logistics, Lund University, Lund. Nilsson, F. (2005), Adaptive Logistics – Using Complexity Theory to Facilitate Increased Effectiveness in Logistics, Division of Packaging Logistics, Lund University, Lund. Nilsson, F. (2006), “Logistics management in practice – towards theories of complex logistics”, International Journal of Logistics Management, Vol. 17 No. 1, pp. 38-57. Nilsson, F. (2007), “Towards a dialectic complexity framework – philosophical reflections”, in Richardson, K. and Cilliers, P. (Eds), Explorations in Complexity Thinking: Pre-Proceedings of the 3rd International Workshop on Complexity and Philosophy, ISCE Publishing, Boston, MA. Nilsson, F. and Darley, V. (2006), “On complex adaptive systems and agent-based modelling for improving decision-making in manufacturing and logistics settings – experiences from a packaging company”, International Journal of Operations & Production Management, Vol. 26 No. 12, pp. 1351-73. Pascale, R.T. (1999), “Surfing the edge of chaos”, Sloan Management Review, Vol. 40, pp. 83-95. Phelan, S.E. (1999), “A note on the correspondence between complexity and systems theory”, Systemic Practice and Action Research, Vol. 12 No. 3, pp. 237-46. Powell, T.C. (2003), “Strategy without ontology”, Strategic Management Journal, Vol. 24 No. 3, pp. 285-91. Power, D., Schoenherr, T. and Samson, D. (2010), “The cultural characteristic of individualism/collectivism: a comparative study of implications for investment in operations between emerging Asian and industrialized western countries”, Journal of Operations Management, Vol. 28 No. 3, pp. 206-22. Prigogine, I. (1997), The End of Certainty – Time, Chaos, and the New Laws of Nature, 1st ed., The Free Press, New York, NY. Richardson, K.A. (2003), “On the limits of bottom-up computer simulation: towards a nonlinear modeling culture”, in Sprague, R.H. Jr (Ed.), Proceeding of the 36th Hawaiian International Conference on System Sciences, Washington, DC. Richardson, K.A. and Cilliers, P. (2001), “Special editor’s note: what is complexity science? A view from different directions”, Emergence, Vol. 3 No. 1, pp. 5-22.

Beyond the systems approach 781

IJPDLM 42,8/9

782

Richardson, K.A., Cilliers, P. and Lissack, M.R. (2001), “Complexity science: a ‘gray’ science for the ‘stuff in between’”, Emergence, Vol. 2 No. 3, pp. 6-18. Rigby, C., Day, M., Forrester, P. and Burnett, J. (2000), “Agile supply: rethinking systems thinking, systems practice”, International Journal of Agile Management Systems, Vol. 2 No. 3, pp. 178-86. Sachan, A. and Datta, S. (2005), “Review of supply chain management and logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 35 No. 9, pp. 664-705. Saghir, M. (2004), A Platform for Packaging Logistics Development – A Systems Approach, Department of Design Sciences, Division of Packaging Logistics, Lund University, Lund. Schwartz, S. (Ed.) (1994), Beyond Individualism/Collectivism: New Cultural Dimensions of Value, Sage, Thousand Oaks, CA. Senge, P.M. (1990), The Fifth Discipline: The Art and Practice of the Learning Organization, 1st ed., Doubleday, New York, NY. Simon, H.A. (1996), The Science of the Artificial, 3rd ed., MIT Press, Cambridge, MA. Skjøtt-Larsen, T., Schary, P.B., Mikkola, J.H. and Kotzab, H. (2007), Managing the Global Supply Chain, 3rd ed., Liber, Copenhagen. Snowden, D. (2002), “Complex acts of knowing: paradox and descriptive self-awareness”, Journal of Knowledge Management, Vol. 6 No. 2, pp. 100-11. Snowden, D. and Stanbridge, P. (2004), “The landscape of management: creating the context for understanding social complexity”, ECO, Vol. 6 Nos 1/2, pp. 140-8. Stacey, R.D. (2001), Complex Responsive Processes in Organizations – Learning and Knowledge Creation, Routledge, London. Stacey, R.D. (2002), “Organizations as complex responsive processes of relating”, Journal of Innovative Management, Vol. 8 No. 2, pp. 27-39. Stacey, R.D., Griffin, D. and Shaw, P. (2000), Complexity and Management – Fad or Radical Challenge to Systems Thinking?, Routledge, London. Sterman, J. (2000), Business Dynamics Systems Thinking and Modeling for a Complex World, McGraw-Hill, New York, NY. Stock, G.N., Greis, N.P. and Kasarda, J.D. (1999), “Logistics, strategy and structure. A conceptual framework”, International Journal of Physical Distribution & Logistics Management, Vol. 29 No. 4, pp. 224-39. Surana, A., Kumara, S., Greaves, M. and Raghavan, U.N. (2005), “Supply chain networks: a complex adaptive systems perspective”, International Journal of Production Research, Vol. 43 No. 20, pp. 4235-65. Svensson, G. (2010), “Teleological approaches in supply chain management: illustrations”, Supply Chain Management: An International Journal, Vol. 15 No. 1, pp. 16-20. Tokar, T. (2010), “Behavioural research in logistics and supply chain management”, International Journal of Logistics Management, Vol. 21 No. 1, pp. 89-103. Towill, D.R. (1999), “Simplicity wins: twelve rules for designing effective supply chains”, The Journal of the Institute of Operations Management, Vol. 25 No. 2, pp. 9-13. Trim, P.R.J. and Lee, Y.I. (2004), “A reflection on theory building and the development of management knowledge”, Management Decision, Vol. 42 No. 3, pp. 473-80. Vafidis, D. (2007), Approaches for Knowledge and Application Creation in Logistics. An Empirical Analysis Based on Finnish and Swedish Doctoral Dissertations Published Between 1994 and 2003, Turku School of Economics, Turku.

van Ackere, A., Larsen, E.R. and Morecroft, J.D.W. (1993), “Systems thinking and business process redesign: an application to the beer game”, European Management Journal, Vol. 11 No. 4, pp. 412-23. von Bertalanffy, L. (1969), General Systems Theory – Foundations, Development and Applications, First revised edition, George Braziller Inc., New York, NY. Waidringer, J. (2001), Complexity in Transportation and Logistics Systems: An Integrated Approach to Modelling and Analysis, Department of Transportation and Logistics, Chalmers University, Go¨teborg. Wysick, C., McKelvey, B. and Hu¨lsmann, M. (2008), “‘Smart parts’ supply networks as complex adaptive systems: analysis and implications”, International Journal of Physical Distribution & Logistics Management, Vol. 38 No. 2, pp. 108-25. Zhao, X., Flynn, B.B. and Roth, A.V. (2007), “Decision science research in China: current status, opportunities, and propositions for research in supply chain management, logistics, and quality management”, Decision Sciences, Vol. 38 No. 3, pp. 368-88. About the authors Fredrik Nilsson is Associate Professor in Packaging Logistics at the Department of Design Sciences, Lund University, Sweden. His research areas are complexity theory and thinking, with current projects in healthcare, packaging and consumer goods supply chains. He has published work in the International Journal of Operations & Production Management, International Journal of Logistics Management and in the ISEC book Explorations in Complexity Thinking. Fredrik Nilsson is the corresponding author and can be contacted at: fredrik. [email protected] Britta Gammelgaard is Professor of Logistics and Supply Chain Management at Copenhagen Business School. Among her main research interests are SCM and logistics innovation and research methodologies in logistics and SCM. Dr Gammelgaard’s work has been published in international journals such as Journal of Business Logistics, International Journal of Physical Distribution & Logistics Management and Transportation Journal. She is a member of the Editorial Boards of JBL and International Journal of Physical Distribution & Logistics Management and Associate Editor of Journal of Supply Chain Management. In 2009, she received the Danish Transportation Research Award (Hedorf’s Fond).

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Beyond the systems approach 783

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0960-0035.htm

IJPDLM 42,8/9

784 Received 19 December 2011 Accepted 16 January 2012

A reviewer’s guide to the grounded theory methodology in logistics and supply chain management research Ila Manuj and Terrance L. Pohlen Department of Marketing and Logistics, University of North Texas, Denton, Texas, USA Abstract Purpose – The purpose of this paper is to analyze previous grounded theory articles and, based on this analysis, to provide a framework to assist reviewers in evaluating grounded theory research and increasing the rigor and credibility of this methodology in logistics and supply chain journals. Design/methodology/approach – An analysis of existing articles appearing in the leading logistics and supply chain journals combined with an extensive review of the grounded theory method literature were used to develop a comprehensive framework for evaluating grounded theory research. Findings – The paper finds that no standard criteria for publication of grounded theory research exists in logistics and supply chain journals. Grounded theory is routinely confused with other qualitative methodologies. Overall, this situation leads to publications that do not adequately address or report on the process for developing a grounded theory. Research limitations/implications – Reviewers can use this paper to establish the quality of grounded theory research. Reviewers who are unfamiliar with or skeptical of the grounded theory method can use the framework to evaluate the rigor and credibility of a grounded theory study rather than rejecting such research. The checklist can be used to provide thorough and constructive reviews to authors. Originality/value – The paper presents a framework that provides a ready reference for reviewers to assess whether the authors have taken appropriate action in selecting a grounded theory methodology, collecting and analyzing data, developing a theory grounded in the data, and for evaluating their research. Existing research is compared with the framework to identify potential shortcomings in the review process. The application of the framework to the review of future articles provides an opportunity to increase the credibility and rigor of grounded theory research in logistics and supply chain management journals. Keywords Grounded theory research, Qualitative research, Guide for reviewers, Distribution management, Supply chain management, Research, Periodicals, Literature Paper type General review

International Journal of Physical Distribution & Logistics Management Vol. 42 No. 8/9, 2012 pp. 784-803 q Emerald Group Publishing Limited 0960-0035 DOI 10.1108/09600031211269758

1. Introduction Grounded theory (GT) has become increasingly accepted by logistics and supply chain management researchers as a valid research methodology (Davis-Sramek and Fugate, 2007). Although logistics and supply chain management have gained recognition as academic disciplines, the supporting theory has largely been borrowed from Authors are listed in the alphabetical order of their last names. Both authors contributed equally to the paper.

other disciplines. Some of these theories, because of their lack of grounding in logistics and supply chain management, may not fit or sufficiently explain logistics and supply chain phenomena. The attraction of GT stems from the need for theory to develop creative perspectives, generate powerful insights into human interaction and business practices, and explain new and even well-researched complex social phenomena (Mello and Flint, 2009; Glaser and Strauss, 1967). GT presents several challenges to reviewers tasked with assessing the rigor and credibility of research. Reviewers have to rely on the authors’ accounts of the research process and explanations of how key evaluation criteria were satisfied. Reviewers must often sift through lengthy descriptions of the research process and determine the appropriate balance to be struck between a parsimonious description of the research and sufficient detail to demonstrate rigor and credibility. Furthermore, there is no comprehensive reference for the GT method. To obtain a thorough understanding of GT, reviewers must rely on several texts or articles focusing on specific aspects of the GT method. Many of these methodological developments have taken place outside the management literature (Jones and Noble, 2007). Researchers frequently confuse GT with qualitative approaches such as case studies or content analysis and apply GT as a generic term to any qualitative approach that employs inductive analysis ( Jones and Noble, 2007). Qualitative approaches frequently employ techniques such as coding and verification. This common terminology has produced situations of methodological slurring (Suddaby, 2006; Goulding, 2002) where researchers use interpretative methods to analyze realistic assumptions. Jones and Noble (2007) argue that GT is at risk of losing its integrity as many researchers have adopted only snippets of the methodology. As a result, GT research has often been perceived as lacking a consistent methodology, depth, and rigor (Mello and Flint, 2009). The purpose of this paper is to develop a framework to assist reviewers in assessing the rigor and credibility of manuscripts employing a GT methodology. The supporting objectives include: analyzing articles appearing in the leading journals which employ GT methodology; identifying shortcomings in these articles regarding the application of the methodology; providing references to high quality GT studies to serve as a guide; distilling knowledge from multiple sources on GT research; and combining this knowledge with the best practices obtained from high quality articles. Four sections comprise the remainder of this paper. The next section identifies the need for a reviewer’s framework. Section 3 provides a reviewer’s perspective on GT development and the implications associated with aligning the method with a consistent paradigm of inquiry. The Section 4 summarizes the steps in GT analysis and identifies the major actions required to rigorously emerge a GT. The final section presents a framework of questions the reviewer should consider when evaluating GT research. 2. The need for a reviewer’s framework To understand the rigor of GT studies in the logistics and supply chain management, a review was conducted of articles published during the past 12 years appearing in the journals ranked among the top five in the discipline (Carter et al., 2009; Gibson and Hanna, 2003; Menachof et al., 2009): Journal of Business Logistics, Transportation Journal, Journal of Supply Chain Management, International Journal of Physical Distribution and Logistics Management, and The International Journal of Logistics Management. The review had two objectives. First, the articles were analyzed for

Grounded theory methodology

785

IJPDLM 42,8/9

786

a rigorous application of the GT method to identify methodological issues, advances, and techniques for inclusion in the reviewer’s framework. Second, articles with rigorous execution were identified as examples to guide reviewers. The review produced an initial sample of 33 articles with “grounded theory” appearing in the title or abstract. Articles combining GT with other qualitative methods were included in the analysis but excluded from being identified as examples. 22 articles employed only GT. Of these, five were selected as examples based on a justification and comprehensive description of the methodology and rigorous analysis. Table I provides a summary of the five articles. The following paragraphs summarize the key issues identified during the review. Introduction and literature review The review suggested “methodological slurring” ( Jones and Noble, 2007) in the majority of articles. Several mentioned GT in the introduction, but the subsequent discussion and methodology did not use GT. In other instances, studies combined GT with other qualitative approaches. These studies typically did not focus on theory development; instead, they presented findings drawn from rich descriptions and comparisons of the phenomenon in a limited number of instances. Methodology The review of the methodology produced two categories of articles. The first included research that mentioned GT but applied only bits and pieces of the method. For example, studies combining GT with case studies adopted techniques such as coding and constant comparative analysis. However, the studies began with a pre-selected number of cases. The coding tended to be descriptive rather than explanatory. Theoretical sampling and saturation were not used, and how a theory emerged from the development of categories or theoretical memos was unclear. In the second category, a GT approach was employed, but these articles did not clearly establish the context under investigation. In situations where the context was specified, a connection generally did not exist between the characteristics of the context and its relevance to the phenomenon. Specifying the context is important because the resulting theory can only explain phenomena within the context of the study. Three major issues were identified with data collection. First, most studies employed in-depth interviews as the main source of data and stated that theoretical sampling was employed. The choice and number of participants was usually explained, but information regarding theoretical sampling of subsequent participants and the rationale for reaching theoretical saturation were weak or missing. For example, interviews stopped when no new information was obtained. This cursory treatment of theoretical sampling and theoretical saturation raises questions regarding methodological rigor. Second, most studies claimed to use multiple sources of data but rarely provided insights regarding how this data was analyzed or shaped the findings. Third, the majority of studies used an interview protocol but failed to provide or describe it which raises concerns regarding the objectivity of the questions, whether the questions biased the findings, and how the interview protocol evolved to support theoretical sampling. Most articles did not adequately explain the analysis process stating merely that guidelines for open and selective coding were followed and constant comparative analysis was used. Justifications for which GT approach (Glaser, Strauss and Corbin,

Purpose is to develop a data-driven descriptive model of logistics outsourcing strategy development process Inconsistency between prescriptive and descriptive models. Outsourcing strategy development process not addressed in literature

Introduction and literature review

Methodology

Use both literature and field research to develop GT GT used for investigating processes where limited theory exists. No specific context. Companies from different industries. Purposive sampling based on extent of outsourcing. Theoretical sampling used. No protocol provided. Interview guide with specific questions, followup questions, and outline of research used. Unstructured interviews Coding followed established GT guidelines. Open coding to construct categories, properties and dimensions. Selective coding of core categories. Discovery of underlying uniformities Working hypotheses. Looking for negative or contradictory evidence Mollenkopf et al. Purpose is theory development within Qualitative method adopted as phenomenon not previously explored (2007) supply chain strategy. from cross-functional perspective and is poorly understood; no Limited attention to theory-based control over the phenomenon, and because other researchers have research in managing returns. used the methodology. Firms in Italy chosen because of European Functional integration is ignored. economic integration and a pan-European regulatory environment. Research seeks to understand linkages Research setting enables study of firm-level and external factors. between marketing and logistics at Modified theoretical sampling. Industry sectors and supply chain strategic and operational levels as they echelons included to maximize diversity. Participants were deal with returns management influential decision makers. Invitations extended until diversity achieved. Bracketing-type technique for reducing bias. Open-ended questions. Interview guide used but not provided. Interviews digitally recorded. Information redundancy achieved via modified theoretical sampling, hence inferred that theoretical saturation reached. No description of process of data analysis Manuj and Purpose is to develop a model of supply GT chosen because objective is to develop a comprehensive model Sahin (2011) chain decision-making complexity. grounded in actual business practice. No specific context mentioned. Relationship between supply chain Initial sample based on participants’ ability to provide relevant data, complexity and decision-making is varied job titles and profiles, articulation skills, and willingness. poorly understood. A theoretical model Further participants selected based on emerging questions. needed to advance research. Identify Semi-structured interview process. Broad questions followed by specific research questions. Use specific and directed questions. Interview protocol provided. literature to identify gaps Interviews continued till theoretical saturation defined as “no new information or insights from additional interviews”. Simultaneous collection, coding, and analysis of interview data. Secondary data sources also used

Mello et al. (2008)

Name and year

(continued)

Trustworthiness evaluated based on credibility, dependability, transferability, confirmability, integrity, fit, understanding, generality, and control. A theoretical model developed and propositions offered

Two sets of criteria: (1) credibility, dependability, transferability, confirmability, and integrity (2) fit, understanding, generality, and control. A causal model is proposed. Four research propositions are offered

Criteria of fit, relevance, workability, modifiability, scope and parsimony. Descriptive model of logistics outsourcing process and additional factors developed

Findings, discussion, and contributions

Grounded theory methodology

787

Table I. Review of selected GT articles

Purpose is to present a theoretical framework for performance-based logistics from perspective of servicedominant logic. SDL suitability to advancing theory in supply chain management and PBL justified. The research problem stated as lack of a theoretical structure of PBL

Methodology

GT appropriate since objective is to emerge a theory. PBL and SDL lack explanatory and predictive framework, studying how complex systems adapt to their environment. Defense context selected. A variety of participants from companies with PBL, no PBL, and converting to PBL were included. Multiple sources including case studies, archival review, practitioner conference interaction, and interviews used. Data collection, coding, and analysis described in detail. Constant comparative method, memoing, and theoretical sampling described. Theoretical saturation described as no new insights, themes, and issues. Additional interviews demonstrated consistent group of constructs and relationships between them. Detailed tables suggestive of extensive memoing Flint et al. (2005) Purpose is to construct a theory of GT used due to social nature of innovation, problematic nature, likely logistics innovation from ground up and importance of processes, and phenomena not deeply explored. No compare to extant literature. specific context selected. Variety of participants from logistics Concept of innovation is defined and service provider and other industries selected. explored in a logistics context. Specific Relied on discovery-oriented, open-ended questions for interviews research question proposed that deals and site visits. Sampling explained in detail starting with with innovation in a logistics context convenience sample with theoretical sampling later. Sample of 33 participants identified as enough for theory building. Interview guide provided and modified based on emergent theory. Guidelines for GT briefly described. Used GT and hermeneutic processes for data analysis. Constant comparative method. Use of software. Theoretical memoing

Introduction and literature review

Table I.

Randall et al. (2010)

No specific evaluation criteria presented. Discussion identifies specific contributions and ties them to existing literature. Managerial implications discussed. Specific future research directions identified

Evaluation based on fit. A detailed theoretical structure for PBL presented. SCM community introduced to SDL concept as a potential foundation for investigating phenomena

Findings, discussion, and contributions

788

Name and year

IJPDLM 42,8/9

Charmaz, and others) was adopted or how the approach influenced the coding and subsequent analysis were starkly missing; studies explaining these elements were rare exceptions. Moreover, the information regarding the use of memos, development of working hypotheses or interim theories, and use of data other than interviews was limited. Findings, discussion, and contributions The review indicated a need to improve explanations regarding how the emergent theory satisfied the criteria for a GT and advanced the understanding of the phenomenon under investigation. The studies typically recognized that the proposed theory needed to conform to several criteria such as fit, relevance, and modifiability (Glaser, 1978). However, the majority of studies simply provided a list and description of key findings, which failed to provide any theoretical insight or establish connections between the new findings and existing theory. Future research was frequently not linked to the context of the research and failed to explain how expanding the context could lead to formal, or eventually grand, theories with greater explanatory power. The researchers rarely addressed whether the paradigm of inquiry and GT approach adopted in the research support follow-on verificational studies and quantitative testing of the hypotheses and theory. The results of this review produced five major implications for theory development which are summarized below. Implications for developing and advancing GT in logistics and supply chain management: (1) Continued methodological slurring due to confusion between GT and qualitative methods. (2) Perceived lack of credibility and rigor of GT within the discipline. (3) Grounded theories not accepted by researchers in other disciplines. (4) Challenging review process resulting from a lack of methodological standards. (5) Difficulty in obtaining acceptance in top journals. 3. A reviewer’s perspective on theory development and adoption of a consistent paradigm of inquiry An initial task for reviewers is to determine whether: GT is appropriate, the authors have adopted an approach consistent with their paradigm of inquiry, and the selected approach has been consistently applied throughout the research. Appropriateness of GT GT can be defined as “a systematic approach to qualitative research that facilitates theoretical abstraction from field data through a focus on comparative analysis” (Mello and Flint, 2009, p. 109). The objective is to develop “an abstract theoretical understanding of the studied experience” (Charmaz, 2006, p. 4). The role of the researcher is not to provide a perfect description of an area but to develop a theory that accounts for much of the relevant behavior. The resulting theory may “be presented as a well-codified set of propositions or in a running theoretical discussion, using conceptual categories and their properties” (Glaser and Strauss, 1967, p. 31). Reviewers should anticipate the emergence of a substantive or formal theory. GT typically produces substantive, or mid-range theory. Substantive theories initially may

Grounded theory methodology

789

IJPDLM 42,8/9

be developed through comparative analysis within the same substantive or empirical area to fill a gap in our theoretical understanding of a phenomenon. The research could develop a formal theory where comparative analysis occurs among several different substantive areas. Reviewers should note that the focus is not on achieving generalizability but on increasing the ability to explain the phenomena under study within and across substantive areas.

790 Systematic approach and consistency with paradigm of inquiry GT employs a systematic approach for analyzing qualitative and quantitative data. A manuscript should describe a series of steps leading to the development of a theory. The GT approach initially developed by Glaser and Strauss (1967) responded to criticism that qualitative research was not methodologically sound to produce credible results. Reviewers can preempt such criticism by ensuring the manuscript contains an explanation of the approach and how theoretical insights emerged from the data. Reviewers must encourage, guide, and require authors to go beyond providing a “neat and tidy” descriptive account of the phenomena to conceptualizing a theory that captures the theoretical “whole” of a basic social process including stages, properties, conditions, and consequences (Glaser, 1978, p. 11). Reviewers should be wary of a “sleight of hand” that produces little insight into the analytical approach, does not add significantly to our understanding of the phenomena under investigation, relies on description, and invites the reviewer to take it on trust that theory somehow emerged from the data without providing an adequate explanation of development of theoretical insights (Barbour, 2001). GT methodology continues to evolve since the initial publication of The Discovery of Grounded Theory. One philosophical paradigm is emergent (following Glaser), and the other is more formulaic and directive (following Strauss). Other variations such as the constructivist approach by Charmaz have also emerged. A reviewer must assess whether the use of GT is congruent with the paradigm of inquiry adopted in the manuscript (Annells, 1996) because divergent philosophical approaches to GT have implications for the conduct of the research in terms of process as well as outcomes of research (Mello and Flint, 2009; Annells, 1996; Walker and Myrick, 2006; Goulding, 2002; Kelle, 2005; Duchscher and Morgan, 2004). Guba and Lincoln (1994) identified four paradigms of inquiry: positivisim, postpositivism, critical theory, and constructivisim. These divergent paradigms guide the interaction of the researcher with the data, introduction of outside concepts and conditions, and the use of subsequent or sequential verification studies. The “classic” school of GT methodology advocated by Glaser has frequently been identified as postpositivist due to its roots in social interactionism and an objectivist orientation (Annells, 1996; Goulding, 2002). Reviewers should find the researchers remain separate from the method, do not introduce outside concepts and constructs, and strictly adhere to the premise that theory should be allowed to emerge from the data. This paradigm accepts subsequent verificational research which is consistent with Glaser (1992) – GT “should be seen in sequential relation” to verificational research with the ultimate aim of both being the “building up of scientific facts”. The key implications for reviewers when evaluating manuscripts adopting classic GT include ensuring that the researchers remained objective and outside the method, allowed the theory to emerge from the data, did not interact with the data by introducing outside concepts or constructs, and accept sequential verificational research.

The approach to GT advocated by Strauss and Corbin (1990, 1998) falls in the constructivist paradigm of inquiry but contains remnants of postpositivism (Annells, 1996). The reviewer should find the researcher interprets and interacts with the data to create knowledge. This approach imposes a coding paradigm to explain the construction of a framework of grounded categories (Kelle, 2005). A conditional matrix is employed which is likely to introduce issues such as class, gender, race, and power into the analysis (Annells, 1996). Verification is not assumed to be possible only through follow-on quantitative analysis but instead can occur through constant comparative data analysis throughout the course of the research (Strauss and Corbin, 1994). The key implications for reviewers when evaluating manuscripts adopting this approach are: the researchers became participants in the research by interacting and interpreting the data; the research introduces outside conditions, concepts, or constructs into the analysis to assist in constructing a theory; and verification may occur through constant comparative analysis. In summary, the reviewer needs to determine whether the GT method employed is consistently applied, and is consistent with the paradigm of inquiry and the philosophical approach adopted in the research. Reviewers will find few differences in the methodology; a few differences were discussed in the previous paragraph (Annells, 1996; Goulding, 2002; Charmaz, 2004 for detailed explanations of these differences). However, the paradigm of learning adopted in the paper will affect the analysis and interpretation of the data, whether the research and resulting GT take an objectivist, positivist stance or a constructivist, interactionist stance, and whether subsequent verificational research is recommended for future research. 4. Major steps in a GT development This section describes the major steps within a GT analysis and actions for each step that should be considered when evaluating a manuscript. The eight steps (Figure 1) are presented in a linear fashion; however, GT proceeds in a non-linear, iterative manner. Step 1: explaining a phenomenon Reviewers should consider why existing theory does not adequately explain a phenomenon. The authors may have reviewed the literature to reveal a lack of relevant theory, the inability of existing theory to fully explain the phenomenon, or the need to develop formal theory from previous substantive theories. Some manuscripts do not contain a literature review. A common misconception is researchers should enter the field devoid of any knowledge (Suddaby, 2006) to minimize preconceptions and bias. The leading grounded researchers have largely debunked this notion (Flint et al., 2005; Charmaz, 2004; Goulding, 2002) and suggest that authors understand the substantive area and can explain and interpret it in their work (Glaser, 1978). Theoretical sensitivity refers to the ability of a researcher to work with the data in both theoretical and sensitive ways (Glaser, 1978). Reviewers need to assess the ability of the authors to “theoretically and conceptually think about the data from a distance, while simultaneously maintaining an in-close level of sensitivity and understanding about the process and their involvement in that process” (Walker and Myrick, 2006, p. 552). Reviewers can take several cues to assess the theoretical sensitivity of the authors. The phrasing of the problem as a general, broad statement that leaves open possible theoretical outcomes suggests the authors understand the substantive area but have left possible alternative explanations open for discovery. The research questions

Grounded theory methodology

791

IJPDLM 42,8/9

792

Figure 1. Major steps in a grounded theory development

should identify the need for theory, further explanation, or interpretation in a substantive area. The authors must avoid verifying existing theory or hypotheses during the research. Step 2: establishing the appropriateness of GT methodology The goal of GT is to conduct research, grounded in empirical data, which increases our understanding of a phenomenon by developing theory with explanatory and predictive power (Glaser and Strauss, 1967, p. 24). Therefore, the justification should indicate an attempt to conceptualize how or why the phenomenon occurs. The reviewer must ensure that the justification goes beyond simply stating that GT was employed due to the exploratory nature of the research, that qualitative research was necessary due to a gap in extant research, or that authors adopted GT to describe a phenomena in detail. Fischer and Otnes (2006) identified four areas that support the use and goals of GT: investigating questions about the nature of a new construct; raising questions about the adequacy of a previously well-established construct; investigating previously unrecognized facilitators or implications of a construct; and addressing questions about the adequacy of prior conceptualizations of facilitators or implications of a construct.

Step 3: selecting a context The reviewers need to consider the context employed within which the phenomenon is investigated. “A context is the set of conditions that give rise to the problems or circumstances to which individuals respond by means of action/interaction/emotions” (Corbin and Strauss, 2008, p. 229). Some practical considerations may guide the selection of context such as access to: willing companies, companies that can provide rich data, and knowledgeable people that are experiencing or have experienced the phenomenon. While convenience factors are acceptable, the reviewers must primarily look for the rationale for fit between the phenomenon, research questions, and the context. Step 4: selecting sources of data Reviewers should detect clear evidence of a multi-step, iterative process for data collection and sample selection. The categories and relationships that will emerge as the research progresses cannot be known in advance. Therefore, the criteria for collecting data during initial sampling depend upon the research questions posed before the investigation commences. This approach is frequently identified to as “purposive sampling” and refers to the calculated decision to sample a specific set of events, settings, or people based on criteria that are decided in advance. Reviewers should check the adequacy of the initial sample by asking questions such as whether the initial participants: fit the context, had visibility over part or the entire phenomenon, were knowledgeable, willing to participate, and experienced with, and engaged in the phenomenon being studied. To judge the reliability of data, reviewers may pose questions regarding how access was gained to obtain the appropriate data for the research. Reviewers must be convinced theoretical sampling was employed for subsequent data collection. Theoretical sampling is the: [. . .] process of data collection where the analyst collects codes, and analyzes the data and decides what data to collect next and where to find them based on the emergent theory (Mello and Flint, 2009, p. 112).

Evidence that coding and constant comparative analysis guided the authors in determining the participants or data sources that were subsequently selected must be sought. Manuscripts often report the number of interviews or data sources used in the research which may indicate the authors did not go beyond an initial sample. A description of how sources were initially and subsequently selected during the research process is as important as reporting the sample size and composition. As codes and categories are developed, the research should reflect their theoretical development with properties and connections to other categories until each category is saturated. Theoretical sampling continues until each category is saturated, elaborated, and integrated into an emerging theory (Glaser, 1992). When a category is saturated, constant comparison will not reveal new properties or dimensions for the category or relationships between categories. Additional information may be sought to fill in gaps in the emerging theory or to further understand differences detected between incidents. Research may employ theoretical sampling numerous times (see Step 7) during the investigation as one category is saturated and the research progresses to saturate other categories or exploring the relationships between them. Reviewers should be wary of the four common mistakes related to use of theoretical sampling: sampling to address initial research questions, sampling to reflect population

Grounded theory methodology

793

IJPDLM 42,8/9

794

distributions, sampling to find negative cases, and sampling until no new data emerge (Charmaz, 2009, p. 100). Initial sampling represents a point of departure; whereas, theoretical sampling is used to saturate categories or concepts for theoretical refinement and elaboration. Sampling to reflect population distributions implies verification or hypothesis testing. Theoretical sampling, on the other hand, focuses on increasing explanatory power and our understanding of a phenomenon. Sampling to obtain negative cases may be inconsistent with GT. When negative cases emerge in the data, theoretical sampling may be applied to understand the negative cases with the resulting theory being modified to incorporate these cases. The reviewer should challenge instances where negative cases are sought out or artificially imported into the analysis. Finally, many authors confuse theoretical sampling with data gathering (Charmaz, 2004, p. 102) and do not satisfactorily achieve theoretical saturation. Step 5: developing the interview protocol The reviewer should check whether the researcher developed an interview protocol or observation guide. The initial set of questions may be based on insights derived from literature, experience, or preliminary field work and should typically be broad and general in scope and reflect authors’ theoretical sensitivity. Since the objective is to emerge a theory from data, reviewers need to look for evidence that these initial questions were reworded, refined, or even discarded over the course of the research (see revision loop in Figure 1). The interview protocol should be open-ended, unstructured, and allow the authors to probe new ideas as they emerge. The interview questions should avoid the forcing of preconceived theories or concepts. Reviewers should find the questions were initially broad but became more focused as the research progressed and theoretical sampling was employed. Reviewers should look for the logic of modifying the interview protocol and find evidence that the interview protocol was modified in order to refine constructs, develop theory, or to increase the depth of understanding of the phenomenon. Step 6: data collection and coding Data collection and coding represent the first stage of the analytic process within GT. The manuscript should explain how data will be captured. If the interview protocol indicates that the permission of interview participants to record the interview was obtained or other sources of data were captured for future use, then it is easier to establish that data can be accurately captured and revisited, and if needed, multiple researchers can review the same data. If the data is not captured in its entirety, then reviewers need confirm how the accuracy of data capture was maintained. Comparative analysis, or constant comparison, represents a core tenet of GT. Comparisons, both within and between incidents and with new and previous data, should have been used to identify similarities, differences, trends and patterns in the data as well as to make modifications of the emerging theory. These comparisons support the generation of abstract categories and properties which can be used to develop a theory explaining the behavior under observation (Glaser and Strauss, 1967). Although presented in a sequential manner, reviewers should be able to determine whether the research actually iterated numerous times across multiple steps (Figure 2). Reviewers need to consider several points when evaluating the coding, memoing, and theoretical sampling phase of the methodology. First, they must determine

Grounded theory methodology Theoretical Sampling

Subsequent steps

795

Memoing/ constant comparison Open/Initial Coding

whether coding occurred at a descriptive or conceptual level. Descriptive coding simply describes a phenomenon and will provide little to no explanatory power. Conceptual coding and the resulting categories will reflect greater theoretical insight such as why or how the phenomenon is occurring. Second, the use of constant comparative analysis should appear in the description of how new codes were created, the formation of selective or theoretical coding, analyses captured in theoretical memos, the discovery of categories and properties, the resulting relationships between categories, and theory development. Reviewers will need to determine how the authors employed theoretical memos to raise their coding to a conceptual level and to emerge categories, to identify tentative relationships between them and to identify the properties and dimensions of each category. Third, constant comparison must be used to guide theoretical sampling. Reviewers will need to determine whether the authors used the data to guide theoretical sampling to obtain relevant data that would increase their insight and understanding. Finally, the constant comparison of data within the research should reflect a focused flexibility to keep transcending the analyses until the problem is theoretically saturated (Glaser, 1978). The reviewer must be convinced that constant comparison and the analysis of data occurred simultaneously and throughout the investigation process. Step 7: diagramming concepts, integrating memos, and developing theory The next phase of the GT methodology focuses on emerging theory from the codes, categories, and memos developed during the previous phase of research. Theoretical saturation represents the stopping point for data collection and the “sample size” for the research. Reviewers should not accept statements that saturation was achieved after accomplishing an arbitrary number of interviews. For example, “Theoretical saturation was achieved after 17 interviews.” The research should report how saturation was reached and not rely on arbitrary numbers, frequency counts, or citing repetition in the data.

Figure 2. Iterative nature of the grounded theory methodology

IJPDLM 42,8/9

796

Saturation occurs when constant comparison emerges no new properties or dimensions for each of the categories (Holton, 2010). “Saturation is not seeing the pattern over and over again. It is the conceptualization of these incidents which yields different properties of the pattern until no new properties emerge” (Glaser, 2001, p. 191). Unfortunately, no published guidelines or tests of adequacy exist for estimating the sample size required to reach saturation. Despite the lack of guidelines, the reviewers should ensure that the process used to reach saturation was rigorous, thorough and transparent. The authors “should specify what they did and how they did it, including how they handled data condensation and interpretation” (Bowen, 2008, p. 149). Memo, or theoretical, sorting represents another process common to the different forms of GT. During this process, the reviewers should learn how the researchers created and defined links between categories (Charmaz, 2006). Sorting frequently produces more memos of a higher conceptual level that aid in producing a generalized, integrated model by which to write the theory. Reviewers may find that the relevant literature was integrated into the theory (Glaser, 1978). As part of their evaluation, reviewers should seek to understand how the authors used sorting to develop relationships between categories and used the memos to produce a theoretical framework and generate an integrated, dense and complex theory. Reviewers will confront very different methods in integrating the categories into a final GT. Glaser (1978) suggests the use of theoretical coding to conceptualize how the categories relate to each other and may be integrated into a theory. To assist in this effort, he developed 18 theoretical coding families, but others may exist. These theoretical codes assist in developing a coherent, analytical story and help move it in a theoretical direction. The theoretical categories are eventually subsumed into a core category that: [. . .] pulls together all the strands in order to offer an explanation of the behavior under study. This is usually done when the theory is written and integrated with existing theories to show relevance and perspective (Goulding, 2002, p. 88).

Strauss and Corbin (1990, 1994) take a very different approach for integrating the final theory. They employ axial coding and a conditional matrix (Strauss and Corbin, 1990), or a conditional/consequential matrix (Corbin and Strauss, 2008), to enable the researcher to become theoretically sensitive in systematically relating conditions, action/interaction and consequences to a phenomenon. The matrix is operationalized by tracing conditional paths through it. Tracing paths of incidents from the level of action/interaction through the various conditional levels to consequences is used to determine how categories relate to one another. Step 8: evaluating GT research Finally, the reviewers should check whether the authors established the credibility and trustworthiness of the study. Several sets of criteria may be employed to evaluate the quality of a GT research. Charmaz (2006) suggests four criteria: credibility, originality, resonance, and usefulness. Glaser (1978) suggests fit, work, relevance, and modifiability. Flint et al. (2002) employ five criteria from interpretive research – credibility, transferability, dependability, confirmability, and integrity (based on Hirschman, 1986; Lincoln and Guba, 1985; Wallendorf and Belk, 1989) and four from GT – fit, understanding, generality, and control (based on Strauss and Corbin, 1990).

The reviewers should look for an explicit description of how the authors addressed at least one set of criteria. This set of criteria, in turn, must fit with the GT approach adopted by the authors. The final product of a GT approach is a theory or theoretical framework. The reviewers need to determine whether the author has successfully explained how the theory is grounded in the data, that is, the methodology supports the development of theory. The reviewers must be convinced that the theory or theoretical framework logically flows from the research effort. In terms of contribution to the body of knowledge, the reviewer should look for clear, meaningful, and significant theoretical contribution within the context in which the theory is proposed. Reviewers need to establish that the authors compared the theory to the extant literature to establish if the proposed theory is different or the proposed theory assists in our understanding of the literature. Any differences need to be described or explained. Reviewers should be convinced that the proposed theory increases our understanding of the phenomenon and provides hypotheses or propositions from the theory to guide future research. The reviewers must also look for managerial relevance and usefulness. Table II summarizes the key considerations for each step. This list of questions can be used by the reviewer to ensure the methodology has been appropriately executed. Reviewers may also find the table useful in making a comprehensive evaluation of the research and the overall contribution. 5. Conclusions GT can play an important role in the development of theory grounded in logistics and supply chain management, but the methodology continues to confront several hurdles which must be overcome. A review of 33 articles in the leading journals detected problems similar to those experienced in other disciplines: the lack of a clear distinction between GT and qualitative analyses, methodological slurring, unclear or missing descriptions of key methodological steps, a missing focus on theory development, and resulting perceptions of a lack of credibility and rigor. However, several articles demonstrated effective execution of the methodology and can guide reviewers in their assessment of manuscripts. Reviewers can play a critical role in ensuring the rigor and credibility of published articles. However, reviewers confront several challenges, including making inferences in light of incomplete details on execution of research, deciding what to include while trying to meet the journal space constraints, and keeping pace with the evolution of the methodology. The framework put forth can serve as a guidepost to assist reviewers in two ways. First, the framework provides the two key characteristics of GT methodology which reviewers can use for assessing whether congruency exists between the research objectives, theory development approach, and the paradigm of inquiry adopted in the research. This discussion sensitizes the reviewers to the iterative, unstructured and fluid nature of the research methodology, such as evolving (rather than fixed) research questions, continually changing (rather than preset) interview protocol, and theoretical sampling (instead of a pre-determined sample). Second, the framework provides an eight step checklist for reviewers to use for assessing the execution and rigor of the GT methodology. Each step of the framework is elaborated in terms of what the reviewers should consider when evaluating the quality of a GT submission. Reviewers should avoid imposing or advocating

Grounded theory methodology

797

Table II. A reviewer’s guide to evaluating the quality of GT research

Does the author identify a process or phenomenon that will be examined and explain why theory is needed to increase our understanding? Does the author describe the lack of an extant theory capable of explaining the phenomenon? Does the author explain that the purpose is to generate a theory rather than achieve generalizability? Does the author address theoretical sensitivity? How does the author identify the theoretical issues and explain how the research will remain open to the emergence of new constructs, relationships and theory? Does the phrasing of the problem statement indicate that the researchers have remained open for discovery? Is the problem statement phrased to indicate the need for a theory? Step 2: establishing Does the author provide justification for GT beyond the need to perform an exploratory study? appropriateness of GT Does the author explain which approach within the GT methodology is accepted? methodology Are the key implications of adopting an approach within the GT methodology consistent with the approach in terms of author’s interaction with the data, introduction of outside concepts and conditions, and the use of subsequent or sequential verification studies? Step 3: selecting a context Does the author explain why the context is appropriate for the purpose of the research? Is the context selected appropriate for building theory of the phenomenon or process being investigated? Step 4: selecting sources Does the author explain how the access will be gained to obtain the appropriate data for the research – , i.e. access to key personnel, data, observation points, of data etc. to conduct the research? Does the author explain how the initial participants were selected? Does the researcher indicate that the sample of firms/participants was identified in advance or was based on data collection? Does the author provide evidence that coding and constant comparative analysis guided the author in determining the follow-on sources or participants that were subsequently selected? Does the author indicate that theoretical sampling was continued until saturation was achieved? Does the author indicate that the purpose of theoretical sampling was to achieve more explanatory power and not to achieve generalizability of the results? Step 5: developing the Are the interview questions phrased to avoid the forcing of preconceived theories or concepts? interview protocol Do the interview questions allow the theory or framework to emerge from, or be grounded in, the data collected during the investigation? Is the interview protocol open-ended and structured in that it allows the author to probe new ideas as they emerge? Does the author indicate that interview protocol was modified to include more focused questions as interview progressed? Does the author provide evidence that interview protocol was modified to refine constructs, develop theory, or to increase the depth of understanding of the phenomenon? Step 6: data collection How is the author capturing data? Are the interviews taped and recorded? Are the interviews transcribed? and coding Is the author employing any tools to assist in the research such as MaxQDA? Does the author explain how the initial coding was performed? Does coding occur while the data collection is being conducted? How do the codes evolve over time? Does the author indicate how memos were used to capture ideas, potential categories, relationships, differences, or similarities in the data? How is constant comparison incorporated into the analysis of the data? (continued)

Questions to ask

798

Step 1: describing the phenomenon

Step

IJPDLM 42,8/9

Step 8: evaluating GT research

Step 7: diagramming concepts, integrating memos, and developing theory

Step

Do the authors provide evidence/examples that codes emerge from the data? Do authors provide evidence that codes are evolving as constant comparisons are made? What does the author report regarding the use of comparisons to refine categories and obtain a richer understanding of the phenomenon under investigation? How are memos used to identify themes that emerge from the data? How does the author report gaps in the investigation? How are the gaps targeted for follow-on investigation, or theoretical sampling? Theoretical saturation How does the author explain theoretical saturation? Does the author go beyond stating that “nothing new” was found and explain that the categories were determined to be saturated when no new theoretical insights or no additional information was obtained regarding the properties of the core categories? Does the author provide examples of how saturation was determined to have been achieved? Theoretical sorting Does the author explain how a theory or theoretical framework was accomplished through sorting, diagramming, or integrating theoretical memos developed during the research? How is the result of this process integrated into a theory or framework? Is the process of integrating the categories into a final GT consistent with the approach adopted (Glaserian, Straussian or any other)? Developing theory Does the author indicate whether theoretical sampling was used to enrich categories, to fill gaps in understanding, or to explore leads identified during the research? Does the author explain how focused coding occurred and how memos were used to develop the focused codes? How are conceptual categories identified and refined during the research? Is evidence from the research provided to support ideas or patterns? How does the author evaluate the research? Does the author use criteria such as Charmaz (credibility, originality, resonance, usefulness), Glaser (work, relevance, modifiability), Flint (credibility, transferability, dependability, confirmability, integrity), or Strauss and Corbin (fit, understanding, generality, control)? Does the author establish credibility with the reviewer? Does the author provide a theory or theoretical framework? Does the author explain the scope, based on the context, for the theory? Has the author successfully explained how the theory is grounded in the data – , i.e. how the investigation supports the development of theory? Does the theory or theoretical framework logically flow from the research effort? Was member checking, or some other technique, used to confirm that the author’s analysis is consistent with the data (participants)? Does the researcher convincingly and clearly establish theoretical contribution? Does the author compare the theory to the extant literature? How does the theory assist in our understanding of the literature? How is this theory different? Why any differences? How does the theory increase our understanding of the phenomenon? Does the author propose any hypotheses or propositions from the theory to guide future research? Does the research establish managerial relevance and usefulness?

Questions to ask

Grounded theory methodology

799

Table II.

IJPDLM 42,8/9

800

an approach when reviewing a manuscript. Instead, they should ensure the authors have clearly identified the approach adopted for research, recognize the implications in their analysis and theory development, and remain consistent with the version of GT methodology employed. Authors may err on the side of providing too many details than risk rejection if the manuscript is perceived as not being rigorous. Or, authors may not include adequate details because of space constraints. Reviewers should be prepared for either approach, to establish a dialog with authors, and to make recommendations on striking a balance between parsimony and credibility in the final version of the manuscript approved for publication. Limitations The framework is not a substitute for in-depth knowledge of GT methodology. Readers may refer to the references section for well-respected and widely accepted sources of knowledge on the methodology. The contribution of this paper is that for researchers educated in the methodology, it provides a comprehensive set of considerations for evaluating a manuscript. For reviewers with limited knowledge, the framework provides a fast and easy reference to start the process of learning about the methodology. Contributions More effective reviews of GT research are expected to help the body of knowledge in three ways. First, reviewers aware of critical methodological issues will act as gatekeepers ensuring that only credible and high quality research is published. This awareness will also prevent qualitative research being evaluated against inadequate, or even worse, incorrect criteria. The discussion on consistency among the paradigm of inquiry, philosophical approach, and method of GT provides a foundation to evaluate GT research. These considerations may lead to higher rigor with which the GT methodology is evaluated and increase the credibility of GT as a valid methodology. Second, reviewers who are skeptical of the rigor of GT methodology can learn about the methodology from the discussion and use the checklist to establish the rigor of a GT study rather than rejecting or discarding research. Third, the discussion and examples provided in this paper and the table could be used to provide thorough and constructive reviews that will enable researchers to improve the quality of their current and subsequent research efforts. Summary The execution of GT research in logistics and supply chain management has yet to receive wide-spread acceptance. The method continues to encounter skepticism due to a perceived lack of rigor and credible findings. However, well-executed GT research can make significant contributions to the discipline through theory development and increasing our understanding and ability to explain key phenomenon. Through a combination of an extensive review of the method and leading GT theory research published in the leading journals, a framework was developed to assist reviewers. By increasing their expertise and focusing their attention on key steps and considerations within the GT method, the framework can assist reviewers in their critique and of GT research. A more robust and thorough review of GT manuscripts will not only increase rigor and credibility but advance the development of theory grounded in logistics and supply chain data and practice.

References Annells, M. (1996), “Grounded theory method: philosophical perspectives, paradigm of inquiry, and postmodernism”, Qualitative Health Research, Vol. 6 No. 3, pp. 379-93. Barbour, R.S. (2001), “Checklists for improving rigour in qualitative research: a case of the tail wagging the dog?”, British Medical Journal, Vol. 322 No. 7294, pp. 1115-17. Bowen, G.A. (2008), “Naturalistic inquiry and the saturation concept: a research note”, Qualitative Research, Vol. 8 No. 1, pp. 137-52. Carter, C.R., Easton, P., Vellenga, D. and Allen, B.J. (2009), “Affiliation of authors in transportation and logistics academic journals: a reevaluation”, Transportation Journal, Vol. 48 No. 1, p. 42. Charmaz, K. (2004), “Premises, principles, and practices in qualitative research: revisiting the foundation”, Qualitative Health Research, Vol. 14 No. 7, pp. 976-93. Charmaz, K. (2006), Constructing Grounded Theory, Sage, Thousand Oaks, CA. Charmaz, K. (2009), “Shifting the grounds: constructivist grounded theory methods”, in Morse, J.M., Stern, P.N., Corbin, J.M., Bowers, B. and Clarke, A.E. (Eds), Developing Grounded Theory: The Second Generation, University of Arizona Press, Walnut Creek, CA. Corbin, J. and Strauss, A. (2008), Basics of Qualitative Research, 3rd ed., Sage, Thousand Oaks, CA. Davis-Sramek, B. and Fugate, B.S. (2007), “State of logistics: a visionary perspective”, Journal of Business Logistics, Vol. 28 No. 2, pp. 1-34. Duchscher, J.E.B. and Morgan, D. (2004), “Grounded theory: reflections on the emergence vs. forcing debate”, Journal of Advanced Nursing, Vol. 48 No. 6, pp. 605-12. Fischer, E. and Otnes, C.C. (2006), “Breaking new ground: developing grounded theories in marketing and consumer behavior”, in Belk, R.W. (Ed.), Handbook of Qualitative Research Methods in Marketing, Edward Elgar, Northampton, MA, pp. 19-30. Flint, D.J., Woodruff, R.B. and Gardial, S.F. (2002), “Exploring the phenomenon of customers’ desired value change in a business-to-business context”, Journal of Marketing, Vol. 66 No. 4, pp. 102-17. Flint, D.J., Larson, E., Gammelgaard, B. and Mentzer, J.T. (2005), “Logistics innovation: a customer value-oriented social process”, Journal of Business Logistics, Vol. 26 No. 1, pp. 113-48. Gibson, B.J. and Hanna, J.B. (2003), “Periodical usefulness: the US logistics educator perspective”, Journal of Business Logistics, Vol. 24 No. 1, pp. 221-40. Glaser, B.G. (1978), Theoretical Sensitivity, The Sociology Press, Mill Valley, CA. Glaser, B.G. (1992), Basics of Grounded Theory Analysis, The Sociology Press, Mill Valley, CA. Glaser, B.G. (2001), The Grounded Theory Perspective: Conceptualization Contrasted with Description, The Sociology Press, Mill Valley, CA. Glaser, B.G. and Strauss, A.L. (1967), The Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine Transaction, Piscataway, NJ. Goulding, C. (2002), Grounded Theory: A Practical Guide for Management, Business, and Market Researchers, Sage, Thousand Oaks, CA. Guba, E.G. and Lincoln, Y.S. (1994), “Competing paradigms in qualitative research”, in Denzin, N.K. and Lincoln, Y.S. (Eds), Handbook of Qualitative Research, Sage, Thousand Oaks, CA, pp. 105-17.

Grounded theory methodology

801

IJPDLM 42,8/9

802

Hirschman, E.C. (1986), “Humanistic inquiry in marketing research: philosophy, method, and criteria”, Journal of Marketing Research, Vol. 23 No. 3, pp. 237-49. Holton, J.A. (2010), “The coding process and its challenges”, The Grounded Theory Review, Vol. 9 No. 1, pp. 21-40. Jones, R. and Noble, G. (2007), “Grounded theory and management research: a lack of integrity?”, Qualitative Research in Organizations, Vol. 2 No. 2, pp. 84-103. Kelle, U. (2005), “Emergence vs. forcing of empirical data? A crucial problem of grounded theory reconsidered”, Forum Qualitative Sozialforschung (Forum: Qualitative Social Research), Vol. 6 No. 2, Art. 27, available at: www.qualitative-research.net/index.php/fqs/article/view/ 467/1001 (accessed August 30, 2011). Lincoln, Y.S. and Guba, E.G. (1985), Naturalistic Inquiry, Sage, Newbury Park, CA. Manuj, I. and Sahin, F. (2011), “A model of supply chain and supply chain decision-making complexity”, International Journal of Physical Distribution & Logistics Management, Vol. 41 No. 5, pp. 511-49. Mello, J. and Flint, D.J. (2009), “A refined view of grounded theory and its application to logistics research”, Journal of Business Logistics, Vol. 30 No. 1, pp. 107-25. Mello, J.E., Stank, T.P. and Esper, T.L. (2008), “A model of logistics outsourcing strategy”, Transportation Journal, Vol. 47 No. 4, pp. 5-25. Menachof, D.A., Gibson, B.J., Hanna, J.B. and Whiteing, A.E. (2009), “An analysis of the value of supply chain management periodicals”, International Journal of Physical Distribution & Logistics Management, Vol. 39 No. 2, pp. 145-66. Mollenkopf, D., Russo, I. and Frankel, R. (2007), “The returns management process in supply chain strategy”, International Journal of Physical Distribution & Logistics Management, Vol. 37 No. 7, pp. 568-92. Randall, W.S., Pohlen, T.L. and Hanna, J.B. (2010), “Evolving a theory of performance-based logistics using insights from service dominant logic”, Journal of Business Logistics, Vol. 31 No. 2, pp. 35-61. Strauss, A. and Corbin, J. (1990), Basics of Qualitative Research: Grounded Theory Procedures and Techniques, Sage, Newbury Park, CA. Strauss, A. and Corbin, J. (1994), “Grounded theory methodology: an overview”, in Denzin, N.K. and Lincoln, Y.S. (Eds), Handbook of Qualitative Research, Sage, Thousand Oaks, CA, pp. 273-85. Strauss, A. and Corbin, J. (1998), Basics of Qualitative Research: Grounded Theory Procedures and Techniques, 2nd ed., Sage, Thousand Oaks, CA. Suddaby, R. (2006), “What grounded theory is not”, Academy of Management Journal, Vol. 49 No. 4, pp. 633-42. Walker, D. and Myrick, F. (2006), “Grounded theory: an exploration of process and procedure”, Qualitative Health Research, Vol. 16 No. 4, pp. 547-59. Wallendorf, M. and Belk, R.W. (1989), “Assessing trustworthiness in naturalistic consumer research”, in Hirschman, E. (Ed.), Interpretive Consumer Research, Association for Consumer Research, Provo, UT, pp. 69-84. About the authors Ila Manuj, PhD is an Assistant Professor of Logistics at the Department of Marketing and Logistics, University of North Texas. Her research interests are in the areas of risk and complexity management in global supply chains. She has published in Journal of Business Logistics, International Journal of Physical Distribution & Logistics Management, International Journal of Logistics Management, and Transportation Journal, co-authored several book

chapters, and presented at several national and international conferences. Ila Manuj is the corresponding author and can be contacted at: [email protected] Terrance L. Pohlen, PhD, is an Associate Professor and Director, Center for Logistics Education and Research, University of North Texas. His research focuses on the costing and financial management of logistics and supply chain management. Recently, he co-authored CSCMP’s The Handbook of Supply Chain Costing. His research has been published in the Journal of Business Logistics, International Journal of Logistics Management, International Journal of Physical Distribution & Logistics Management, and the Transportation Journal. He chairs the professional certification in transportation and logistics and is an AST&L board member.

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Grounded theory methodology

803

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0960-0035.htm

IJPDLM 42,8/9

Rigor in qualitative supply chain management research Lessons from applying repertory grid technique

804

Keith Goffin, Jawwad Z. Raja, Bjo¨rn Claes, Marek Szwejczewski and Veronica Martinez

Received 29 November 2011 Accepted 16 December 2011

Cranfield School of Management, Cranfield, UK Abstract Purpose – The purpose of this paper is to share the authors’ experiences of using the repertory grid technique in two supply chain management studies. The paper aims to demonstrate how the two studies provided insights into how qualitative techniques such as the repertory grid can be made more rigorous than in the past, and how results can be generated that are inaccessible using quantitative methods. Design/methodology/approach – This paper presents two studies undertaken using the repertory grid technique to illustrate its application in supply chain management research. Findings – The paper presents insights into supply chain research that otherwise would not have emerged using traditional methods. Both studies derive a comprehensive list of empirical categories of constructs, many of which have not been identified in the extant literature. Moreover, the technique demonstrates that frequently-mentioned constructs are not necessarily the most important. Research limitations/implications – The paper demonstrates how quantitative calculations can strengthen qualitative research. Importantly, from the authors’ experience of using the technique the paper details how to focus on demonstrating validity, reliability, and theoretical saturation. Originality/value – It is the authors’ contention that the addition of the repertory grid technique to the toolset of methods used by logistics and supply chain management researchers can only enhance insights and the building of robust theories. Qualitative studies that adopt the technique cannot only provide rich insights, but also counter the common criticism aimed at qualitative research – that of failing to provide clear and transparent accounts of the analysis process and how findings are generated from the data set. Keywords Repertory grid technique, Qualitative methods, Manufacturer-supplier relationships, Servitization, Supply chain management, Supplier relations Paper type Research paper

1. Introduction To further our knowledge of supply chain management (SCM), research needs to be based on a balance of quantitative and qualitative approaches (Boyer and Swink, 2008; Burgess et al., 2006) and more qualitative studies are needed (New, 1997; Na¨slund, 2002; Voss et al., 2002). For qualitative methods to be accepted, researchers need to demonstrate the degree of rigor that leading journals – including IJPDLM – rightly International Journal of Physical Distribution & Logistics Management Vol. 42 No. 8/9, 2012 pp. 804-827 q Emerald Group Publishing Limited 0960-0035 DOI 10.1108/09600031211269767

The authors would like to acknowledge the support of the EPSRC/IMRC under grant number [IMRC 144], which is supporting Product Service Systems research at Cranfield University. They would also like to thank Dorota Bourne, the special issue editors and three anonymous referees for their insightful comments.

demand (Goodyear, 1990; Knemeyer et al., 2001; Seuring, 2008). Rigor is just as important in qualitative as in quantitative studies (Drucker-Godard et al., 2001) and the classic texts on qualitative research stress the importance of validity and reliability (Miles and Huberman, 1994; Yin, 2009). Therefore, how qualitative methods can be applied rigorously to SCM research is a pertinent and important issue. This paper describes two studies, which used an innovative qualitative approach to investigate important SCM issues: the relationships between manufacturers and suppliers; and perceptions of customer support. Case study methodology was used, including interviews based on repertory grid technique (RGT), which helps respondents articulate their views on complex topics. Although we emphasize a specific technique in this paper, our discussion demonstrates how qualitative methods can be conducted in a rigorous way. We also show how such approaches produce results that cannot be obtained using quantitative methods. Readers should note that we will not present the full findings of our two studies (but full references are given to earlier publications). This paper is focused on the importance of methodological rigor in qualitative research. The paper is structured as follows: we discuss rigor, validity and reliability in qualitative research in Section 2. This is followed by a brief overview of RGT (Section 3). We then discuss our first study, which investigated manufacturer-supplier relationships, in Section 4. We explain the aims of the study, how we applied RGT, and the key results for supply chain researchers and scholars of methodology. Section 5 describes our second study, in the emerging field of servitization – how manufacturers are increasingly focusing on service offerings to complement their products. Section 6 compares and contrasts the two studies: we discuss how reliability and validity can be enhanced in qualitative studies, including how theoretical saturation can be demonstrated using a Pareto analysis. The final section summarizes the implications for SCM researchers. 2. Rigor, validity and reliability The rigor vs relevance debate has been a recurring theme in management studies over the last decade (Starkey and Madan, 2001; Vermeulen, 2005; Gulati, 2007; Palmer et al., 2009; Polzer et al., 2009) and SCM researchers have also entered this discussion (Flynn, 2008; Mentzer, 2008). Although research should be relevant, it has been stressed that relevance must not come at the expense of rigor (Bartunek et al., 2006). It has been argued that “high quality research must use the most rigorous research methods possible” (Flynn, 2008, p. 66). Sophisticated methods can make research more rigorous (Mentzer and Flint, 1997) but a clear demonstration of rigor is more important than the method used (Flynn, 2008; Mentzer, 2008; Mentzer and Flint, 1997). The conventional criteria for evaluating methodological rigor are construct validity, internal validity, external validity and reliability (Halldo´rsson and Aastrup, 2003; Riege, 2003; Yin, 2009). Some have argued that these four criteria apply to both quantitative and qualitative research (Ellram, 1996; Mentzer and Flint, 1997; Yin, 2009), whereas others have claimed research needs to be evaluated according to the paradigm within which it is rooted (Healy and Perry, 2000). Consequently, Guba and Lincoln (1989) argue that better criteria for evaluating qualitative research are confirmability, credibility, transferability and dependability. In Table I, we juxtapose the conventional criteria with those advocated by Guba and Lincoln since Riege (2003) argued that all eight criteria are useful in demonstrating the “quality, validity, and reliability of case study research” (p. 85).

Applying repertory grid technique 805

IJPDLM 42,8/9

Criteria for evaluating methodological rigor Construct validity Confirmability

806 Internal validity Credibility

External validity Transferability

Reliability

Dependability

Table I. Criteria for assessing methodological rigor

Explanation Establishes appropriate operational measures for theoretical concepts being researched (Mason and Bramble, 1989) Assesses whether the interpretation of data is drawn in a logical and unprejudiced manner (Riege, 2003) and frEe of researcher bias. Integrity of the findings is assured by objectively establishing a link between data and findings Asks whether the approximate truth about inferences is from cause-effect or causal relationships (Trochim and Donnelly, 2008). Do chosen variables accurately explain the issue under research (Dane, 1990)? Refers to the degree to which research findings were verified by interviewees or peers as realities may be interpreted in multiple ways. The purpose of this test is to demonstrate that the inquiry was carried out in a credible way (Riege, 2003) Refers to the generalizability of findings or the degree to which the conclusions from a study statically hold for other individuals, in other contexts, and at moments in time (Trochim and Donnelly, 2008) Refers to the degree to which the understanding obtained in one study can be transferred to explain phenomena observed in other contexts through analytical generalization. Erlandson et al. (1993) argue that full generalizability to other settings is impossible as no two contexts are identical. However, according to them a comprehensive understanding of one context justifies making useful interpretations about similarities and differences in other contexts. Refers to the degree to which a procedures and measures provide consistent results (Trochim and Donnelly, 2008). In other words, reliability is equated to attaining the same results through multiple applications of the same measuring instrument by different researchers and/or at different times (Drucker-Godard et al., 2001) In assessing reliability, all stages of the research process – including data collection, coding, and all other processes of preparing and analyzing data – need to be described as accurately as possible to attain a high degree of transparency. In qualitative research this is described as providing a dependability audit that outlines the process followed that allows for trackability (Halldo´rsson and Aastrup, 2003).

Note: The additional aspects that particularly need to be evaluated for qualitative studies are italicized

The most common type of qualitative research is the case study method but it is often criticized as being applied in a way that lacks rigor (Ellram, 1996; Seuring, 2008) More specifically, Seuring (2008) claimed that much qualitative research presented in SCM journals does not adequately explain how the data were analyzed. In order to assess qualitative research in SCM, we reviewed the articles published in leading journals over the last decade. In conducting this review, we extended the work of Sachan and Datta (2005), who analyzed leading journals from 1999 to 2003. Table II summarizes the different methodologies used in SCM journals from 1999 to 2010, and shows that the number of publications based on case study research increased from 16.1 to 20.3 percent. Given the current popularity of qualitative approaches, the need for qualitative rigor has never been greater.

Period

Research methods

1999-2003 (Sachan and Datta, 2005)

Survey Simulation Interviews Math model Case study Conceptual models Others Total Survey Simulation Interviews Math model Case study Conceptual models Others Total Survey Simulation Interviews Math model Case study Conceptual models Others

2004-2010

Increase/decrease

JBL a n % 48 9 3 12 4 4 12 92 49 12 12 7 15 7 49 151

52.2 9.8 3.3 13.0 4.3 4.3 13.0 100.0 32.5 7.9 7.9 4.6 9.9 4.6 32.5 100.0 219.7 21.9 4.6 28.4 5.6 0.3 19.5

SCMIJ b n % 33 3 15 8 37 11 40 147 85 3 14 12 85 34 81 314

22.4 2.0 10.2 5.4 25.2 7.5 27.2 100.0 27.1 1.0 4.5 3.8 27.1 10.8 25.8 100.0 4.7 21.0 25.7 21.6 1.9 3.3 21.4

IJPDLM c n % 72 10 12 26 30 13 40 203 82 15 14 21 54 29 78 293

35.5 4.9 5.9 12.8 14.8 6.4 19.7 100.0 28.0 5.1 4.8 7.2 18.4 9.9 26.6 100.0 2 7.5 0.2 2 1.1 2 5.6 3.6 3.5 6.9

Total n % 153 22 30 46 71 28 92 442 216 30 40 40 154 70 208 758

34.6 5.0 6.8 10.4 16.1 6.3 20.8 100.0 28.5 4.0 5.3 5.3 20.3 9.2 27.4 100.0 2 6.1 2 1.0 2 1.5 2 5.1 4.2 2.9 6.6

Notes: aJBL – Journal of Business Logistics; bSCMIJ – Supply Chain Management: An International Journal; cIJPDLM – International Journal of Physical Distribution & Logistics Management Source: Updated and enhanced from Sachan and Datta (2005)

One qualitative method that allows rigor to be achieved (and demonstrated) in exploratory studies is RGT. 3. RGT in management research Originally developed by the psychologist Kelly (1955), RGT is a form of structured interviewing. It enables respondents to articulate their views on complex issues (Hussey and Hussey, 1997) and it is particularly appropriate in exploratory studies where the constructs are unclear (Goffin, 2002). For example, in supplier management research, the technique can be used to stimulate interviewees to compare and contrast the performance of their suppliers, rather than asking them direct questions. RGT encourages respondents to think deeply and is an effective approach to identify key constructs. Moreover, RGT brings a quantitative angle to qualitative data, which has been recommended (Eisenhardt, 1989). Table III shows over the last ten years at least 78 management studies utilized RGT. However, only six studies published in the leading operations and SCM journals used the technique (8.9 percent). From these, only one SCM study used RGT – our own work focused on partnerships (Goffin et al., 2006; Lemke et al., 2003). We will next describe this research, which we will refer to as Study 1 (Table IV).

Applying repertory grid technique 807

Table II. Research methods used in SCM research (1999-2010)

IJPDLM 42,8/9

808

Table III. Use of RGT in management research 1999-2010

Management information systems, knowledge management (MIS, KM) Marketing Organisational behavior/studies, human resource management, industrial relations (OB/OS, HRM IR) General management & strategy (gen. man. & stra.) Operation research, management science, production and operations management (OR, MS, POM) – see Table 4 Tourism Public sector management (PSM) International business (IB) Psychology Finance & accounting (F&A) Total

24 14

30.4 17.7

14 8

17.7 10.1

6 4 3 2 2 1 78

8.9 5.1 3.8 2.5 2.5 1.3 100.0

Source: Ebsco and ProQuest (ABI Inform)/classification based on ABS2010

Authors

Journal

Keywords

1. Yan et al. (2009) 2. Phillips (2006)

Computers & Industrial Engineering Journal of Multi-Criteria Decision Analysis IEEE Transactions on Engineering Management Computers & Operations Research Journal of Operations Management. International Journal of Physical Distribution & Logistics Management International Journal of Operations & Production Management

Product management Multiple decision making criteria Information resources management Decision support systems Manufacturer-supplier relationships

3. Tan and Gallupe (2006)

Table IV. Operations research, management science, production & operations management papers based on RGT

Number of papers Percentage

Functional area

4. Scheubrein and Zionts (2006) 5. Goffin et al. (2006) Lemke et al. (2003) 6. Duberley et al. (2000)

Manufacturing processes

Source: Ebsco and Proquest (ABI Inform)/classification based on ABS2010

4. Study 1: partnership research (1997-2006) 4.1 Introduction This research, which was conducted over ten years, focused on manufacturer-supplier relationships. The value of effective supplier management is well known: it can reduce costs, improve quality, and enhance product design (Monczka et al., 1993; Primo and Amundson, 2002). To achieve these advantages, practitioners have recognized the efficiency of working in partnerships with suppliers (Fretty, 2001; Kerns, 2000). At the time that our study commenced in 1997, the key characteristics of such relationships were unclear (Goffin et al., 1997), although the literature indicated that manufacturers choose their suppliers based on their price, quality and delivery performance (Goffin et al., 2006). 4.2 Research aims Addressing the lack of clarity in the literature on the nature of manufacturer-supplier relationships, our research question was:

RQ1. What do managers perceive as the main attributes of close manufacturer-supplier relationships? Previous supply chain relationship research relied on large surveys (Ellram and Hendrick, 1995; Mudambi and Schru¨nder, 1996; Saxton, 1997), which did not uncover the attributes of partnerships. A fresh approach was needed. 4.3 Choice of RGT Interviews can be used to study supplier relationships (McCutcheon and Stuart, 2000) but our pilot investigation showed that direct questions have serious limitations. This is shown by Table V, which summarizes the most frequent answers from pilot interviews with purchasing managers. Fourteen managers answered that their relationships were “good” and nine that they were “partnerships”. Such quotes indicate that the answers were superficial, the term “partnership” has become debased and ambiguous (Brennan, 1997; McCutcheon and Stuart, 2000). To acquire a deeper understanding of “partnerships”, RGT was selected.

Applying repertory grid technique 809

4.4 Sample For our investigation, we chose an exploratory sample of German medium-sized Mittelstand companies. From a list of 110 high performing companies, we contacted 28 companies in the electronics and engineering sectors, of which 18 agreed to participate. After the pilot telephone interviews with purchasing managers, we made on-sites visits and used RGT to probe deeply into the nature of manufacturer-supplier relationships. In total 39 repertory grid interviews were conducted (typically two per company, resulting in over 46 hours of interview data).

Characteristic

Example quotes/details

1. Good

“If you asked me directly, I can tell you that the relationships with our suppliers are very good” (Purchasing Manager in Engineering) “Our relationships are partnership-like, which sums it up pretty much” (Purchasing Manager in Process) “We are co-operative and this in objective terms. So, working together does not only mean speaking about things – joint actions are needed [. . .] It is a co-operative effort and meetings are scheduled on a regular basis” (Purchasing Manager in Engineering) “Our relationships are normal and based on what any purchasing manager would reasonably expect” (Purchasing Manager in Electronics) “In my experience, our relationships are open which allows us to address issues in an uncomplicated fashion” (Managing Director in Engineering) For example, “goal oriented” (four mentions)

14

Total

49

2. Partnership-like 3. Co-operative

4. Normal/reasonable

5. Open

6. Seven other categories

Source: Answers from 34 managers to a direct question, from Goffin et al. (2006)

Mentions 14

9 5

4

3 Table V. The five most frequently mentioned characteristics of manufacturer-supplier relationships

IJPDLM 42,8/9

810

4.5 Applying the RGT In this section we will describe one grid elicited from one respondent (Interviewee 16-2). This respondent named nine of his company’s suppliers: these included three “close”, three “distant”, and three “average” relationships. Each supplier’s name was written on a blank postcard (pre-numbered); these were the elements of the technique. The interviewee was then presented with a random set of three elements (termed a triad in repertory grid terminology) and asked: “Why is working with two of these suppliers similar and different from working with the third?” The interviewee’s response – termed a construct – was that one supplier had reliable “delivery on-time”, whereas the shipments from the other two were often “delayed”. After identifying the first construct, each of the nine suppliers was rated against it on a five-point scale (with 1 “very good”, 3 “neutral”, and 5 “very poor”). Further triads are used to elicit other constructs. The interviewee was not allowed to repeat constructs and so was forced to think more deeply about supplier relationships as the interview progressed. After identifying each construct, open-ended questions (such as “Can you explain that further?”) were used to clearly understand its meaning (Drew, 1995). Table VI shows the complete grid with the ten constructs elicited from the interviewee, including “delivery reliability on-time”, “product quality” and “early information”. The ratings of the elements on “early information” show that the respondent differentiated strongly between the performances of their suppliers on this construct. In contrast, the ratings of the elements on “product quality” range only from 1 to 3. Variability is the measure of the spread of ratings on a particular construct and gives an indication of an interviewee’s salient constructs (Smith, 1986). The RGT elicits constructs in an unbiased way, as they are based on the elements (i.e. suppliers) and the constructs (i.e. attributes of supplier relationships) nominated by the interviewee. So each grid gave detailed insights into one respondent’s views on manufacturer-supplier relationships. 4.6 Analysis The 39 repertory grid interviews elicited a total of 411 constructs (far more than the 49 answers from the pilot interviews). Several respondents mentioned the same constructs and these were grouped into categories using a coding process, based on our interpretation of each interviewee’s explanations of their constructs. One of our research team conducted this process and constantly referred to the interview transcripts. The emergent categories (Goffin et al., 2006) were discussed with the other members of the research team and this produced a list of 26 categories (Table VII). The most frequently mentioned categories were “flexibility” (mentioned by 23 respondents) and “delivery performance” (20 respondents). The constructs elicited covered a range of factors including suppliers’ organizations (e.g. size and capability); the relationship itself (e.g. its duration); and the strategic value of the relationship to the manufacturer. The wide range of attributes illustrates the complexity of manufacturer-supplier relationships, which did not emerge from the conventional (pilot) interviews. 4.7 Results It was important to identify the salient characteristics of manufacturer-supplier relationships from the 26 categories of constructs. Key constructs can be identified firstly by their frequency of mention. Therefore, only the constructs mentioned by at least

2 *4 * 4

3 4 3 3 3

*5 *

3 *3 *

*3 *

1 2 1 1

2 3

*1 *

2

*1 *

1 2 *2 * 3

*3 *

2 2 3 *3 *

*2 *

4

2 1 *1 * 1

3 *2 *

2 *2 *

3 *3 *

3 3 3 3 *3 *

3 3 4 *4 * 3

1 1 *1 * 1 1

*4 *

*4 *

*1 *

Note: Respondent 16-2: a manager in the electronics sector

7. Early information 8. Confident 9. Takes us seriously 10. Flexible

1. Delivery reliability on-time 2. Product quality high 3. Price level low 4. Competence high 5. Contact(s) always the same 6. Makes decisions

Constructs

*3 *

2

*3 *

2

4 3 *3 * *3 * 3 3 3 3 *2 * 3

2 1 *2 * 1 *1 * 2

4 1 3 *3 *

3 3 *2 * 2 2 *5 *

Delayed Low High Low Constantly changing Does not decide anything No information Clumsily Arrogant Inflexible

Elements – suppliers Supplier Supplier Supplier Supplier Supplier Supplier Supplier Supplier Supplier 1 2 3 4 5 6 7 8 9 Close Distant Distant Average Close Average Distant Close Average Poles

Applying repertory grid technique 811

Table VI. An example of a repertory grid

IJPDLM 42,8/9

812

Table VII. The attributes of partnership

No.

Categories of constructs

E1 E2 E3 E4 E5 E6 E7 E8 E9

Flexibility Delivery performance Personal relationship Special product capability Quality Dependency Size of organization Relationship maintenance Volume of turnover as bought from the supplier Price level Feedback New product development Complaint handling Location

E10 E11 E12 E13 E14

Frequency (% of respondents)

Average normalized variability (ANV)

Key category

23 20 18 17 16 15 14 13 13

(59) (51) (46) (44) (41) (38) (36) (33) (33)

Low Low High High Low High High High High

No No Yes Yes No Yes Yes Yes Yes

11 11 10 10 10

(28) (28) (26) (26) (26)

Low High High High Low

No Yes Yes Yes No

Note: Factors mentioned by 25 percent or more of respondents and with a “high” ANV are italicized Source: Full explanations of these constructs are in Goffin et al. (2006)

25 percent of the respondents (i.e. by ten or more) are listed in Table VII. Frequency, however, is only one indication of importance and it has the limitation that the most obvious rather than the most important constructs will be frequently mentioned. Constructs which differentiate strongly between the different suppliers have a high variability (i.e. a wide spread of ratings) and so the variability of the constructs in each grid was calculated. Because the variability measure is directly affected by the number of constructs in the individual grid, constructs from grids with different numbers of constructs need to be normalized. The normalized variability figures for each construct were then averaged across the grids to produce the average normalized variability (ANV) – see Goffin et al. (2006) for a full explanation. Constructs with a “high” ANV are more important than those with a “low” ANV. It can be seen that “flexibility” (construct category E1: “E” designating an empirically derived construct) was mentioned by nearly 60 percent of interviewees but had a low ANV. This shows that flexibility was often mentioned but most suppliers had similar performance on this construct. In contrast, the construct “special product capability” (E4) had a high frequency (17) and a high ANV and so can be designated as a key attribute of manufacturer-supplier relationships. Table VII raises some interesting points: “flexibility” (E1), “delivery performance” (E2), “quality” (E5), and “price level” (E10) were all frequently mentioned but none had a high ANV. So, despite the fact that these factors were frequently mentioned (and are often cited in both the academic and practitioner literature), all of the suppliers had similar performance when rated against these factors. In other words, our results suggest that “flexibility”, “delivery performance”, “quality” and “price” are hygiene factors and suppliers need to provide a full service and not just good performance on these factors. Using RGT, our research enabled us to identify nine key categories (Table VII) and allowed us to compare these results with the extant literature. Figure 1 is a Venn diagram of the attributes of supplier relationships identified (“L” indicating a construct

Extant Literature

L2 Competitive advantage L3 Co-operation L6 Exchange of resources L8 Information sharing, communication L9 Integration of logistics L11 Non-tender price agreement L14 Sharing risks/rewards L15 Single or limited no.of suppliers L16 Training of supplier L18 Trust L19 Value of resource access L20 Voluntary

Current Study

L1; E19 Commitment L4; E2 Delivery performance L7; E21 Joint problem solving L10; E25 Long-term L12; E10 Price level L13; E3 Quality

E1 Flexibility E14 Location E15 Customer oriented E16 Importance of supplier (to Mfr) E17 Openness E18 Price changes E20 Reliability E22 Competence E23 Organizational culture E24 Classification as a customer E25 Duration E26 Designated contact

Applying repertory grid technique 813

Key Attributes

L5; E6 Dependency L17; E12NPD

E3 Personal relationship E4 Special product capability E7 Size of organization E8 Relationship maintenance E9 Volume of turnover E11 Feedback E13 Complaint handling

Source: Goffin et al. (2006)

identified in the literature and “E” indicating those uncovered in our study). Overall, it can be seen that the repertory grid study identified many attributes not covered by the literature (a full explanation of Figure 1 is given in Goffin et al., 2006). 4.8 Implications of study Study 1 has implications for managers and researchers. Managers now know that suppliers are not only judged and selected on their price, quality, and delivery performance – those are hygiene factors from the manufacturer’s perspective. The rich data generated in our repertory grid interviews also developed a practical model of how manufacturer-supplier relationship can be more effectively managed. The implications of Study 1 for researchers are multi-facetted. On one level they show the need for more probing research into how manufacturer-supplier relationships develop. On another level, the results indicate the efficacy of RGT: it can provide clarity when the extant literature fails to define key constructs unambiguously; and it can also uncover cases when frequently mentioned constructs are not the most important ones. 4.9 Summary of Study 1 The results from the 39 interviews provided three main insights: (1) RGT enabled managers to fully articulate their views on manufacturer-supplier relationships; (2) the most important aspects of relationships were not those most frequently mentioned; and (3) certain constructs were missing from the extant literature.

Figure 1. Venn diagram of the attributes of supplier relationships

IJPDLM 42,8/9

814

5. Study 2: servitization research (2009 ongoing) 5.1 Introduction There is an increasing trend amongst manufacturers to develop after-sales services to support their product offerings (Oliva and Kallenberg, 2003; Wise and Baumgartner, 1999). Services such as maintenance and repair have traditionally been viewed as secondary to the business of selling products (Bowen et al., 1989). However, many manufacturers are now changing their business model and integrating service and product offerings (Vandermerwe and Rada, 1988; Neely, 2008; Desmet et al., 2003; Baines et al., 2009). Developing a customer-focused combination of products and services is termed servitization (Vandermerwe and Rada, 1988). It is important for manufacturers to understand how customers perceive product-service combinations (Brady et al., 2005) and the need to understand customer perceptions of value is a recurring theme in the literature (Woodruff and Gardial, 1996; Flint et al., 2002). The traditional view of “value” focuses on the exchange dimension that predicates the presence of product/service attributes, and performances against those attributes, for which the customer is prepared to pay (Woodruff, 1997). More recent work however argues for a move away from an exchange perspective to understanding the outcome from the offering provided – that is, value-in-use (Vargo and Lusch, 2004). 5.2 Research aims Whilst support services for manufactured goods are not a new phenomenon, servitization challenges firms as to how they integrate products and services (Oliva and Kallenberg, 2003; Wise and Baumgartner, 1999). Although the importance of understanding the customer’s perception of value-in-use is recognized (Flint and Mentzer, 2006; Vargo and Lusch, 2004, 2008), the literature remains unclear about how to assess it. Our study addressed this gap through an exploratory study focused on the research question: RQ2. What do customers perceive as value-in-use of servitized offerings? 5.3 Choice of RGT An initial pilot study showed semi-structured interviews to be ineffective: respondents only mentioned superficial factors, such as “value for money”, “quality” and “reliability”. To enable customers to better articulate their perceptions of value-in-use, we adopted RGT as part of our case study methodology. 5.4 Sample Through a manufacturer of automation equipment, we gained access to four of their customers. These four organizations are original equipment manufacturers (OEM) of pharmaceuticals, insulation materials, specialist metal products, and medical equipment. Data collection took place in the UK and Ireland at the four manufacturing sites of the companies and a total of 33 repertory grid interviews were conducted. 5.5 Applying the RGT Following a similar approach to Study 1, each respondent was asked to name six to ten suppliers of industrial maintenance services and products. The seven generic elements of customer support outlined by Goffin and New (2001) were used in the elicitation of names of suppliers, these include: installation, user training, documentation, maintenance

and repair, on-line support, warranty and upgrades. Then using Kelly’s method of presenting a triad of cards (1955, 1963) we asked: “Can you think of any ways in which two of these suppliers are similar to each other and different from the third in terms of the outcomes you get?” Different triads were then presented to the interviewee and the same question asked each time, to elicit a range of constructs. In total, 272 constructs were elicited from 33 interviews. 5.6 Analysis A more rigorous data analysis process was followed for Study 2 and this will be described in detail. The four stages of this analysis were coding the elicited constructs, producing a reliability table, conducting reliability checks, and a Pareto analysis to demonstrate theoretical saturation. 5.6.1 Coding the elicited constructs. Two sets of construct cards were produced (Jankowicz, 2004). In each set, every one of the 272 constructs elicited in the interviews was listed on a card (Figure 2). This included the construct name and a description or quote describing its meaning taken from the interview transcripts. Subsequently two researchers who conducted the interviews (A and B) carefully inspected the interview transcripts and grids. Both researchers then worked independently and grouped constructs into categories. Researcher A grouped the 272 construct cards into 46 categories, whereas Researcher B grouped them into 43 categories. 5.6.2 Drawing the reliability table. The different categories derived by Researchers A and B were then compared to identify similarities and differences using a reliability table (Jankowicz, 2004). In this, the columns and rows were labelled, respectively, with Researchers A’s 46 categories and Researcher B’s 43 categories (Figure 3). In the cells of the table are the 272 individual constructs. The columns and rows in this table were then rearranged so that the shared categories appeared on the diagonal. Where individual constructs appear off the diagonal, they indicate differences in opinion between Researcher A and B. For example, it can be seen that construct D1-5 was categorized by Researcher A as “responsiveness”, whereas Researcher B categorized it as “reliability”. 5.6.3 Reliability checks. To introduce more rigor into our categorization process, multiple researchers checked the categories and performed inter-rater reliability (IRR) checks (Zwick, 1988). This approach consisted of four stages: Stage 1. The reliability table from the initial coding showed an IRR (Researcher A vs Researcher B) of 48 percent, for all the constructs across 22 categories Company D

Applying repertory grid technique 815

Interviewee name: XXX Construct D1-5 Construct Consistent response Counter pole Not consistent in response

Description/quote “tend to be a mixture, there’s no consistency in the response, some things take longer then we feel they should, some things exceed our expectations but overall things take longer than we would like” (p. 13).

Figure 2. Construct card

IJPDLM 42,8/9

816

Figure 3. Reliability table from the coding of Study 2

(see diagonal line in Figure 3). According to Miles and Huberman (1994) such IRR values are typical at this stage and demonstrate the need for clearer category definitions. Stage 2. This involved a lengthy process of the two researchers discussing the constructs that were positioned off the diagonal of the reliability table. These discussions led to clearer category definitions (Miles and Huberman, 1994), with construct validity being ensured through the researchers constantly consulting the interview transcripts. The process resulted in 29 enhanced category definitions. Stage 3. In line with Jankowicz (2004), the 272 constructs were recoded into one of the 29 enhanced construct categories by Researchers A and B (again working independently and in parallel). The outcome was a second reliability table which showed an IRR of 86 percent (Table VIII, column 4), which exceeded the 80 percent suggested by Miles and Huberman (1994). Stage 4. Subsequently, two other Researchers (C and D), who were not otherwise involved in the research, independently coded the 272 constructs into the 29 enhanced categories. This tested the robustness of the coding process. Researchers C’s and D’s coding showed IRRs of 78 and 83 percent (Table VIII, columns 5 and 6), demonstrating the reliability of the coding. Through our qualitative coding (categorization) process, we were able to ensure reliability. However, the question remained whether our four case studies and 33 repertory grid interviews were sufficient to achieve theoretical saturation. 5.6.4 Theoretical saturation – Pareto analysis. In qualitative research, an important issue is whether theoretical saturation is reached. Saturation is reached either when no new categories or very few additional ones emerge from case studies (Strauss and Corbin, 1998). Researchers need to ensure sufficient data have been collected to produce an accurate picture of the phenomenon under investigation before exiting the field (Glaser and Strauss, 1967). It is particularly important for researchers to provide clear evidence that saturation is achieved (Bowen, 2008) and so we conducted a Pareto analysis. The Pareto in Figure 4 was compiled by plotting the total number

Stages of the coding/reliability checks

Measure

Stage 1: coding Researcher A vs Researcher B

Time required for each stage

10 hours to prepare construct cards

8 hours for 8 hours per researcher to coding re-code

16 hours to prepare the 29 enhanced construct category definitions

817

10 hours for coding

4 hours for 4 hours for 6 hours for the reliability the reliability the reliability table table table N/A

86

78

45

100

40

90

83

Table VIII. Stages of the coding (categorization) process

70

30

60 25 50 20 40 15 30 10

Percentage of total categories

80

35

20

C om pa ny

C om pa ny

C om pa ny

Case studies

D

0 C

0 B

10

A

5

C om pa ny

Total number of distnict categories per case

Inter-rater reliability (IRR) (%)

8 hours for the coding by each researcher 10 hours for the reliability table 48

Stage 3: Stage 4: independent checks re-coding Researcher C Researcher D vs vs Researcher A Researcher Researcher vs A/B A/B Researcher B

Stage 2: defining construct categories Researcher A and Researcher B

Applying repertory grid technique

Figure 4. Pareto demonstrating category saturation

IJPDLM 42,8/9

818

of distinct construct categories elicited at each company. It can be seen that in Company A 23 distinct categories emerged, in Company B 13 additional categories, in Company C 5, and in Company D only one more category emerged. In this respect, Figure 4 demonstrates that the interviews at Company D only added one new construct category. From this we conclude the 33 repertory grid interviews were sufficient to identify the relevant construct categories and further case studies would not have produced new categories. 5.7 Results The interviews produced 272 constructs that were coded into 29 enhanced categories. Of these, those that are mentioned by at least 25 percent of respondents are listed in Table IX. In line with Study 1, the frequency of mentions and the ANV for the categories were calculated. From Table IX it can be seen that the construct category “specialist knowledge” (construct category E1) had both a high frequency (25) and a high ANV and so is a key attribute of value-in-use. The other key attributes were: “accessibility” (E3), “relational dynamic” (E4), “range of product and service offering” (E5), “delivery” (E8), “pricing” (E9) and “locality” (E11). “Responsiveness” (E2) was mentioned 22 times but a low ANV score, indicating that most suppliers are responsive (and are rated similarly). Other construct categories mentioned frequently but with a low ANV included: “understanding customer business” (E6), “service orientation” (E7), “convenience” (E10), “quality of repairs” (E12), “feedback and reporting” (E13) and “communication” (E14). 5.8 Implications of study The findings from this study have implications for practitioners and researchers. Practitioners now know the attributes that are important to customers of servitized offerings. Importantly, this is the first study to identify the perceived attributes of value-in-use.

No. Categories of constructs

Table IX. Categories of constructs on value-in-use

E1 E2 E3 E4 E5 E6 E7 E8 E9 E10 E11 E12 E13 E14

Specialist knowledge Responsiveness Accessibility Relational dynamic Range of product and service offerings Understanding customer business Service orientation Delivery Pricing Convenience Locality Quality of repairs Feedback and reporting Communication

Frequency (% of Average normalized respondents) variability (ANV) Key category 25 22 18 15 14 14 13 13 12 11 11 11 11 9

(76) (67) (55) (45) (42) (42) (39) (39) (36) (33) (33) (33) (33) (27)

High Low High High High Low Low High High Low High Low Low Low

Yes No Yes Yes Yes No No Yes Yes No Yes No No No

Note: Factors mentioned by 25 percent or more of respondents and with a “high” ANV are italicized

For researchers, Study 2 highlights the need to understand value-in-use from the customer perspective. The results show RGT is useful for exploratory research in defining key constructs through a rigorous and transparent process. 5.9 Summary of Study 2 The repertory grid interviews provided a number of important insights: the technique allowed the construct value-in-use to be better understood; the seven key attributes identified are shown to be important considerations for manufacturers in the design and delivery of their product-service offerings; the reliability of the results obtained was demonstrated through the checks performed; and theoretical saturation was demonstrated with a Pareto analysis. 6. Implications for qualitative research The benefit of having conducted two studies using RGT is that our experience with the approach and the two data sets can shed light on how to apply qualitative research in SCM. In this section, we present what we consider to be our key lessons learned. Table X summarizes the findings of the two studies. We give only an overview of the insights gleaned on the SCM issues (readers who are interested in these findings should refer to the references given). As this paper is focused on methodological findings, Table X provides a comprehensive overview of these. The criteria from the literature for assessing qualitative research (summarized in Table I) was used to reach these conclusions. Therefore, the second column in Table X provides conclusions against the criteria “construct validity/confirmability”; “internal validity/credibility”; “external validity/transferability”; and “reliability/dependability”. From Table X, it can be seen that we came to four conclusions about construct validity and confirmability from our experience of RGT. We found that RGT generates more meaningful constructs than those that emerge from direct questions (as shown by comparing the pilot with the full results of Study 1). This is particularly helpful when investigating concepts such as manufacturer-supplier relationships where there is ambiguity in the ways terms such as “partnership” are used. Our rich data set provided a depth of understanding on each of the constructs (Goffin et al., 2006). We also found that RGT reduces bias in that the constructs can be elicited without terms such as “partnership” being mentioned. Another interesting point about RGT is that it provides construct definitions that are tied to their context. Therefore, it should be noted that even if the same or similar word is elicited in different studies, it does not necessarily refer to the same thing. For example, “location” in manufacturer-supplier relationships does not necessarily have the same meaning as “locality” in the value-in-use of servitized offerings. Finally, we found that RGT is particularly useful for exploratory work, where the constructs have not yet been clearly identified. Considering the internal validity and credibility of our qualitative research, we reached five conclusions. In Study 1 our coding was conducted by one researcher (and then discussed by the other team members), whereas in Study 2 we went further and the coding was conducted by two researchers in parallel. This enhanced approach ensured that the interpretation of data was more consistent. We also found that RGT is a highly structured approach and so it can be easily understood and applied by different researchers. Another conclusion was that the frequency of mention of a construct is not necessarily a good measure of importance: “regrettably, frequency of mention does

Applying repertory grid technique 819

Table X. Overview of the SCM and methodological contributions of Studies 1 and 2 Internal validity/credibility Having multiple researchers code the data ensured that the interpretation of the data were consistent and thus more credible As the RGT is a very structured form of interviewing, it means that the way the (qualitative) data are collected and analyzed is transparent, thus adding credibility to the findings RGT allows a differentiation to be made between frequently mentioned factors and the factors that are most important (using the ANV measure) Pareto analysis can be used to demonstrate theoretical saturation and thus increase internal validity In our case studies we used multiple sources of data and so the results from the RG interviews could be triangulated against company documentation, etc. External validity/transferability RGT is most useful for exploratory work but, with small samples, the external validity is a limitation. However, given the high construct validity of the constructs derived by RGT, survey research can take these constructs and test them with larger populations to achieve statistical generalizability (this shows the necessity for a plurality of methods) Although the external validity is a limitation, the rich, unambiguous construct definitions can be easily checked in other contexts, to see whether they are transferable (i.e. it can also be more easily recognized when they are not applicable). Similarly, as the construct definitions are so rich, the findings of RGT exploratory can be accurately compared to constructs in the extant literature Reliability/transferability The reliability of qualitative coding can be improved through multiple coding rounds and the calculation of IRR for the coding conducted by different members of the research team Having independent researchers conduct a check on the coding also helps to make the complex analysis of qualitative data more reliable The transparency of the RGT (in the data collection, coding, and analysis process) means that it is easy for researchers to conduct replication studies

Construct validity/confirmability RGT probes deeper than direct questions and identifies more meaningful constructs The interview transcripts provide rich explanations of each construct identified. As these are in the interviewees’ own language, they ensure high construct validity Researcher bias is reduced as the question posed in RGT (e.g. why is work ins with two of these suppliers similar and different from working with the third?) is always the same and it contains neither preconceived concepts, nor value judgements RGT is particularly effective for exploratory research where the key constructs of the area need to be identified. The definitions of the construct categories (using multiple researchers’ coding) are comprehensive and unambiguous. Therefore, they can be reliably compared and contrasted to the extant literature

820

Study 1: manufacturer-supplier relationships Key SCM findings There are nine key attributes of manufacturersupplier relationships (not all of which were in the extant literature) In supplier selection, “price”, “flexibility”, “quality”, and “delivery” are hygiene not orderwinning factors Study 2: the value-in-use of servitized offerings Key SCM findings RGT allowed the complex construct value-in use to be understood In evaluating servitized offerings, “specialist knowledge”, “accessibility”, “relational dynamic”, “range of product and service”, “delivery”, “pricing” and “locality” are found to be some of the key constructs through which value is attributed

Conclusions on applying RGT to SCM research

IJPDLM 42,8/9

not appear a good surrogate for importance” (Griffin and Hauser, 1993, p. 19). Some constructs were readily mentioned by respondents but are not that important. The use of the measure ANV, however, allowed the most important constructs to be identified. Another key learning concerned theoretical saturation. Strauss and Corbin (1998) warn researchers that ceasing data collection before theoretical saturation is achieved may lead to a theory that is not “fully developed in terms of density and variation” (p. 292). Much of the literature on qualitative methodology recommends researchers to immerse themselves in the data until no new data “seem” to emerge (Patton, 2002; Crabtree and Miller, 1999). We believe that it is not sufficient for researchers to rely on their intuition but that they need to clearly demonstrate the internal validity of their work. Like Goffin and Koners (2011), we used a Pareto analysis to demonstrate theoretical saturation. Interestingly, we could not identify any research other than our own where Pareto analysis is applied to demonstrate theoretical saturation[1]. This is perhaps a symptom of the philosophical divide between quantitative and qualitative approaches, where some qualitative researchers shy away from numerical analysis to demonstrate the validity of their findings. It is salient to note that some social scientists recognize the value of both qualitative and quantitative methods: “Another characteristic of good ethnography is its inclusion of both quantitative and qualitative data” (LeCompte and Schensul, 1999, p. 18). Therefore, we also concluded that multiple sources of data can increase internal validity (and we applied this in our cases studies). On the issue of external validity and transferability, we reached two conclusions. The samples used in both studies were over the 30 interviews recommended in the literature, but is still a limitation. For this reason, our studies need to be viewed as exploratory work which has elicited constructs and generated clear definitions which can be validated using surveys of larger samples. We also believe that the rich data, where multiple respondents gave their explanations of constructs and several researchers coded the data, means that it is easier for other researchers to check whether the constructs we have identified apply (or do not apply) to their samples – that is, a degree of transparency has been achieved. Critics of qualitative research often question the reliability of findings (cf. Na¨slund, 2002; Knemeyer et al., 2001) and so reliability and dependability need careful consideration. We followed the recommendations of Miles and Huberman (1994) and created definitions of the categories that emerged. Study 2 went further and used reliability tables and IRR calculations to create more robust construct (category) definitions. The use of independent researchers to check the coding reduced the chance of systematic bias in the interpretation of the results by the research team. Moreover, a clear trail of evidence is available for reanalysis by others (Miles and Huberman, 1994). In our opinion, reliability checks should form an integral part of qualitative research. 7. Discussion, limitations and future research We believe that qualitative and quantitative methods should be used in combination rather than in isolation. Qualitative researchers need to embrace some quantitative approaches, whereas quantitative researchers need construct definitions with high validity, such as those generated by RGT. It is our contention that the addition of RGT to the toolset of methods used by logistics and SCM researchers can help generate more insights on which more robust theories can be based. By adopting the RGT researchers may also counter the common criticism

Applying repertory grid technique 821

IJPDLM 42,8/9

822

yielded at qualitative research for failing to provide clear and transparent accounts of the analysis process (Gephart, 2004; Seuring, 2008; Bluhm et al., 2011; Pratt, 2008). While the RGT is extremely useful, it is not without limitations. One such limitation when analyzing multiple grids is that constructs elicited from respondents are not structured in any way. This means relationships between personal constructs are unclear when they are grouped across the respondents. Lastly, the small sample sizes in both studies described limits the ability to make generalizable industry recommendations. In our studies to date, we only used the RGT to understand one side of the dyad. Future research in SCM should analyse both sides of the dyad or even the entire supply chain. Furthermore, Sachan and Datta (2005) point towards the need to understand behavioural issues that are affected by factors like culture, relationships, trust and power in relation to the supply chain and not simply at the function or firm level. The RGT is useful to understand such issues in combination with other methods. Lastly, there is a need for longitudinal research: it would be interesting to see how manufacturer-supplier relationships and value-in-use attributes evolve over time. .

8. Summary This paper discussed two SCM studies using RGT. By comparing and contrasting these two studies, a number of conclusions were reached on how RGT can be applied in SCM research. First, it was demonstrated that RGT is effective for exploratory research where key constructs need to be identified, and where empirically derived constructs need to be compared to those in the literature. Second, RGT differentiates between frequently mentioned factors and those that are most important. Third, a Pareto analysis may be applied to demonstrate theoretical saturation (Goffin and Koners, 2011). Lastly, the reliability of qualitative coding can be improved through multiple coding rounds and the calculation of IRR checks. Performing such quantitative calculations can strengthen qualitative research (and numerical reliability checks can be applied to other qualitative data, not just RGT data). This special issue called for “papers that investigate and reflect, and perhaps theorize, on the creation of qualitative research in logistics and SCM studies” (Gammelgaard and Flint, 2010). We have shared with readers how our experiences of two studies have not only provided us with insights into SCM but also prompted us to deeply consider how we conduct our research. Over the course of the two studies we have learnt much about how to demonstrate validity, reliability, and theoretical saturation of qualitative research. Qualitative researchers need to make their methods more transparent (Bluhm et al., 2011; Gephart, 2004; Pratt, 2008) and, it is often stressed that, in presenting qualitative research, scholars must demonstrate a trail of evidence (Miles and Huberman, 1994; Yin, 2009). This is the central challenge for qualitative research because if the trail of evidence is not convincing, qualitative researchers will always be subject to a trial of evidence. Note 1. A full text search using “theoretical saturation” and “Pareto” on the ABI/Inform and EBSCO databases produced papers where the Pareto technique had been applied to demonstrate theoretical saturation. We first used Pareto to check for theoretical saturation in Goffin et al. (2010).

Baines, T.S., Lightfoot, H.W., Benedettini, O. and Kay, J.M. (2009), “The servitization of manufacturing: a review of literature and reflection on future challenges”, Journal of Manufacturing Technology Management, Vol. 20 No. 5, pp. 547-67.

Applying repertory grid technique

Bartunek, J.M., Rynes, S.L. and Ireland, R.D. (2006), “What makes management research interesting, and why does it matter?”, Academy of Management Journal, Vol. 49 No. 1, pp. 9-15.

823

References

Bluhm, D.J., Harman, W., Lee, T.W. and Mitchell, T.R. (2011), “Qualitative research in management: a decade of progress”, Journal of Management Studies, Vol. 48 No. 8, pp. 1866-91. Bowen, D.E., Siehl, C. and Schneider, B. (1989), “A framework of analysing customer service orientations in manufacturing”, Academy of Management Review, Vol. 14 No. 1, pp. 75-95. Bowen, G.A. (2008), “Naturalistic inquiry and the saturation concept: a research note”, Qualitative Research, Vol. 8 No. 1, pp. 137-52. Boyer, K.K. and Swink, M.L. (2008), “Empirical elephants – why multiple methods are essential to quality research in operation and supply chain management”, Journal of Operations Management, Vol. 26 No. 3, pp. 337-48. Brady, T., Davies, A. and Gann, D.M. (2005), “Creating value by delivering integrated solutions”, International Journal of Project Management, Vol. 23 No. 5, pp. 360-5. Brennan, R. (1997), “Buyer/supplier partnering in British industry, the automotive ad telecommunications sectors”, Journal of Marketing Management, Vol. 13 No. 8, pp. 759-75. Burgess, K., Singh, P.J. and Koroglu, R. (2006), “Supply chain management: a structured literature review and implications for future research”, International Journal of Operations & Production Management, Vol. 26 No. 7, pp. 703-29. Crabtree, B. and Miller, W. (1999), Doing Qualitative Research, 2nd ed., Sage, Newbury Park, CA. Dane, F.C. (1990), Research Methods, Brookes/Cole, Pacific Grove, CA. Desmet, S., van Dierdonck, R. and van Looy, B. (2003), “Servitization: or why services management is relevant for manufacturing environments”, in van Looy, B., Gemmel, P. and van Dierdonck, R. (Eds), Services Management: An Integrated Approach, Pearson Education, Harlow. Drew, P. (1995), “Conversation analysis”, in Smith, J.A., Harre´, R. and Langenhove, L. (Eds), Rethinking Methods in Psychology, Sage, London, pp. 64-79. Drucker-Godard, C., Ehlinger, S. and Grenier, C. (2001), “Validity and reliability”, in Thietart, R.-A. et al. (Eds), Doing Management Research, Sage, London. Duberley, J., Johnson, P., Cassell, C. and Close, P. (2000), “Manufacturing change”, International Journal Of Operations & Production Management, Vol. 20 No. 4, pp. 427-40. Eisenhardt, K. (1989), “Building theories from case study research”, Academy of Management Review, Vol. 14 No. 4, pp. 532-50. Ellram, L.M. (1996), “The use of the case study method in logistics research”, Journal of Business Logisitics, Vol. 15 No. 2, pp. 93-138. Ellram, L.M. and Hendrick, T.E. (1995), “Partnering characteristics: a dyadic perspective”, Journal of Business Logistics, Vol. 16 No. 1, pp. 41-64. Erlandson, D.A., Harris, E.K., Skipper, B.L. and Allen, S.D. (1993), Doing Naturalistic Inquiry: A Guide to Methods, Sage, Newbury Park, CA.

IJPDLM 42,8/9

824

Flint, D.J. and Mentzer, J.T. (2006), “Striving for integrated value chain management given a service-dominant logic for marketing”, in Lusch, R.F. and Vargo, S.L. (Eds), The Service Dominant Logic of Marketing, M.E. Sharpe, London, pp. 139-49. Flint, D.J., Woodruff, R.B. and Gardial, S.F. (2002), “Exploring the phenomenon of customers’ desired value change in a business-to-business context”, Journal of Marketing, Vol. 66 No. 4, pp. 102-17. Flynn, B.B. (2008), “Having it all: rigor versus relevance in supply chain management research”, Journal of Supply Chain Management, Vol. 44 No. 2, pp. 63-7. Fretty, P. (2001), “Partnering pays”, Industrial Distribution, Vol. 90 No. 5, pp. 13-16. Gammelgaard, B. and Flint, D. (2010), “Call for papers: special issue on qualitative research in Logistics & Supply Chain Management”, International Journal of Physical Distribution & Logistics Management, available at: www.emeraldinsight.com/products/journals/call_ for_papers.htm?id¼2086 Gephart, R.P. (2004), “Qualitative research and the Academy of Management Journal”, Academy of Management Journal, Vol. 47 No. 4, pp. 454-62. Glaser, B.G. and Strauss, A.L. (1967), The Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine Publishing Company, Hawthorne, NY. Goffin, K. (2002), “Repertory grid technique”, in Partington, D. (Ed.), Essential Skills for Management Research, Sage, London, pp. 199-225. Goffin, K. and Koners, U. (2011), “Tacit knowledge, lessons learnt, and new product development”, Journal of Product Innovation Management, Vol. 28, pp. 300-18. Goffin, K. and New, C. (2001), “Customer support and new product development – an explorative study”, International Journal of Operations & Production Management, Vol. 21 No. 3, pp. 275-301. Goffin, K., Lemke, F. and Koners, U. (2010), Identifying Hidden Needs: Creating Breakthrough Products, Palgrave Macmillan Academic publishers, Basingstoke. Goffin, K., Lemke, F. and Szwejczewski, M. (2006), “An exploratory study of ‘close’ supplier-manufacturer relationships”, Journal of Operations Management, Vol. 24 No. 1, pp. 189-209. Goffin, K., Szwejczewski, M. and New, C. (1997), “Managing suppliers: when fewer can mean more”, International Journal of Physical Distribution & Logistics Management, Vol. 27 No. 7, pp. 422-36. Goodyear, M. (1990), “Qualitative research”, in Birn, R., Hague, P. and Vangelder, P. (Eds), A Handbook of Market Research, Kogan Page, London. Griffin, A. and Hauser, J.R. (1993), “The voice of the customer”, Marketing Science, Vol. 12 No. 1, pp. 1-27. Guba, E. and Lincoln, Y.S. (1989), Fourth Generation Evaluation, Sage, Newbury Park, CA. Gulati, R. (2007), “Tent poles, tribalism, and boundary spanners: the rigor-relevance debate in management research”, Academy of Management Journal, Vol. 50, pp. 775-82. Halldo´rsson, A. and Aastrup, J. (2003), “Quality criteria for qualitative inquiries in logistics”, European Journal of Operational Research, Vol. 144, pp. 321-32. Healy, M. and Perry, C. (2000), “Comprehensive criteria to judge validity and reliability of qualitative research within the realism paradigm”, Qualitative Market Research: An International Journal, Vol. 3 No. 3, pp. 118-26. Hussey, J. and Hussey, R. (1997), Business Research, Macmillan Press, London. Jankowicz, D. (2004), The Easy Guide to Repertory Grids, Wiley, Chichester.

Kelly, G.A. (1955), The Psychology of Personal Constructs, Routledge, London. Kerns, C.D. (2000), “Strengthen your business partnership: a framework and application”, Business Horizons, Vol. 43 No. 4, pp. 17-22. Knemeyer, M.A., Ponzurick, T.G. and Logar, C.M. (2001), “A qualitative examination of factors affecting reverse logistics systems for end-of-life computers”, International Journal of Physical Distribution & Logistics Management, Vol. 32 No. 6, pp. 455-79. LeCompte, M. and Schensul, J.J. (1999), Designing and Conducting Ethnographic Research, Altamira Press, Lanham-Seabrook, MD. Lemke, F., Goffin, K. and Szwejczewski, M. (2003), “Investigating the meaning of supplier-manufacturer partnerships: an exploratory study”, International Journal of Physical Distribution & Logistics Management, Vol. 33 No. 1, pp. 12-35. McCutcheon, D. and Stuart, F.I. (2000), “Issues in the choice of supplier alliance partners”, Journal of Operations Management, Vol. 18 No. 3, pp. 279-301. Mason, E.J. and Bramble, W.J. (1989), Understanding and Conducting Research, 2nd ed., McGraw-Hill, New York, NY. Mentzer, J.T. (2008), “Rigor versus relevance: why would we choose only one?”, Journal of Supply Chain Management, Vol. 44 No. 2, pp. 72-7. Mentzer, J.T. and Flint, D.J. (1997), “Validity in logistics research”, Journal of Business Logistics, Vol. 18 No. 1, pp. 199-216. Miles, M.B. and Huberman, M. (1994), Qualitative Data Analysis: An Expanded Sourcebook, Sage, London. Monczka, R.M., Trent, R.J. and Callahan, T.J. (1993), “Supply base strategies to maximize supplier performance”, International Journal of Physical Distribution & Logistics Management, Vol. 23 No. 4, pp. 42-54. Mudambi, R. and Schru¨nder, C.P. (1996), “Progress towards buyer-supplier partnerships: evidence from small and medium sized manufacturing firms”, European Journal of Purchasing & Supply Management, Vol. 2 Nos 2/3, pp. 119-27. Na¨slund, D. (2002), “Logistics needs qualitative research – especially action research”, International Journal of Physical Distribution &Logistics, Vol. 32 No. 5, pp. 321-38. Neely, A. (2008), “Exploring the financial consequences of the servitization of manufacturing”, Operations Management Research, Vol. 1 No. 2, pp. 103-18. New, S.J. (1997), “The scope of supply chain management research”, Supply Chain Management, Vol. 2 No. 2, pp. 15-22. Oliva, R. and Kallenberg, R. (2003), “Managing the transition from products to services”, International Journal of Service Industry Management, Vol. 14 No. 2, pp. 160-72. Palmer, D., Dick, B. and Freiburger, N. (2009), “Rigor and relevance in organization studies”, Journal of Management Inquiry, Vol. 18 No. 4, pp. 265-72. Patton, M.Q. (2002), Qualitative Research and Evaluation Methods, 3rd ed., Sage, Thousand Oaks, CA. Phillips, L.D. (2006), “How Raiffa’s RAND memo led to a multi-criteria computer program”, Journal of Multi-Criteria Decision Analysis, Vol. 14 Nos 4-6, pp. 185-9. Polzer, J.T., Gulati, R., Khurana, R. and Tushman, M.L. (2009), “Crossing boundaries to increase relevance in organizational research”, Journal of Management Inquiry, Vol. 18, pp. 280-6. Pratt, M.G. (2008), “Fitting oval pegs into round holes: tensions in evaluating and publishing qualitative research in top-tier North American journals”, Organizational Research Methods, Vol. 11 No. 3, pp. 481-509.

Applying repertory grid technique 825

IJPDLM 42,8/9

826

Primo, M.A.M. and Amundson, S.D. (2002), “An exploratory study of supplier relationships on new product development outcomes”, Journal of Operations Management, Vol. 20 No. 1, pp. 33-52. Riege, A.M. (2003), “Validity and reliability tests in case study research: a literature review with ‘hands-on’ applications for each research phase”, Qualitative Market Research, Vol. 6 No. 2, pp. 75-86. Sachan, A. and Datta, S. (2005), “Review of supply chain management and logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 35 No. 9, pp. 664-705. Saxton, T. (1997), “The effects of partner and relationship characteristics on alliance outcomes”, Academy of Management Journal, Vol. 40 No. 2, pp. 443-61. Scheubrein, R. and Zionts, S. (2006), “A problem structuring front end for a multiple criteria decision support system”, Computers & Operations Research, Vol. 33 No. 1, pp. 18-31. Seuring, S.A. (2008), “Assessing the rigor of case study research in supply chain management”, Supply Chain Management: An International Journal, Vol. 13 No. 2, pp. 128-37. Smith, M. (1986), “An introduction to repertory grids – part one”, Graduate Management Research, Vol. 3 No. 1, pp. 4-17. Starkey, K. and Madan, P. (2001), “Bridging the relevance gap: aligning stakeholders in the future of management research”, British Journal of Management, Vol. 12, pp. S3-S26 (special issue). Strauss, A. and Corbin, J. (1998), Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, 2nd ed., Sage, London. Tan, F.B. and Gallupe, R. (2006), “Aligning business and information systems thinking: a cognitive approach”, IEEE Transactions On Engineering Management, Vol. 53 No. 2, pp. 223-37. Trochim, W.M.K. and Donnelly, J.P. (2008), The Research Methods Knowledge Base, 3rd ed., Cengage Learning, Mason, OH. Vandermerwe, S. and Rada, J. (1988), “Servitization of business: adding value by adding services”, European Management Journal, Vol. 6 No. 4, pp. 314-24. Vargo, S.L. and Lusch, R.F. (2004), “Evolving to a new dominant logic for marketing”, Journal of Marketing, Vol. 68, pp. 1-17. Vargo, S.L. and Lusch, R.F. (2008), “Service-dominant logic: continuing the evolution”, Journal of the Academy of Marketing Science, Vol. 36 No. 1, pp. 1-10. Vermeulen, F. (2005), “On rigor and relevance: fostering dialectic progress in management research”, Academy of Management Journal, Vol. 48 No. 6, pp. 978-82. Voss, C., Tsikriktsis, N. and Frohlich, M. (2002), “Case research in operations management”, International Journal of Operations & Production Management, Vol. 22 No. 2, pp. 195-291. Wise, R. and Baumgartner, P. (1999), “Go downstream: the new profit imperative in manufacturing”, Harvard Business Review, September/October, pp. 133-41. Woodruff, R.B. (1997), “Customer value: the next source for competitive advantage”, Journal of the Academy of Marketing Science, Vol. 25 No. 2, pp. 139-53. Woodruff, R.B. and Gardial, S.F. (1996), Know Your Customer: New Approaches to Understanding Customer Value and Satisfaction, Blackwell, Oxford. Yan, W., Chen, C. and Chang, W. (2009), “An investigation into sustainable product conceptualization using a design knowledge hierarchy and Hopfield network”, Computers & Industrial Engineering, Vol. 56 No. 4, pp. 1617-26.

Yin, R.K. (2009), Case Study Research, Design and Methods, 5th ed., Sage, London. Zwick, R. (1988), “Another look at interrater agreement”, Psychological Bulletin, Vol. 103 No. 3, pp. 374-8. About the authors Keith Goffin is Professor of Innovation and New Product Development at Cranfield School of Management in the UK. Previously, he worked for 14 years in management roles at Hewlett-Packard. He has published extensively in journals such as the Journal of Product Innovation Management and the International Journal of Operations & Production Management. In 2010 he published (with F. Lemke and U. Koners) Identifying Hidden Needs: Creating Breakthrough Products (Palgrave) and the second edition of his best-selling book Innovation Management: Strategy and Implementation Using the Pentathlon Framework (with R. Mitchell) was also published. Keith Goffin is the corresponding author and can be contacted at: [email protected] Jawwad Z. Raja is Senior Research Fellow at the Centre for Business Performance, Cranfield School of Management. Previously, he worked at the School of Construction Management and Engineering, University of Reading. His current research interests focus on servitization, value co-creation, human resource management and communities of practice. He received his PhD from the University of London. Bjo¨rn Claes is Senior Research Fellow at the Centre for Business Performance at the Cranfield School of Management. His research focuses on the areas of product service integration (servitization), change management, supply chain performance measurement and supply chain behaviour. He received his PhD in Management from IE Business School in Madrid, Spain. Marek Szwejczewski is Director of the Best Factory Awards and Reader in Innovation and Process Management at Cranfield School of Management. Prior to joining Cranfield he worked for several years in various industry sectors, ranging from retailing to telecommunications and prior to coming to Cranfield was Marketing Manager with Motorola. He has authored numerous articles and reports on supplier management, innovation, manufacturing performance, and manufacturing strategy. He has advised a wide range of organisations, including the Department of Trade and Industry, Pilkington and Schering. His current research interests are world class manufacturing, technology transfer, innovation management, and international manufacturing strategy. Veronica Martinez is Principal Research Fellow in the Centre for Business Performance, Cranfield School of Management. Her research interests revolve around the field of measuring and managing business performance, supply chain management and value creation in strategy management. She has published in International Journal of Operations & Production Management, amongst others, and is author of several industry reports.

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Applying repertory grid technique 827

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0960-0035.htm

IJPDLM 42,8/9

Using the “documentary method” to analyse qualitative data in logistics research

828

Alexander Trautrims and David B. Grant Logistics Institute, University of Hull, Hull, UK

Received 2 December 2011 Accepted 16 January 2012

Ann L. Cunliffe Anderson School of Management, University of New Mexico, Albuquerque, New Mexico, USA, and

Chee Wong Logistics Institute, University of Hull, Hull, UK Abstract Purpose – This paper aims to examine the use of a qualitative data analysis technique, the documentary method, in the development of knowledge in logistics. The value of the method is illustrated through an example of its application in a study of in-store logistics processes at six leading European retail stores. Design/methodology/approach – Extant literature is outlined regarding philosophical underpinnings of the documentary method and is followed by an explanation of the method and its application. Finally, an illustration is provided of its adaptation and use in a logistics research project. Findings – Drawing on a social constructionist approach, the documentary method can add to the development of logistics research by providing rich descriptions of actual practice, problems and issues in logistic processes – compared with the stated goals of such processes. Research limitations/implications – The documentary method is not suitable for all areas of logistics research and will need certain adaptations and adjustments when transferred into particular research contexts. In addition, the research question, philosophical stance, and knowledge of qualitative methodologies will ultimately determine the appropriateness of the technique. Originality/value – The paper presents the first application of the documentary method in the field of logistics. Keywords Qualitative research, Social constructionism, Documentary method, Interview analysis, Distribution management, Research methods, Information management Paper type Conceptual paper

International Journal of Physical Distribution & Logistics Management Vol. 42 No. 8/9, 2012 pp. 828-842 q Emerald Group Publishing Limited 0960-0035 DOI 10.1108/09600031211269776

Introduction Acceptance of qualitative methods in logistics research has increased in recent years despite the dominance of quantitative methods. However, their application is still in its infancy and the wide range of available qualitative methods still relatively underexplored compared to other disciplines. Accordingly, this special issue is devoted to understanding and exploring how qualitative techniques can contribute to logistics research. To this end, this paper introduces a qualitative data analysis technique based on a social constructionist perspective – the “documentary method” – that originates from educational sciences. The paper discusses how this qualitative method can enrich our knowledge of the practical issues involved in logistics processes. The structure of the paper is as follows: First, extant literature is outlined

regarding the philosophical underpinnings of this method, followed by a positioning and explanation of the documentary method. Third, the paper discusses how data can be collected and analysed using this method through an example from a logistics research project, and finally, draws conclusions about its value to logistics research. We argue that the nature of the documentary method enables researchers to capture rich details of the micropractices of logistic processes while also providing an analytical protocol that draws generalisable principles from unique cases. It can therefore supplement a logistic researcher’s methods toolkit. Overview of qualitative research in logistics According to Mentzer and Kahn (1995) the logistics discipline is dominated by quantitative and positivist research, perhaps arising from the nature of logistics as an applied science with strong engineering roots. In addition, Stock (1997, p. 515) argued that as a comparatively young discipline, logistics does not have a rich methodological heritage and therefore relies “on the usage of concepts, definitions, theories, rules and principles from other disciplines”. Stock (1990, p. 4) also called for more involvement with other disciplines to broaden the horizon of logistics research and enable a logistics researcher “to escape the traditional perspective of the discipline”. Since then this horizon has broadened, with qualitative research perhaps entering a more established stage as evidenced by Sachan and Datta’s (2005) analysis of almost 450 articles from three leading logistics journals published between 1999 and 2003. They demonstrated that while the majority of publications still use quantitative methods, there is a shift towards more qualitative techniques. They concluded such increased acceptance might be caused by the growing maturity of the discipline and the focus on more “why” and “how” questions; consistent with the “boundary-spanning” issues represented in Kent and Flint’s (1997) sixth era in the development of logistics research. The discipline’s move towards other areas of the supply chain, rather than just goods flow and technical operations, is perhaps a response to the positivist focus on structural issues which often ignores those parts of the supply chain where human factors and interaction are an integral part of the process. This move is international, as Arlbjørn et al.’s (2008) analysis of Nordic PhD research in logistics demonstrates. They found researchers gravitating towards theory development rather than theory testing, and using more qualitative case-based approaches as identified earlier by Vafidis (2007). A number of researchers have suggested that logistics research can be enriched by the use of inputs from other disciplines and by qualitative methods such as case studies and action research (Na¨slund, 2002; Na¨slund et al., 2010). Yet at least two issues remain controversial in terms of their effect on the credibility of research: the “methodological soundness” and “rigour” of qualitative approaches (Halldo´rsson and Aastrup, 2003). Some researchers promote rigour through methodological triangulation and a combination of qualitative and quantitative, or inductive and deductive methods (Mangan et al., 2004). Within the broader and more established discipline of the sociology of knowledge, scholars acknowledge that rigour is based on being consistent in terms of using methods of data collection, analysis and theorising appropriate to a particular research paradigm (Cunliffe, 2010a). The paper draws on this premise to explain how the documentary method enriches understanding of how human perceptions and behaviours influence the operation and effectiveness of the logistics process.

Using the “documentary method” 829

IJPDLM 42,8/9

830

Research paradigms and philosophical underpinnings The paradigm debate, which addresses the philosophical underpinnings of various approaches to research, foreshadows the quantitative-qualitative question by claiming that the first question researchers should ask is not what method to use, but what paradigm to work from (Morgan and Smircich, 1980). A paradigm is “an entire constellation of beliefs, values and techniques, and so on, shared by the members of a given community” (Kuhn, 1996, p. 175), and paradigms change when an old Weltanschauung (“worldview”) cannot explain anomalies and research results. New paradigms can assist in highlighting new problems and looking at old problems from a different perspective, but can take time to become accepted by the research community as researchers tend to remain in paradigms that best match their personal Weltanschauung and contest new paradigms as being “non-rigorous”. At the most basic level, the two extremes are objectivism and subjectivism; with objectivism often being equated with quantitative and subjectivism with qualitative methodologies. The objectivist end of the continuum assumes there is an external reality into which humans are socialized that can be studied in a neutral way from an outsider/expert perspective. Researchers study structures, systems and processes as concrete realities (the “what” of social reality), and often use methods such as surveys, coded structured interviews, protocol analysis, grounded theory and realist-based ethnography to generate quantifiable and/or categorized data that can be generalized across contexts. Subjectivist researchers view realities as socially constructed and maintained in everyday interactions, and that people are reflective and intentional actors. They study culturally-situated meanings – participants’ perceptions and understanding of a situation – by using methods such as narrative analysis, participative action research, autoethnography and interpretive ethnography (Cunliffe, 2010b) to generate qualitative data. While their research focus lies on “how” realities are constructed by actors in a particular context, subjectivist researchers believe that an understanding of practical experience, issues and problems in one context offers insights of relevance to other contexts. There are many other positions that researchers can take between these two extremes (see Cunliffe (2010a) for a more in-depth explication), and continuing debate lies around the issues of the relevance and rigour of research conducted from the various paradigms. If, as subjectivist researchers believe, they are part of the process of meaning-making and therefore cannot be neutral observers, then how can their research be rigorous and trustworthy? This issue will be addressed later in the paper in relation to the documentary method. Broadening the field: the documentary method This paper suggests that a qualitative data analysis technique currently used in educational studies – the documentary method – has value for logistics research because it can offer insights into how human perceptions and actions influence the implementation and effectiveness of logistics processes. This method falls into Kent and Flint’s (1997, p. 26) “boundary-spanning” era because it focuses on understanding behavioural issues including “inter-functional cooperation and coordination”. Developed by Bohnsack (1989), initially to research the evolution of biographies in youth cultures in Germany, the documentary method is situated towards the subjectivist end of the continuum. It broadly falls within a social constructionist perspective

(Cunliffe, 2008), being based on the premise that our sense of what is real is a practical accomplishment, achieved through the contextual, embodied, ongoing interpretive work of people (Garfinkel, 1967) rather than in objective structures and systems. Based on the work of Bourdieu (1977) and Schwandt (2003), its philosophical roots lie in continental philosophy, in particular the Frankfurt School. Bohnsack developed the method as a way of accessing human action and the “system of meanings to which it belongs” (Schwandt, 2003, p. 296) and representing the actors’ explicit and tacit interpretations of their world by presenting “types” or patterns of meaning. Bohnsack (2010) considers that there are two levels of knowledge: the reflexive or theoretical knowledge and the practical or incorporated knowledge. The latter knowledge is also called “atheoretical knowledge” by Mannheim (1984) or “tacit knowledge” by Polanyi (1967) and assumes that there is an underlying mental structure in actions. This knowledge is “owned” by the research participants or actors and thus differs from an objectivist stance where the researcher alone has a privileged access to this knowledge. The documentary method spans the boundaries of these two levels of knowledge by tapping into the practical tacit knowledge of actors, which may even be outside the awareness of the participants themselves, as a means of extrapolating more general theoretical knowledge. It does so by focusing on the following (Bohnsack, 2010): . Practice – what are the common experiences, meanings, and routine practices that emerge and are maintained in the interactions of people around logistic processes? . On answering the question “how are logistics processes enacted in everyday interactions?” . On obtaining data based on actors’ experiential knowledge and interpreting, reconstructing and explicating the actors’ tacit meanings. . On the researcher constructing types from these meanings. The researcher should not try and remove all prejudgements, but rather take them into consideration as part of the process of making sense of others’ experience. . On participants’ language and ‘frames of orientation’ as a means of making actions and events understandable to others. Participants express themselves in a way that includes information about their perception of reality without necessarily explicitly mentioning or even defining such understanding. The researcher extracts this underlying knowledge from actors’ stories and language. Based on the above, the documentary method offers a way of bridging the relevance and rigour gap (Kieser and Leiner, 2009) – the claim that research cannot be both relevant (having practical value) and rigorous (building robust theory through controlled research design protocols). It focuses on relevance by exploring participant meaning, and rigour by providing a protocol for developing generalized types in a systematic way. Bohnsack addresses the issue of rigour arguing that actors’ subjective interpretations have to be transformed into an objective mode before they can be analysed. He uses a “reconstructive methodology” in which actors’ constructs are disassembled and the researcher interprets and reconstructs the constructs in three stages: telling, interaction and discourse. Hence, reconstruction considers not only subjectively intended meanings, but also has to consider the objective structure of meaning (Wagner, 1999). In this way, the documentary method bridges the gap between purely interpretive methods (e.g. some narrative methods, autoethnography, interpretive ethnography,

Using the “documentary method” 831

IJPDLM 42,8/9

832

hermeneutic analysis) which focus only on participant meanings and more positivist methods (e.g. surveys, coded interviews, boundary/object analysis) which take abstract generalizations from the data based on the researcher’s preconceived framework. This is particularly relevant to logistics where the focus often lies on the system and its elements and “how” the system is enacted in practice is often ignored. From a social constructionist perspective, systems are created, maintained and changed in the interpretations, tacit understandings and interactions of people. The documentary method provides a way of reconstructing tacit knowledge into atheoretical knowledge that can be generalized to other logistics contexts. By addressing rigour and relevance, the documentary method could be appropriately used as part of ethnographic, case study and grounded theory methodologies. The documentary method generates data from narrative interviews with organizational members. The narrative interview aims to collect data from open communication, which gives opportunities to participants to describe their everyday interactions and practices, their views and perceptions. This does not mean that data collection happens in a totally unstructured way, but that the interviewee responds to an initial narrative generating question, followed by the researcher probing certain aspects (Nohl, 2006). From a social constructionist perspective participants’ knowledge lies within their discourse and therefore data analysis uses interview transcripts to extract issues arising from the participants’ responses. Specifically, the analysis focuses on narrative pieces of the transcripts which are particularly rich (Nohl, 2006). From a qualitative perspective every narrative, and therefore also those derived through interviews, consists of two meaningful aspects: what is said, and how it is said (Fontana and Frey, 2003): Each narrative has two parts; a story and a discourse. The story is the content, or chain of events. The story is the ‘what’ in a narrative, the discourse is the ‘how’. The discourse is rather like a plot, how the reader becomes aware of what happened, the order of appearance of the events (Sarup, 1996, p. 17).

Consequently, data analysis takes the existence of these two levels into account. This differs from other qualitative methods such as conversation analysis and discourse analysis because the former focuses on the “how” – the mechanics of a conversation (e.g. the organization and sequencing of talk, turn taking), while the latter focuses on the “what” – the language and the structures of meaning. Thus, the documentary method is for researchers interested in exploring how ‘reality is shaped by respondents and what that reality might be. The narrative interview is analyzed and reconstructed in order to reveal the meanings expressed by the respondent, drawing on the “‘what” and the “how” of the interpretive process. This occurs as the researcher differentiates between respondents’ intrinsic understandings and those that emerge after probing. This is done by explicating common and uncommon expressions, phrases and language, rather than using positivist methods such as coded interviews and word counts (Meuser, 2006). Bohnsack differentiates between participants’ immanent (i.e. inherent) meaning and the documentary meaning, which is the researcher’s explication of meaning from the text. Thus, the interpretation has to take place in three steps as shown in Figure 1: the rephrasing interpretation of immanent meanings; the researcher’s reflective interpretation and reconstruction of meanings, topics, and their connection back to

Narrative Interviews with research participants

Transcribe interviews

Using the “documentary method” 833

Rephrasing interpretation

Analysis

Reflective interpretation

Comparative analysis

Create a typology

the context and the research aim; and a comparative analysis. Only in the reflective interpretation is the development of discourse and reality analysed (Wagner, 1999). Specifically, in the rephrasing interpretation (“what” is said in the interview), the researcher looks for topic changes and rephrases the text. By rephrasing the interview, the documentary method reveals different and more generalizable meanings that objectify processes within the data (Berger and Luckmann, 1966). The rephrasing interpretation is then followed by the reflective interpretation of “how” the topic is treated and how reality is constructed. To gain this insight, the researcher reconstructs the frame in which reality is created (Nohl, 2006). This frame is then used in the third step, the comparative analysis. The comparative analysis looks at how the interviewees responded to certain questions, the way they tackled the problem or the issue and in which sequence they constructed their response. Rigour is addressed by conducting sufficient interviews to include all aspects and dimensions of knowledge that are necessary for the comparison of interview content and the analytical development of relevant concepts and categories. This will depend on the topic under study and the organizational context. The aim is “conceptual representativity” (Przyborski and Wohlrab-Sahr, 2008) (having a range of views, job positions and perspectives, etc.) rather than an accurate representation of an objective reality. Different types are based on differences and commonalities of the characteristics and variables of an object can be extracted through them (Nohl, 2006). The researcher then creates a typology by grouping types together. Within one typology elements should be as similar to each other as possible and therefore one looks for similarities

Figure 1. Mapping the documentary method

IJPDLM 42,8/9

834

and aims for a high internal homogeneity; while at the same time aiming for external heterogeneity (Kluge, 1999). The aim for external heterogeneity proliferates the suitability for exploratory studies as it requires a wider coverage of different research objects/subjects. Adaptation of the documentary method to logistics: an illustrative case Having explained the philosophical origin and the process of the documentary method, it should not be forgotten that the method was developed for a research study about youth cultures in the social sciences, using open interviews for the data collection. Such an open interview structure might not be chosen by logistics researchers for practical reasons of access and the time available at the research site. The authors want to present an adaptation of the documentary method to the field of logistics, which puts more importance on the rephrasing interpretation and less on the discourse compared to Bohnsack’s (1989) original method. The adaptation was made in this case because the researchers were interested in the logistics processes themselves, which need the “what” facts from the rephrasing interpretation, as well as the tacit knowledge carried by logistics employees. Therefore, rephrasing interpretation and reflective interpretation became more interwoven in this study than suggested by Bohnsack. This highlights two advantages of the documentary method: it is adaptable to the specific research question and focus, and it is of particular benefit if the focus lies on practice – in particular the relationship between employees and logistic processes (i.e. relevance). A brief description of the study follows, bearing in mind that it is used purely for illustrative purposes: as a context from which to explicate the application, adaptation and value of the method. Two illustrative examples are drawn from a larger exploratory study of in-store logistics processes in retail companies. The aim of the study was to investigate how employees in a retail setting understand and interact with replenishment processes to improve on-shelf availability and reduce out-of-stocks. The documentary method was particularly appropriate to this research as it combined an interpretation of employees’ tacit knowledge with the ability to draw more generalizable insights about logistics systems and processes. Data was collected in 30 interviews with employees involved in-store replenishment across hierarchical levels from shop floor to senior management at six market-leading retailers, operating in four different retail sectors in three different European countries. The selection of interviewees was based on the criterion that they were informed and experienced respondents, which in this case means they were directly involved in the design or execution of in-store replenishment, and as such included store managers, head office merchandise managers or customer service assistants. Six companies were chosen to give a heterogeneous sample of retailers. While the documentary method originally calls for open interviews to allow respondents to express themselves fully and to set the flow of the interview, such a fully open interview was considered infeasible for this study. Negotiating access for interviews at retail companies was difficult enough, and one had to consider that participating companies would not accept the long interview durations usually anticipated with open interviews. This can be a disadvantage of the method, but one open to adaptation. Further, retail stores usually have a very hands-on operational work mentality, and the level of trust that would be needed for a fully open and deep interview seemed to be not achievable within such a short time frame. Also, after discussion with research

stakeholders, a sample size of 30 interviews was agreed and the conduct of in-depth interviews would have generated an unmanageable amount of data for the research project. Under these limitations, the study amended the research approach to using semi-structured interviews. The interviews were structured around three main topics derived from literature. Within these topic areas the flow was dominated by the participant and the researcher only used probes where appropriate and applicable. To give the respondents the freedom to express their own views and to make them lead the interview direction, the questions were open and rather general. The interviews were audio recorded and transcribed. Although the analysis of the full study takes the environment and interview situation into consideration, for the purposes of this paper the main focus is on the interview transcripts. For the analysis story-telling pieces are particularly valuable. These can be identified by certain narrative structures, such as the progression of events, for example connected through expressions such as “and then [. . .]”. The rephrasing part of the analysis looks at “what” the interviewee said without changing its meaning. The following example is cut out of a 70 lines response from a head office project manager at a British grocery retailer to the initial interview question: “What’s happening to a product from the point when it arrives at a store?” It provides a short example of rich description in that it offers a detailed comprehensible story of experience and events: So whoever is receiving it should check the high value against the delivery note and really check it off line by line. [...] And then you’ll have seen the back-door book, where they receive everything. So if it’s temperature checked, they’ll attempt to check it and just log it all down, all the driver’s details, his van details, time of delivery, temperature and who signed for it, that’s in the book. And then from there it depends on what replenishment activities you’ve got on. So whether generally in a big store, I’d say we replenish at a night, so when we’ve not got that many customers in. In a smaller store, I think it’s just load it in at backdoor, load it straight out onto shop floor and start merchandising. Rules are you bring your old stock forward and replenish to the back. [...] We have rules for presentation levels of what should you have on at all times in terms of presentation. And then you just replenish from the back. Bring the old stuff forward and put it to the back. And then we have the daily disciplines, that you’ve probably seen in store, like the gap checking or other reports, we have like low stock reports, high stock reports.

At the rephrasing analysis this translates into: “High value deliveries should be compared against the delivery sheet. All delivery details need to be put down in the back-door book. At a large store, replenishment happens at night, when fewer customers are in the store. At smaller stores the delivery is brought directly into the shop floor for replenishment. Rules exist for the presentation levels and to fill up from the back. Daily tasks are stock checks and gap checks”. Not only does the rephrasing analysis compress the content; it also forces the researcher to think about what the interviewee is really saying. What the interviewee is really saying can sometimes be debatable and the researcher might want to discuss the analysis with other researchers that are involved in the project. As an illustration the text piece “we have the daily disciplines” is considered. A standard dictionary would describe “disciplines” as “rules”. However, how binding rules are would depend on the way they are enforced. Therefore, the meaning of “disciplines” can vary between an order that must be adhered to and a recommendation. In this case the researcher chose to use it as “tasks” as the central organisation wanted stores to perform these checks, but in another

Using the “documentary method” 835

IJPDLM 42,8/9

836

part of the interview the respondent actually states that the organisation has no control whether the stores do it. Thus, the rephrasing into “tasks” was chosen. Through the rephrasing analysis the research finds out about the content or the “what” of the interview. Afterwards, the reflective analysis looks at the “how” level of the same text sample: {a} So whoever is receiving it should check the high value against the delivery note and really check it off line by line. [...] And then you’ll have seen the back-doorbook, where they receive everything. So if it’s temperature checked , they’ll attempt to check it and just log it all down, all the driver’s details, his van details, time of delivery, temperature and who signed for it, that’s in the book. {b} And then from there it depends on what replenishment activities you’ve got on. So whether generally in a big store, I’d say we replenish at a night, so when we’ve not got that many customers in. In a smal ler store, I think it’s just load it in at backdoor, load it straight out onto shop floor and start merchandising. Rules are you bring your old stock forward and replenish to the back. [...] {c} We have rules for presentation levels of what should you have on at all times in terms of presentation. And then you just replenish from the back. Bring the old stuff forward and put it to the back. And then we have the daily disciplines, that you’ve probably seen in store, like the gap checking or other reports, we have like low stock reports, high stock reports.

Looking at the sample text above, we have divided it based on content into three parts: the store receives a delivery; replenishment; and product availability on-shelf. Throughout the response the matter of control is one issue that arises. Words that express control are underlined in the above text sample; expressions that signal ambiguity or qualification are italicised. Parts {a} and {c} are dominated by control; whereas part shows more expressions of vagueness. This analysis is based on the researchers’ perception of the text and the situation; although the researchers could probe for the right understanding during the interview itself, it may be recommended to triangulate these perceptions in discussion with other researchers. The researchers may therefore interpret that the organisation’s central office, which the interviewee belongs to, has less control over the structure of replenishment processes that are going on in the store than about other areas of the business. The point where and when products are delivered is strongly controlled, however the flow of the product in the store goes out of the central office’s control; it then regains control by putting control procedures and measures in at the shelf. However, as we know from the rephrasing analysis, these control measures may or may not be followed by the store. From this excerpt, one can say that the respondent who is at the retailer’s organisational centre does not have a sufficient overview of the replenishment process that is happening in the stores. The delivery reception, which is the boundary between

central organisation and the store, can still be controlled. After that the store is in charge and the central organisation can only obtain a final measure of results and is unable to control the process execution itself. In order to identify typologies, one has to see the analysed data in contrast to other interviews. The second example text piece is therefore chosen from an interview with a head office store process manager at a Continental European do-it-yourself (DIY) retailer. The text piece is cut from his 51 lines answer to the same initial question as in the first example “What’s happening to a product from the point when it arrives at a store?” [...] We separate into eight to twelve [...] logistics streams. Depending on which logistics stream the product belongs to, it is either detail checked at reception or it is brought into the store without further checks. [...] Using the supplier and the log stream, the receiver knows whether the product needs to be checked or not. If it does not get detail checked; it is only overseen whether the number of packs is correct, whether the packs are undamaged. Then the entire thing is ticked off in the system [...] Going into sale means that it is brought to the individual departments by the reception [...] Does an entire pallet go into the store, then it happens very often that it is a mixed pallet. For example, it contains not only items for the plumbing area; it can also contain five or six buckets of paint. [...] Then the pallet is brought to the area of the product that needs to be taken off the top first. [...] We work with different colours. So that every one of us knows, when there is a yellow writing, he knows precisely which area it is. [...] We have separated our sales area into modules. Every metre is modulised [...].

The rephrasing analysis results in: “The company has eight to 12 logistics streams. The logistics stream determines to which detail a product needs to be checked at the store. After the product is given free, it is brought to the shop floor by the goods reception. Mixed pallets are brought to the department to which the most upper product on the pallet belongs to. The replenishment system uses a company-wide colour coding and standardised modules for its shelf plans”. Again, the reflective analysis follows for this text sample: [...] We separate into eight to twelve [...] logistics streams. Depending on which logistics stream the product belongs to, it is either detail checked at reception or it is brought into the store without further checks. [...] Using the supplier and the log stream, the receiver knows whether the product needs to be checked or not. If it does not get detail checked; it is only overseen whether the number of packs is correct, whether the packs are undamaged. Then the entire thing is ticked off in the system [...] Going into sale means that it is brought to the individual departments by the reception [...] Does an entire pallet go into the store, then it happens very often that it is a mixed pallet. For example, it contains not only items for the plumbing area; it can also contain five or six buckets of paint. [...] Then the pallet is brought to the area of the product that needs to be taken off the top first. [...] We work with different colours. So that every one of us knows, when there is a yellow writing, he knows precisely which area it is. [...] We have separated our sales area into modules. Every metre is modulised. [...]

Using the “documentary method” 837

IJPDLM 42,8/9

838

Much like the first excerpt, this second piece also contains words of control. Nevertheless, the first step of control when the product arrives uses harder words such as check and control. The second cloud of control is more about standardisation and ownership, with words like “every” and “knows”. The words underlined with a straight line show variance in the processes, where things can go further in different ways. The bold expresses when the product is moved forward through the process. In contrast to the first piece, the DIY interviewee showed that he is aware of all the different ways a product goes throughout the replenishment process and he explains in detail the processes that are applied in each case. This may be due to his position in the central office, where his main task is to monitor adherence to centrally given logistics processes in the stores and distribution centres. The interview response is structured in a way that the description of a product movement occurs with a cloud of words that express process separation. The step of potential process separation is then followed by words that suggest control/standardisation. Through this pattern of alternating movement options with process control, the company achieves standardisation and can afford variety in their processes without drifting into confusion. The second piece also does not use any words of uncertainty, i.e. where the participant is not completely sure what happens. Thus, process ownership and central control is achieved very differently in the two examples. In example one, the grocery retailer tries to gain control by monitoring reports and by giving rules. The DIY retailer in example two instead gives employees a manual for a system in which they operate. The processes are organised and prepared in a way that the store employees handle them independently in accordance with process guidelines. Overall, the two different types evolving from these two cases are that the grocery retailer aims to control processes by controlling results, whilst the DIY retailer controls the processes by owning them and controlling that the processes are conducted properly. As this example is only based on two smaller pieces from interviews, the entire data set would have to be investigated to confirm this typology, however that is out of scope for the purposes of this paper. Conclusions This paper offers an example of a qualitative research method drawn from education studies – the documentary method – that can be adapted and used for logistics research. Based on a social constructionist perspective, the method combines a subjective and objective stance: “subjectivity” in terms of garnering and interpreting participants’ perceptions, comments and narratives and interpreting how they construct their realities and “objectivity” in terms of the researcher’s construction of types and comparative analysis of typifications that allows a degree of generalization. The value of the method to logistics research is that it seeks to use the tacit knowledge of actors to help the researcher understand the complexities and operation of logistics processes. In other words, it directly addresses the issue of “relevance” to practice. The documentary method focuses not just on “what” the process or system is, or what it is intended to be – which is often the outcome of objectivist research, particularly when the researcher relies on observation and process documents – but also on “how” the process is constructed, perceived and enacted by those involved in the process. This offers a basis for identifying the issues and problems involved in the everyday operation of logistics. As demonstrated in the two examples, logistics

processes do not always operate as intended and control measures may or may not be followed by the store. The researchers gained an immanent perspective of those actors right at the centre of the process and could thus work with them to resolve problems. Compared to surveys and structured interviews, which usually centres around concerns of the researcher rather than those of employees, the documentary method provides rich data of relevance to practice because it engages the tacit knowledge of actors. This can provide a way of unearthing issues, concerns and problems that otherwise may not be immediately obvious to the researcher. Instead of testing theory or researcher-based hypotheses, this method develops theory inductively and systematically, which leads to knowledge development in logistics by providing a basis upon which researchers can re-theorize and practitioners re-structure processes. The two examples illustrate how control is understood differently at the two retail companies (the grocery retailer looked at results as a measure of control; the DIY retailer controlled the adherence to process execution guidelines by monitoring the process execution itself). This can be used to theorize control from different perspectives. Methodologically, the paper contributes to a call to use more qualitative methods in logistics research (Na¨slund, 2002; Mangan et al., 2004), Stock’s (1997) specific call for methodology knowledge transfers from other disciplines to the field of logistics, and represents research at the boundary-spanning era highlighted by Kent and Flint (1997). While the documentary method is not suitable for all research questions, particularly sheer mathematical problems in logistics, it is most suitable for exploratory research which focuses on answering the “how” and “why” questions. Further, this method suits social and human resources aspects in processes, which is becoming more important in research (Kent and Flint, 1997; Neumann and Dul, 2010). The method may also be criticized, from a positivist perspective, as lacking in rigour or validity discussed by Mentzer and Flint (1997). However, from a subjectivist perspective rigour is judged in different ways; for example Golden-Biddle and Locke (1993) argue similarly to Halldo´rsson and Aastrup (2003) that the “rigour” or trustworthiness of ethnographic work is based on the three criteria of: (1) Authenticity – has the researcher provided enough detail to show they have been there and gathered and analysed data in a discipline way? (2) Plausibility – does the connection between the description/data and the conceptualization make sense? (3) Criticality – does the text cause the reader to see the situation differently? These criteria can be applied as indicators of trustworthiness relating to the use of the documentary method, which offers rich description, a systematic analysis, and a development of types through a reconstructive methodology. As cautioned above the appropriateness of a selected methodology and data analysis method very much depends on the research question. The two examples provided herein were used to illustrate how the documentary method may be adapted to the logistics discipline to offer different conceptualizations of practice. In contrast to more positivist methodologies, which are seen to be rigorous but not necessarily relevant, the documentary method bridges the rigour-relevance gap by systematically developing theory from practice. This paper aims to encourage the transfer and use of new methods as a means of widening the knowledge base and relevance of logistics research.

Using the “documentary method” 839

IJPDLM 42,8/9

840

References Arlbjørn, J.S., Jonsson, P. and Johansen, J. (2008), “Nordic research in Logistics & Supply Chain Management: an empirical analysis”, International Journal of Physical Distribution & Logistics Management, Vol. 38 No. 6, pp. 452-74. Berger, P.L. and Luckmann, T. (1966), Social Construction of Reality: A Treatise in the Sociology of Knowledge, Anchor Books Doubleday, New York, NY. Bohnsack, R. (1989), “Generation, Milieu und Geschlecht – Ergebnisse aus Gruppendiskussionen mit Jugendlichen” (“Generation, milieu and sex – results from group discussions with adolescents”), in Fuchs, W., Kohli, M. and Schu¨tze, F. (Eds), Biographie und Gesellschaft, Band 8, LeskeþBudrich, Opladen, pp. 3-421. Bohnsack, R. (2010), “Documentary Method and Group Discussions”, in Bohnsack, R., Pfaff, N. and Weller, W. (Eds), Qualitative Analysis and Documentary Method in International Educational Research, Barbara Budrich, Opladen, pp. 99-124. Bourdieu, P. (1977), Outline of a Theory of Practice, Cambridge University Press, Cambridge. Cunliffe, A.L. (2008), “Orientations to social constructionism: Relationally-responsive social constructionism and its implications for knowledge and learning”, Management Learning, Vol. 39, pp. 123-39. Cunliffe, A.L. (2010a), “Crafting research: Morgan and Smircich 30 years on”, Organizational Research Methods, available at: http://orm.sagepub.com/content/early/2010/07/23/ 1094428110373658 (accessed 20 August). Cunliffe, A.L. (2010b), “Retelling tales of the field: in search of organizational ethnography 20 years on”, Organizational Research Methods, Vol. 13, pp. 224-39. Fontana, A. and Frey, J.H. (2003), “From Structured Question to Negotiated Text”, in Denzin, N.K. and Lincoln, Y.S. (Eds), Collecting and Interpreting Qualitative Materials, London, Sage, The Interview, pp. 61-106. Garfinkel, H. (1967), “The practice of reflexive sociology”, in Garfinkel, H. (Ed.), Studies in Ethnomethodology, Prentice-Hall, Englewood Cliffs, NJ, pp. 1-34. Golden-Biddle, K. and Locke, K. (1993), “Appealing work: An investigation of how ethnographic texts convince”, Organization Science, Vol. 4 No. 4, pp. 595-617. ´ . and Aastrup, J. (2003), “Quality criteria for qualitative inquiries in logistics”, Halldo´rsson, A European Journal of Operational Research, Vol. 144 No. 2, pp. 321-32. Kent, J.L. and Flint, D.J. (1997), “Perspectives on the evolution of logistics thought”, Journal of Business Logistics, Vol. 18 No. 2, pp. 15-29. Kieser, A. and Leiner, L. (2009), “Why the rigour-relevance gap in management research is unbridgeable”, Journal of Management Studies, Vol. 46 No. 3, pp. 516-33. Kluge, S. (1999), Empirisch begru¨ndete Typenbildung: Zur Konstruktion von Typen und Typologien in der qualitativen Sozialforschung (Empirically Based Construction of Types: About the Constructions of Types and Typologies in Social Research), LeskeþBudrich, Opladen. Kuhn, T.S. (1996), Structure of Scientific Revolution, The University of Chicago Press, Chicago, IL. Mangan, J., Lalwani, C. and Gardner, B. (2004), “Combining quantitative and qualitative methodologies in logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 34 No. 7, pp. 565-78. Mannheim, K. (1984), Konservatismus. Ein Beitrag zur Soziologie des Wissens (Conservatism – A Contribution to the Sociology of Knowledge), Suhrkamp, Frankfurt aM.

Mentzer, J.T. and Flint, D.J. (1997), “Validity in Logistics Research”, Journal of Business Logistics, Vol. 14 No. 1, pp. 27-42. Mentzer, J.T. and Kahn, K.B. (1995), “A framework of logistics research”, Journal of Business Logistics, Vol. 16 No. 1, pp. 231-50. Meuser, M. (2006), “Rekonstruktive Sozialforschung” (“Reconstructive social research”), in Bohnsack, R., Marotzki, W. and Meuser, M. (Eds), Hauptbegriffe qualitativer Sozialforschung (Main Concepts of Qualitative Social Research), UTB Verlag, Stuttgart, pp. 140-2. Morgan, G. and Smircich, L. (1980), “The case for qualitative research”, Academy of Management Review, Vol. 5, pp. 491-500. Na¨slund, D. (2002), “Logistics needs qualitative research – especially action research”, International Journal of Physical Distribution & Logistics Management, Vol. 32 No. 5, pp. 321-38. Na¨slund, D., Kale, R. and Paulraj, A. (2010), “Action research in supply chain management – a framework for relevant and rigorous research”, Journal of Business Logistics, Vol. 31 No. 2, pp. 331-55. Neumann, W.P. and Dul, J. (2010), “Human factors: spanning the gap between OM and HRM”, International Journal of Operations & Production Management, Vol. 30 No. 9, pp. 923-50. Nohl, A. (2006), Interview und dokumentarische Methode –Anleitungen fu¨r die Forschungspraxis (Interview and Documentary Method – Manuals for the Research Practitioner), VS Verlag fu¨r Sozialwissenschaften, Wiesbaden. Polanyi, M. (1967), Tacit Dimension, Routledge & Kegan Paul, London. Przyborski, A. and Wohlrab-Sahr, M. (2008), “Qualitative Sozialforschung – Ein Arbeitsbuch (Qualitative Social Sciences Research – A Workbook)”, Oldenbourg, Munich, pp. 25-51. Sachan, A. and Datta, S. (2005), “Review of supply chain management and logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 35 Nos 9/10, pp. 664-705. Sarup, M. (1996), Identity, Culture and the Postmodern World, Edinburgh University Press, Edinburgh. Schwandt, T.A. (2003), “Three epistemological stances for qualitative inquiry: interpretivism, hermeneutic and social constructionism”, in Denzin, N.K. and Lincoln, Y.S. (Eds), The Landscape of Qualitative Research: Theories and Issues, Sage, London, pp. 292-331. Stock, J.R. (1990), “Logistics thought and practice: a perspective”, International Journal of Physical Distribution & Logistics Management, Vol. 20 No. 1, pp. 3-6. Stock, J.R. (1997), “Applying theories from other disciplines to logistics”, International Journal of Physical Distribution & Logistics Management, Vol. 27 Nos 9/10, pp. 515-39. Vafidis, D. (2007), Approaches for Knowledge and Application Creation in Logistics: An Empirical Analysis Based on Finnish and Swedish Doctoral Dissertations Published Between 1994 and 2003, Turku School of Economics, Turku. Wagner, H. (1999), “Rekonstruktive Methodologie: George Herbert Mead und die qualitative Sozialforschung” (“Reconstructive Methodology:George Herbert Mead and Qualitative Social Research”), in Bohnsack, R., Lu¨ders, C. and Reichertz, J. (Eds), Qualitative Sozialforschung (Qualitative Social Research), Band 2, LeskeþBudrich, Opladen, pp. 7-96.

Using the “documentary method” 841

IJPDLM 42,8/9

842

Further reading Burrell, G. and Morgan, G. (1979), Sociological Paradigms and Organisational Analysis: Elements of the Sociology of Corporate Life, Ashgate, Aldershot. About the authors Alexander Trautrims (BA Hons, MSc, PGDip, PhD) is a Lecturer in Logistics and Supply Chain Management in the Logistics Institute at the University of Hull, Kingston-upon-Hull, UK. Before coming to the University of Hull, he received his MSc in Logistics and Supply Chain Management at the Logistics Research Centre, Heriot-Watt University, Edinburgh, UK. His research focuses on retail logistics, particularly issues regarding on-shelf availability and store processes. For his PhD he investigated in-store processes across European retailers. Alexander Trautrims is the corresponding author and can be contacted at: [email protected] David B. Grant (BComm, MBA, MSc, PhD) is Professor and Director at the Logistics Institute, University of Hull, UK. His research interests include customer service, satisfaction and service quality, retail logistics, logistics and supply chain relationships, and sustainable logistics. His recent applied research has investigated on-shelf availability and out-of-stocks, total loss and waste in food retailing, forecasting and obsolete inventory, service quality of internet retailers, and consumer logistics and shopping convenience. He has over 125 publications in refereed journals, books and conference proceedings, is on six journal editorial boards, and is a member of the Council of Supply Chain Management Professionals Education Strategies Committee and the British Retail Consortium’s Storage and Distribution Technical Advisory Committee. Ann L. Cunliffe (BA Hons, MPhil, PhD) is ASM Alumni Professor at the University of New Mexico and Visiting Professor at Leeds University, Hull University and Strathclyde University. She obtained her MPhil and PhD degrees from Lancaster University. Her research interests lie in examining the relationship between language and collaborative, responsive and ethical ways of leading and managing organizations. She is currently Editor-in-Chief of Management Learning and organizer of the Qualitative Research in Management and Organization conference. Chee Wong (BEng, MSc, PhD, PGCHE) is a Senior Lecturer in Operations Management from the University of Hull Business School and Logistics Institute. He teaches international logistics, logistics and supply chain management, and operations management. He researches supply chain strategy, sustainable logistics, third party logistics, global logistics, and operations strategy.

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0960-0035.htm

Towards a methodology for studying supply chain practice

Studying supply chain practice

Benedikte Borgstro¨m Technology Management and Economics, Industrial Marketing, Chalmers University of Technology, Go¨teborg, Sweden and Jo¨nko¨ping International Business School, Entrepreneurship, Strategy, Organization and Leadership, Centre of Logistics and Supply Chain Management, Jo¨nko¨ping, Sweden

843 Received 12 September 2011 Accepted 20 September 2011

Abstract Purpose – The purpose of this paper is to describe and analyse a methodology for studying the practice of logistics and supply chain management (SCM), namely the mystery methodology. Design/methodology/approach – Many SCM and logistics researchers share methodological presuppositions concerning the “reality status” that are usually unspoken and deviating from presuppositions of the methodology here investigated. By proposing an alternative methodology, the paper stimulates further ideas that will advance the discussion of research methodologies in SCM. Findings – The methodology facilitates exploration and elaboration of anomalies in theory and in practice. The mystery construction process facilitates SCM research in three ways: as a consistent methodology for practice research; for learning and responsiveness to new insights; and with the problem of bounding the case. Research limitations/implications – The methodology is delimited by its constructivist/ interpretivist assumptions in order to provide accurate representations. It makes possible richer insights into, and the meaning of, SCM phenomena in which social action can be understood in an intelligible way. Practical implications – Construction of mysteries opens up for learning during the research process by refining the research question and the literature base. Under the assumption that the researcher is knowledgeable about the literature in a variety of areas, the methodology implies rigour and relevance in SCM research. Originality/value – This paper is grounded in contemporary supply chain integration problems and develops the discipline further with its alternative approach in which practice of action is in focus. Keywords Supply chain practice, Mystery methodology, Qualitative research, Research methods, Supply chain management Paper type Research paper

Introduction In this paper, I draw on the fact that qualitative research is gaining acceptance to advance knowledge of logistics and supply chain management (SCM) (Craighead et al., 2007; Halldo´rsson and Arlbjørn, 2005; Na¨slund, 2002; Sachan and Datta, 2005; Solem, 2003). The movement includes a more interpretative notion of theory and subjectivist methodological assumptions (Solem, 2003) and engagement in inductive as well as abductive research (Spens and Kova´cs, 2006), in which understanding and sense-making are generative for logistics and SCM knowledge (Nilsson, 2005). Some methodological knowledge development challenges of logistics and SCM remain. Na¨slund (2002) argues that logistics and SCM problems are often messy real-world problems in which the process of action needs to be a main subject of research.

International Journal of Physical Distribution & Logistics Management Vol. 42 No. 8/9, 2012 pp. 843-862 q Emerald Group Publishing Limited 0960-0035 DOI 10.1108/09600031211269785

IJPDLM 42,8/9

844

Perspectives that open up for such social practice engagement will strengthen logistics and SCM research in terms of new research directions and managerial relevance (Gammelgaard, 2004), and there are very few, if any, examples of these in the field (Craighead et al., 2007; Sachan and Datta, 2005). Social practice perspectives have increased in importance in Business Administration studies in order to understand, for example, organizing rather than organizations (Czarniawska, 2004). A practice study implies that methods and units of analysis are used to understand how a process of change is performed and what it means (Chia and MacKay, 2007). Both action research and ethnography are proposed in order to get close to, understand and make sense of logistics problems (Nilsson, 2006; Na¨slund, 2002). Almost any method or combination of methods is appropriate as long as it is used in a contextualized, reflexive and interpreted manner (Chia and MacKay, 2007; Hatch and Yanow, 2008). Consequently, methodological consciousness is critical to develop logistics and SCM knowledge. There are many ways to address methodology, and in suggesting one way that I have found useful I hope to stimulate further ideas that will advance the discussion of research methodologies in SCM (in line with Halldo´rsson and Arlbjørn, 2005). I aim to describe and analyse a methodology for studying the practice of logistics and SCM, namely the mystery methodology. The logic of mystery methodology is consistent with constructivist ontological assumptions and interpretative epistemological assumptions (Alvesson and Ka¨rreman, 2007a). The methodology relates to abduction in that theorizing is inspired by empirical material and theory in order to explore a phenomenon in a reflexive manner (Alvesson and Ka¨rreman, 2007b; Carroll et al., 2008; Dubois and Gadde, 2002; Ka¨rreman and Alvesson, 2009; Spens and Kova´cs, 2006; Spicer et al., 2009). In the following I will describe the essentials of a practice study and what it means in terms of rigour and relevance to the SCM field. Thereafter the mystery methodology and its working method are defined and analysed. Then specific challenges related to a supply chain study of integration practice are further elaborated, because the approach implies a more interpretive notion of theory as well as engagement in ongoing supply chain problems. A typical supply chain issue relates to integration, and the theoretical underpinned explanation of the approach is next made concrete by an illustration of the initial steps in a research process of integration in practice (a car manufacturing supply chain). Then, a discussion follows of the mystery construction as a theorizing device in general and in relation to practice of supply chain integration. Thereafter contributions, limitations and future research opportunities are identified regarding how the methodology works to carry out rigorous SCM research that is based on relevant topics. A methodology for supply chain practice Halldo´rsson and Arlbjørn (2005) argue that the manufacturer perspective is most common in SCM studies and question whether theorizing generated by the manufacturer’s perspective also applies equally among other members of the supply chain. Perspectives delimit and enable usage of SCM concepts. It is possible to add other perspectives by different types of actors as focal units of analysis or to use different types of supply chains as such units. Another possibility is to use social practice as a focal unit of analysis. These and other perspectives are possible in a constructivist-interpretivist epistemology in which knowledge is context-specific and

possible to interpret (Hatch and Yanow, 2008). Meaning is constructed, and it is important to be transparent and give a clear representation of the construction process. Alvesson and Ka¨rreman’s mystery methodology draws on Asplund’s (1970) discussion about the ability to see a thing as something and interpret its meaning. Social theories are rarely used in logistics and SCM research. Dynamics and complexity of real-world problems pose methodological challenges (Na¨slund, 2002) that social practice studies might respond to by giving meaning. A practice study aims to understand how a process of change is performed and what it means, i.e. its performation and multiplicity (Chia and MacKay, 2007; Law, 2004, 2008). The social in sociologically informed practice studies is important because from that perspective neither the object “supply chain integration” nor the object “SCM” exists as a starting point; these are objectified over time (Carter et al., 2008a). Along that line of argumentation, SCM and integration should be seen as the projection of practices with their dynamics and complexity. By seeing SCM as a discourse, it is possible to critically examine meanings such as the official one, that of top management and that of others (Carter et al., 2008b). The researcher’s role would then be to construct interpretations demonstrating dynamics and complexity of SCM that originate in different meanings. The contribution is better understanding of the phenomena. How can we achieve that? A practice study is a specific process study (Chia and MacKay, 2007). Few practical guidelines are available from sociological sources, such as Latour (1987), Giddens (1979) and Law (2004). The social practice approach takes off from a definition of practices, a set of assumptions and a great interest in empirical matters. Practices are seen as consistent patterns of actions, and the level of the practice is the only matter of relevance. Practices influence and are influenced by individuals, by organizations, by the environment and by the concept that development should be the unit of analysis for practice researchers (Chia and MacKay, 2007). The critical power of practice in relation to supply chain integration is in using practice as an epistemology to articulate integration knowledge as a practical accomplishment. In the epistemology of practice, knowledge is situated and activities are contextual (Gherardi, 2009). Is it possible to expect rigour of a methodology that makes sense of practice? Rigour and relevance of the approach An approach based on the epistemology of practice and the logic of mystery methodology is rare in the contemporary SCM field. Therefore, SCM research objectives in terms of rigour and relevance will be discussed before the mystery methodology is explained. Is it possible for researchers to execute rigorous SCM research that is based on relevant and important topics (Carter, 2008a)? Flynn (2008) proposes that it is a question of engaging in ambiguous matters supported by literature bases from a variety of areas, including operations/SCM, marketing and strategic management. By building on useful foundations, we can develop high quality academic research papers with engagement in questions that are of critical importance for supply chain professionals. Flynn (2008) strongly believes that rigour and relevance are possible with regard to contemporary SCM problems. The potential is in framing. Interesting research questions depend on a strong theoretical foundation. A rigorous and thorough study has a match between a body of strong theory and the ambiguous empirical matter used to guide research. Theory development via such a dialogue between empirical material and theory is based on an abductive logic.

Studying supply chain practice

845

IJPDLM 42,8/9

846

A small but growing proportion of logistics and SCM research has an abductive research approach and an interpretative notion of theory, which facilitates theory building in the field (Solem, 2003; Spens and Kova´cs, 2006). Abductive research is often reported as being deductive, which poses problems for the logistics and SCM discipline because a failure to describe the study in a rigorous manner means that it is difficult to assess how conclusions were reached (Spens and Kova´cs, 2006). Rigour of an abductive research approach means that choices and links in the research process need to be transparent and coherent (Dubois and Gibbert, 2010). Mentzer (2008) argues that rigour is about constant examination of the claims that research makes. Awareness is needed, because in the use of the appropriate theories and methods, it is important to recognize that all research is flawed. Depending on methodological research strategy applied, different research objectives are at stake. Mystery construction falls short when it comes to generalizability of findings. The research objective of mystery construction is to propose a more useful understanding for practitioners and scholars. The better understanding is in the form of “aha” – experiences in contrast to precise measurements. In line with Flynn (2008) and Mentzer (2008) argues that the job is to address timely issues with timeless theories and perspectives. Relevance is in meaningful framing of upcoming problems and rigour by a conscious research process. The methodology of mystery construction is empathetic for SCM researchers, in order to execute rigorous supply chain practice research and is next to be explained. Mystery construction The process of creating a mystery and solving the mystery is used to illustrate the abductive process and clarify involved decisions (Alvesson and Ka¨rreman, 2007a). Overall, the mystery methodology is a combination of critical reflection of outliers in a theory, theory-driven disclosure and a specific procedure of working with knowledge breakdowns for the sake of theorizing based on an open and emergent research process (Figure 1). The research process can be associated with any kind of methodology that includes fieldwork research, but empirically varied and rich studies or participatory research are more likely to lead to discovery of breakdowns (Alvesson and Sko¨ldberg, 2000). The approach encourages theoretical development through scrutinizing what is known. It works as interplay between theory, researcher subjectivity (in the form of interpretative repertoire) and empirical findings. The theoretical contribution is in the move from breakdown to mystery and the process involves explicitness of: . relevance of findings; . problems in theoretical background; and . proposed new understanding. Relevance of findings The research methodology (Figure 1) takes off from the researcher with an open attitude based on knowledge in the area. In an initial state of a study when the researcher is familiarizing him- or herself with the setting, unexpected empirical material is encountered. The material cannot plausibly be accounted for by available theory, so the process starts with a breakdown of what is known. This is surprising in the meaning that a previous concept or an assumption is falsified (Weick, 1989).

Studying supply chain practice

847

Source: Alvesson and Kärreman (2007a, p. 1271)

The surprise is a breakdown in research understanding. More systematic work is needed both by going back and forth to the research site and by engaging in additional literature. Thus, abduction is used to interpret and construct the mystery in a sensitive and reflexive way, which is a systematic combination of theory, empirical material and imaginative resources (Dubois and Gadde, 2002). Problems in theoretical background After finding the surprise, the researcher is to formulate propositions or preliminary interpretations of theory and outline a broad relevance of the finding, i.e. to point out why it is surprising and to identify related problems and some hints of a new understanding through a mystery formulation (Alvesson and Ka¨rreman, 2007a). The abductive research process in which openness is of special importance might construct and solve the mystery. Openness to alternative interpretations is facilitated by familiarity with a set of perspectives, concepts and themes as a kind of interpretive repertoire (Alvesson and Sko¨ldberg, 2000). Thus, reinterpretation of problems is used to create new understanding of the phenomena. The novel insights from the combination is to solve the mystery by using new concepts, new theoretical frameworks or new metaphors (Alvesson and Ka¨rreman, 2007a). Another outcome is to formulate new research tasks and reformulate the mystery. Developing the (re)solution of the mystery is important to theoretical abstraction and to understanding the productive practical use of it. Theorizing is viewed as theoretical reasoning in which theoretical frameworks as well as empirical work are parts used (see discussion in Sutton and Staw (1995) and Weick (1995).

Figure 1. The research process: decision tree for mystery-focused research

IJPDLM 42,8/9

848

Proposed new understanding The outcome, i.e. the theories, should provide understanding and be useful. Subjectivity is important in mystery discovery since the discovery is a constructed dialogue between a researcher’s interpretative repertoire and empirical findings. The expression of what is interesting is essential and relates to situational problems, connections to practice and theoretical perspective. Interesting insights are the source that is critically reflected upon based on theory and empirical material. In this sense, they are a substitute to validation as they indicate that a research assumption is reconsidered (Lindblom, 1987; Weick, 1989). Interest is the informed feeling, as, for example, when a researcher returns from an interview and tells the others in the research group, in a few sentences, about what was intriguing (rather than giving a summary of the interview). Sometimes these signs of mystery stand up to scrutiny and can be formulated as a mystery. The outcome is based on a practice-oriented problem that is interpreted with various existing theoretical concepts and further explored for knowledge. What was learnt from that analysis is the contribution. Its boundaries emerge during the research process. The boundaries are based on the problem and follow the action, which is of relevance for practice studies. Thus, the outcome is a thick description based on and bounded by the mystery development. Assessing the theoretical contribution of mystery-focused research The possibilities offered by the mystery methodology are related to a consistent logic and practice-based bounding of the phenomenon. First, the consistency relates to avoiding mixing realism and constructivism methodology – frequently, both theoretical and empirical work contains arguments built on internally inconsistent logic among the constructive/interpretative ontological and epistemological presuppositions and method used (Hatch and Yanow, 2008). A method that tries to mirror social reality by more and more empirical material is consistent with realist presuppositions, while the mystery methodology is interpretative both in terms of data generation during the fieldwork and in terms of analysis construction (Alvesson and Ka¨rreman, 2000, 2007a; Ka¨rreman and Rylander, 2008). Second, the practice-based bounding of the phenomenon relates partly to the starting point of the mystery and partly to the constructive approach. Mystery construction is a way to create insightful theories for understanding. In a complex and dynamic setting of problem-centred research of real-world logistics, this methodology is an opportunity to open up for learning during the research process and creativity in the search for meaning. Meaning, talk and practice interact and have to be interpreted (Alvesson and Ka¨rreman, 2000). Research methodologies that initially bracket the phenomena are likely to miss influential interactions with the surroundings. A concept such as integration in practice involves different actors over time and co-evolves with other processes. Methodological creativity of constructive methods in order to understand happening, especially in a complex supply chain context, is needed. Some restrictions of the mystery methodology relate to its emergent research process. While some qualitative approaches are wide in scope from the start and deliberately attempt to categorize and compare, this approach attempts to explore a challenged assumption, that is, it is emergent and narrow from the start. It emerges along the way that the researcher’s interpretative repertoire allows it to take. Another researcher would most likely use a different interpretative repertoire. Also, there is no obvious starting

or end point. What is surprising and challenging in one perspective is of minor importance in another. Further, what is intelligible in one perspective is obscure in another. This is in line with the constructive/interpretative presuppositions but challenges a common view that empirical or theoretical objective saturation must be achieved. How, then, will the mystery construction facilitate SCM research? A concrete example is to investigate what implies a practice-based construction about supply chain integration. Methodological implications for a study of integration in practice Different ways of seeing There is an increased interest in integrative practices in logistics and SCM literature (Fabbe-Costes and Jahre, 2007; Fawcett and Magnan, 2002; Sandberg, 2007; Storey et al., 2006; Tan, 2002). Despite the increased interest, little attention is directed to practice studies (Gammelgaard, 2004) in contrast to development in related streams of literature such as organizing (Czarniawska, 2004), strategizing/strategy-as-practice (Johnson et al., 2007) and marketing-as-practice (Araujo et al., 2008). The lack of practice-based SCM studies is problematic in order to understand and make sense of the criticized concept of supply chain integration (Fabbe-Costes and Jahre, 2007). Practice-based SCM studies do, however, imply some methodological challenges that relate to their qualitative and in particular their constructivist-interpretivist character. For example, it is difficult to publish qualitative research, probably because plausible evaluation criteria are not generally accepted (Dubois and Gibbert, 2010; Na¨slund, 2002; Pratt, 2008). The problem is that a qualitative study can be used for many different knowledge claims; those that are based on a realist approach are more accepted than those that are based on an interpretivist approach (Dubois and Gibbert, 2010; Na¨slund, 2002). The difference between realist and interpretivist approaches might be illustrated by a painting metaphor of two very different painters, namely Rembrandt and Pollock and their modes of painting (Hatch and Yanow, 2008). Rembrandt paints an object as an exact representation in clear contrast to Pollock’s multifaceted action painting of effects of an object. Both painters do high quality painting because the view of the object is consistent with the painter’s techniques. The paintings are evaluated based on what they accomplish and how they are painted. Evaluation criteria used for realist approaches are built on their internally consistent logic and cannot be superimposed on the interpretative approach. The point that Hatch and Yanow (2008) make is that there are differences between realists and interpretivists regardless of whether they paint or do research. A case study that aims to understand practice (a Pollock way of research) needs a constructivist-interpretivist methodology, in order to follow the action in a consistent way. Seeing integration as a process of action challenges the capturing of integration, the boundaries of integration and, consequently, the conceptualization of integration. In addition, the supply chain focus implies more methodological challenges that relate to the messy real-world problems of supply chain integration. Supply chain integration related challenges The practice perspective is challenging because of interrelatedness in development. Acknowledging that supply chain integration is from practices rather than being a property of an organization or a supply chain opens up for at least four dimensions of integration that complicate its study (Table I).

Studying supply chain practice

849

IJPDLM 42,8/9

Supply chain integration in practice-related challenges

850

Integration as projection of practices (Carter et al., 2008a, b) rather than as related to an organization’s properties (Chia and MacKay, 2007)

Methodological challenges Consistency (draw on sociologically informed practice studies)

A practice study is neither a case Boundaries of the case of a fixed entity, such as an organization, nor a case of a process but needs to be bounded by the process of action. The unit of analysis is the case

Table I. Methodological challenges

In practice the projected supply chain integration is situational. The interrelatedness of micro and macro issues is involved based on the flat ontology (Chia and MacKay, 2007; Carter et al., 2008b)

Capturing integration

The assumption of interrelatedness and the flat ontology make micro, meso and macro analysis and static categorizations such as linear explanations inconsistent in the epistemology of practice

Analysis and theorizing

Response offered by interpretative construction Integrative development is constructed, for example, by focusing on an idea and following the action or by focusing on a projection’s interactions Flexibility in approach. Progressive boundaries emerge with empirical issues and interpretative frames and are an explicit researcher decision. Transparency of the research process that decides the case Explorative approach in which the interrelatedness of people, their action, projections of integration and context can be followed together with its effects. Thereby the performative supply chain integration together with outcomes is up to scrutiny Theorizing with outcomes in terms of learning and temporal meaning-making. Situational learning for “aha” experiences while being aware of its limitations (related to boundaries and capturing) in which the content and process of supply chain integration development are seen as interrelated, dynamics and complexity are explored and ambiguity is critically examined

First of all, the how and why of integration are treated as interdependent issues. The contents of integration and integrative processes are related. Second, there is a dynamic dimension involving, for example, how most organizations face both success and failure in their integration and coordination. Third, action in supply chains might be directed towards multiple goals such as efficiency and effectiveness, sometimes in processes that interact and influence each other. Finally, the dimension of complexity is needed in order to understand the interrelatedness of macro phenomena and micro phenomena (such as relations among group, organizational and network issues). These four dimensions can appear overwhelming, especially in view of the research norm that says that “the researcher is in control, producing a linear, coherent study, where research questions, framework, fieldwork, empirical results, and conclusions follow a rational procedure” (Alvesson and Ka¨rreman, 2007a, p. 1277). A methodology that follows these intertwined dimensions to make sense and create meaning of them is needed, in order to achieve consistency.

Integration within and beyond the organization would be difficult to fit into traditional ethnography with its distinct categorizations. In practice, categories such as micro, macro, process and content are artificial, but the transformation of the phenomena is of importance, such as how coordination develops. Most often boundaries in supply chain network studies delimit important aspects of practice. Halldo´rsson and Arlbjørn (2005) investigated the methodology of 71 SCM articles and found that most studies were designed with the manufacturer as a navigating point. What are the alternatives used in social studies? Latour (2005), for example, argues that it is the action that should be followed rather than an actor, and therefore boundaries are flexible. Summarizing, supply chain integration in practice poses specific challenges that correspond to methodological challenges of different ways of seeing (Table I) that are important to a rigorous supply chain integration study. Next, the mystery methodology is put up against the methodological implications for a study of integration in practice, in order to assess the methodology’s plausible responses. Response from mystery-focused exploration The mystery methodology is a way to see the unexpected and a methodology for creative theorizing based on abduction. The mystery methodology is an opportunity that requires the study to be loosely bounded, i.e. progressive, emergent and open, to empirical material as well as to theory. Flexibility is important to develop and explore new ideas, and the process of gathering additional empirical material is used to construct and solve mysteries. The construction and solving of the mystery is an explicit decision process. An empirical or theoretical surprise continues to surprise, despite further examination, and becomes a breakdown of knowledge. Flexible research design is epistemologically consistent with sensitive construction (Alvesson and Ka¨rreman, 2007a). Research methodologies that initially bracket the phenomena are likely to miss influential interactions with the surroundings because the brackets diminish possibilities to be open and flexible to learning and influences over time. Integration is ongoing and might involve different actors over time, and it might co-evolve with other processes. If the closely interwoven dualities of micro/macro and process/content are separated, then some of the meaning of the integration is lost. The mystery methodology follows the knowledge breakdown into those dualities, important to the construction, without fixed boundaries. Thus, the performance of integration is captured. To bound integration as a mystery is to focus on knowledge breakdowns and exploration of these. The bounding process of a breakdown is abductive in the sense that the researcher has a theoretical stance, is open to empirical impressions and engages in a process of activating stories and theories. The empirical impressions are based on rich knowledge about the observed (Van Maanen, 1988). The fieldwork is a social process of intellectual understanding of a culture and a never-ending story of sense-making of new insights into dramas for exploration. Traditional triangulation of cases, interview accounts and methods might undermine a study’s explorative purpose (Alvesson and Ka¨rreman, 2000; Dubois and Gadde, 2002). The mystery methodology can facilitate exploring a complementary meaning of action. Exploration is to learn about situatedness. For some research agendas, “singular views” may be more insightful than “common themes” (Llewellyn and Northcott, 2007). Then, empirical material is used for inspirational purposes in line with what Alvesson (2003) denotes a reflexive pragmatist approach to empirical material. To be reflexive,

Studying supply chain practice

851

IJPDLM 42,8/9

852

in Alvesson’s terms, is to consciously see the subject from different angles, and to be pragmatic is to use the material for some purpose while being aware of its limitations. Working with mystery construction means that data collection is a search in different sources in an almost journalistic manner; the empirical material is constantly framed and constructed. Analysis and theorizing are intertwined in an abductive process of constructing and solving the mystery. Consequently, the resulting empirical material is thick by inquiry (Cook and Brown, 1999). The thickness of the description is determined in the research process in terms of direction and number of interviews based on what is known and what seems to be under-investigated. Weick (1989) argues that the quality of the theory becomes a function of the knowledge breakdown, the assumptions made and the abductive process of how the assumptions are examined. The few abductive logistics and SCM studies undertaken tend to be described as if they were deductive, which reduces the possibility to evaluate them in a fair way (Na¨slund, 2002; Spens and Kova´cs, 2006). Dubois and Gibbert (2010) argue that abductive case studies, used as illustration and inspiration to develop a theoretical idea, should not be judged based on quality criteria of deductive case studies. They are fundamentally different. Instead, an abductive study needs transparency. Transparency would reduce complexity of the case and walk the reader through the argument in order to make the logic, reasoning and causalities clear (Dubois and Gibbert, 2010). If consistency exists in the interplay between theory, case method and empirical phenomena, then the transparency shows the rigour (Flynn, 2008; Spens and Kova´cs, 2006). The interest in the mystery is a substitute for validation during the construction of theories (Weick, 1989), in which practitioners, peers and colleagues play an important role in the investigation for rigour and relevance (Flynn, 2008; Mentzer, 2008). The reflexive pragmatist approach is purposeful to reduce the gap between research and practice by such validation. At this point, I have argued that mystery construction is a creative research process to make sense of supply chain integration, which in my case includes a practice perspective. Practice involves dynamics of supply chain integration, duality in and of actions and complexity of outcomes, and therefore the alternative explorative approach is feasible. Surprises are described as starting points in mystery construction. They are based on a confrontation of what would be expected based on theoretical assumptions and what is. They are a starting point in the quest to close potential gaps between practice and theory. The next section is an elaboration of what I found surprising from field research, based on the debate in SCM literature of the doing and being of integration (Bagchi et al., 2005; Fabbe-Costes and Jahre, 2007; Fawcett and Magnan, 2002). Illustration of surprise from fieldwork The surprise and consequential knowledge breakdown of the concept of supply chain integration is based on a long-term case study of a car manufacturer. The empirical material is foremost interviews but also secondary material, such as presentations and newspaper articles and other studies of the car manufacturer, i.e. empirical material collected by others. In addition, I engaged in a cooperative project with the car manufacturer and two automotive industry interest organizations. My objective was increased understanding of integration and for that reason I joined their process of implementing an information system (supply chain monitoring). The collaborative work gives an additional type of empirical material: to sensitize involved feelings related to the project’s ups and downs.

The intent of the project was implementing an information system that would integrate the supply network in an information structure and facilitate information sharing of supply and demand in the order fulfilment process. The car manufacturer’s objective was stated as to facilitate the suppliers’ efficiency and effectiveness, and their own development of their built-to-order strategy. Originally, the built-to-order strategy was implemented in a radical change process in the 1990s. The importance of both the distribution network and the supply network, especially the first tiers of suppliers, then increased. Procedures and norms of planning, call-off and deliveries were agreed upon and these largely remain operative. The implementation was postponed, which was somewhat unexpected. The project had been developed by suppliers and car manufacturers to tackle a common problem that was significant. However, the suppliers were already taking part in many different projects with the car manufacturer and found it difficult to secure resources for another. The car manufacturer could hardly finance the whole project, and resources therefore became problematic. Investment costs were difficult to get approved in the car manufacturer organization and this time the investment proposal was turned down. However, the project continued because the needs remained, as did the belief in the idea. The implementation was used for another problem and was still seen as prioritized for its original purpose. The unexpected development of the project might be seen as an indicator of the restricted possibilities to get momentum for integrating the business network. And, more integration facilitates the build-to-order strategy, which is the formal strategy. This project scenario indicates that integrative practice is not necessarily stepwise implemented. Being in touch with involved actors led to further questions. What has formed the integration is related to how actors in different positions in different organizations in the supply chain explain the rationale for and the development of the strategy. Perceived or anticipated consequences of strategy development influence actions that develop into interactions and often negotiations. The interactions and negotiations occur within and between the various organizations involved. In the case of the stopped project, interactions between the logistics function and the purchasing function were more difficult than the interactions and negotiations with the suppliers. Functional objectives had evolved differently and became a barrier to the integration. Change affects different parts of the supply chain differently and these adapt and try to take advantage of the change. This might be a reason why measurement of integrative practices gives such ambiguous results. The intertwinement of content and process seems to play an important role in the dynamics of development of integration. Such interconnectedness is not much discussed in SCM literature. Although these insights are not surprising to the practitioner, the empirical material is a surprise in terms of a knowledge breakdown in Alvesson and Ka¨rreman’s (2007a) terminology. It is the initial decision in the research process (Figure 1). In my research process, the surprises are a result of reflections of being close to the field and confronting this material with additional questions. The knowledge breakdown relates to a normative conception of supply chain integration that is difficult to find in empirical studies. The surprises challenge assumptions from SCM literature and conceptualizations of integration. They become knowledge breakdowns from practice, their meaning is unclear and so are the implications. A surprise involves a constant examination of the problem in action. For that reason flexibility in case boundaries and in the interpretative repertoire of alternative theories is needed. What method should be applied is based on

Studying supply chain practice

853

IJPDLM 42,8/9

854

the refined questions. The knowledge breakdown constitutes a starting point for the research process, which could be seen as a narrow origin. It is a real problem that will be followed in thought trials of further theoretical and empirical studies for better understanding of integration in practice. Qualitative SCM research is faced with a dynamic and complex reality, in which the refined conceptualization is important. Explorative research needs to make use of a flexible design because a predetermined case boundary might exclude interesting insights by design. Flexibility in method, theory and case boundary facilitates the constant examination of the problem. On mystery formulation Surprises from the empirical material are meaningful to relate to theories and concepts in order to spur a further theoretical and empirical study, which is a decision in the research process (Figure 1). The choice is based on the practice-based problem and my interpretative repertoire. Strategy-as-practice and logistics literature seems to give an intelligible picture of the development in the illustration. Methodologically, Alvesson and Ka¨rreman (2007a) propose that at this point it is time to formulate a mystery. My mystery concerns the dynamic and complex intertwinement of supply chain integration content and process. Dynamics and complexity evolve as the issue of the knowledge breakdown is followed in its development. The integration process and the content of integration interact. The dynamics and complexity are more accurately appreciated by following the development in its situated context. To follow the action means that context is evolving. Table II is a start of a thought trial where contextual levels and content/process are seen as inspiration and tentative elements of importance. In order to learn about the development, I have to make sense of the interrelatedness in development and refine the questions along with acquired learning. This is to base knowing on knowledge in a productive inquiry in order to get a resolution (Cook and Brown, 1999).

Practice of integration Hierarchical levels of context are intertwined (flexible boundaries)

Intertwinement of Process

Institutionalized processes: supply chain monitoring Organizational processes: supply chain monitoring not as planned but emergent problems of coordination Actors’ content activities: Actors’ process activities: daily planning based on changes in planning and flexibility and change in changes in selling cars (change of priorities) that orders that involve the focal firm (many actors) involve the focal firm’s (many actors) and the and the industrial industrial network’s network many simultaneous processes

Institutional field Institutionalized practices strategies: SCM prescriptions Organizational Organizational strategies: actions partnerships

Activities/ practices

Table II. An example of a mystery construction (tentative)

Content

Undoubtedly, the research process of mystery methodology seems intuitive to explorative researchers and has certainly been used before the metaphorical label of mystery construction was applied. The vocabulary of breakdown and mystery is proposed to work as generative metaphors in a constructive/interpretative methodology (Alvesson and Ka¨rreman, 2008) rather than as an anomaly in the replication of a theory in a realist-objective view (McKinley, 2008). The virtue of the methodology metaphor is partly that it enforces the inquiry and inspires the researcher and partly that it has potential to legitimize research processes following the development where, for example, the decision tree could be used for clarity and transparency of the research process for publication purposes. Making methodological situatedness more explicit will highlight the extent to which methods decisions involve choices and consequential consistency, which is important for reviewers (Hatch and Yanow, 2008). The dynamic and complex intertwinement is a flow (Table II) for making sense of its happening. The categorization in a table as in Table II is only for analytical reasons. Practice of integration relates to learning across the rows and columns. Sense-making of the dynamics and complexity involved is meaningful to practitioners’ reflection of development as well as academia’s theorizing. Theory construction is about making sense, i.e. making assumptions more explicit, making representations more accurate and detailed (Weick, 1989). A mystery might, for example, be described as formulated thought trials (Weick, 1989). The mystery needs systematic and abductive work to be refined into theory (Alvesson and Ka¨rreman, 2007a). The assumptions that my surprises build upon had to be challenged, and consequently I formulated a few indicators that I investigated further empirically and theoretically. These related to literature of SCM/logistics, industrial networks and strategy that concerns dynamics, complexity and practice. Alvesson and Ka¨rreman (2007a) discuss what to do with the mystery, i.e. what kinds of outcomes are possible. Logically, either the mystery is solved and the mystery and a solution are contributions in terms of intelligible understanding and psychological plausibility, or the mystery is not solved and then the (unsolved) mystery becomes the contribution in terms of, for example, analysed ambiguities and complexities. Weick (1989) argues that theory construction is stated by more explicit assumptions, more accurate representations and more detailed representations in our thought trials. Summarizing this far, mystery construction is the theoretical contribution of well-grounded thought trials. The flexibility and the learning attained by adding new questions to the empirical material and to the literature are a part of the theorizing that is facilitated by the mystery metaphor. In my study, the empirical material is a result of choices during the research process of what is needed to create intelligible meaning. The mystery is built of various constructed plots and their twists and interactions. The plots provide illumination, insight into and understanding of the content/process intertwinement. Could traditional ethnography provide the same results? Ethnography that takes off from a bounded case, for example, the car manufacturer, deliberately decides what the case is about. In the mystery methodology the case emerges with the research process, and in my case that means that the industrial network will act in different ways over time. The dynamic involvement of suppliers and customers in development becomes elaborated. Traditional ethnography tends to call attention to the role of organizational actors rather than other actors (Jarzabkowski and Spee, 2009). The mystery methodology opens up

Studying supply chain practice

855

IJPDLM 42,8/9

856

to include interorganizational as well as organizational influences on integration in terms of its content and process; it includes who and what influences integration and it relates development to dynamics and complexity. Traditional ethnography is delimited to explain integration by variance theorizing or by process theorizing. Variance theorizing compares and searches systematically for differences and similarities in several instances of a case in order to understand; often the frame of reference and the case are predetermined. Thus, over time, the case delimits the problem rather than the problem delimiting the case. In process theorizing, outcomes and causes are seen as reciprocally dependent in the evolution of a process. A process is outlined as a case. Process theorizing involves dynamics and complexity. Rigour in variance theorizing depends on its generalizability and precision, while in process theorizing it is related to insights into the phenomena (Mentzer, 2008). Practice theorizing involves dynamics and complexity. A practice-based problem is not bound to researchers’ categories of what a case is, such as a process or a firm. Neither variance theorizing nor process theorizing is practicable to follow the problem. Some of my results relate to the tensions that occur to interorganizational actors, to interdependencies and to interactions. By following the problem in the mystery methodology I select empirical parts that are interesting instead of harvesting the whole field; this is in line with other practice-based approaches, such as the actor-network approach (Hagberg, 2008; a study of organizing and of an idea’s development over time). The focus of the purposeful study is on following the problem; such practice theorizing is related to access, the suitability of empirical sources and ethics (Johnson et al., 2007). The constructivist/interpretivist researcher might use the mystery methodology for practice-based studies. The narrow research initiation and emergent research process relate to process theorizing. As to Langley’s (1999) scheme of different types of process studies it would be seen as a bracketing approach that is a way of analysing differences in a process over time. The rigour of supply chain practice theorizing is in better understanding of the problem (Flynn, 2008; Mentzer, 2008). Thus, theorizing relates to explanations based on gained understanding. The explanations, i.e. the constructed mystery or the constructed mystery solution, are pragmatic in the sense that scientific propositions can never keep up with social change (Lindblom, 1987). Timely issues can be addressed in different ways in order to create better understanding of the phenomenon (Mentzer, 2008). Scientific propositions are theoretical disclosures whose explanations are founded in their conceptual appeal, their logical structure or their psychological plausibility (Alvesson and Ka¨rreman, 2007a). Rigour and trustworthiness is gained through transparency of the research process that needs to illustrate consistency of choices made (Dubois and Gibbert, 2010; Spens and Kova´cs, 2006). Choices involve decisions of theories and methods that influence rigour and relevance (Flynn, 2008; Mentzer, 2008), and therefore explicitness as in Figure 1 is needed in order to evaluate coherence. The mystery is a concern particularly to participating practitioners. Their interest in new insights into what they are doing is a reason for their participation in research. Therefore, it is important to construct a mystery that is crucial to practitioners as well as to researchers, and to close the gap between what is practised, what business schools teach and what we research (Whittington, 2006). Working with mysteries with constant examination in the research process of, for example, supply chain integration is one way to close the

gap between SCM practice and research. Whether the mystery is solved is a question of perspective. Regardless of whether a “solution” is provided, the research has provided a new way to reflect on and analyse the problem (Carroll et al., 2008). Concluding discussion Mystery methodology is based on sensitive construction (Alvesson and Ka¨rreman, 2007a). It draws on an ontology in which the ambiguous and constructed nature of empirical material allows scope for a freer and bolder way of interacting with empirical material (Alvesson and Ka¨rreman, 2000). The mystery construction process facilitates SCM research in three ways: (1) as a consistent methodology for practice research; (2) for learning and responsiveness to new insights; and (3) with the problem of bounding the case. Consistent methodology If an SCM phenomenon is viewed as holistic, dynamic and context-sensitive, it is possible to learn about implications of these assumptions if the research process and its methodological assumptions are consistent with that view. The proposed methodology facilitates rigour and relevance in terms of thorough and constant examination of a supply chain issue (Flynn, 2008; Mentzer, 2008), especially since thought trials and decisions are delineated and made transparent regarding how and in what context conclusions are reached (Mentzer, 2008). Mystery methodology bridges the relevance gap based on learning and openness to the study object, which is of importance to a perspective that seeks new insights. The illustration of supply chain integration as a phenomenon shows practice in which integration neither follows a stepwise pattern nor is feasible to measure. Instead, the practice study enables insights into intertwinement of micro and macro levels as well as content and process issues, which diminishes the gap between research and practice by sense-making of encountered messiness. The mystery methodology is a structure in which the mystery is studied in an emergent research process that involves meaning, practice and talk of the problem (Alvesson and Ka¨rreman, 2000). Surprises are the origin of mystery construction. In theorizing, the surprise gets a meaning by use of concepts. Mystery construction involves theoretical disclosures whose explanations are founded in their conceptual appeal, their logical structure or their psychological plausibility (Alvesson and Ka¨rreman, 2007a). The mysteries are temporary in the sense that they are social phenomena that shift in character over time as new dynamics and complexity are developed. In this way mystery construction is finished more for pragmatic reasons than for the reason that “we know everything about this now”. Mystery construction relates to insights into a problem – and authenticity and plausibility of the mystery are seen to be more important than the ability to generalize regarding the constituting activities. The methodological consistency might further develop contemporary supply chain research by rigour and relevance. Learning Mystery methodology facilitates SCM research by its potential for learning and openness to new insights. The mystery metaphor is suitable for explorative researchers who are

Studying supply chain practice

857

IJPDLM 42,8/9

858

knowledgeable about the literature in a variety of areas in order to refine questions during the research process. Initially in the research process surprises from fieldwork are anecdotal indicators for further theoretical and empirical framing. Flynn (2008) proposes that rigorous and relevant research is achieved by constantly thinking about which of our theoretical weapons would provide the best way to attack the research question. In line with that proposition the construction of a mystery relates to several meaningful plots; these are built up of insights from empirical and theoretical investigation. A plot might describe integration in terms of how one outcome relates to another, because micro outcomes are interwoven with macro outcomes in practice (Chia and MacKay, 2007). There are multiple goals and a duality in goals that the mystery construction illuminates. Other plots are meaningful and we need to stay open-minded to the research question and the literature base (Flynn, 2008). In the illustration of supply chain integration in the car industry, the constructed contemporary social phenomenon takes off from a thin insight or surprise that is a knowledge breakdown; it opens up for and includes activities, actions, practices and outcomes, and contributes to the knowledge of supply chain integration. This explorative process is of an interpretative/constructional character. The empirical material is a vehicle to develop sensible theory constructions for academia and for practitioners. The mystery methodology does not exclude other methodologies that are in use. Case boundaries The mystery methodology alleviates the problem of bounding the case. Among other things, it opens up for interorganizational research. My empirical matter relates to interorganizational as well as organizational influences on supply chain integration. The mystery methodology facilitates interorganizational research by widening the boundary that determines who and what influences development and by relating the development to dynamics and complexity. Case boundaries are flexible, which facilitates the study of logistics and SCM phenomena that change over time (Na¨slund, 2002). Flexibility also relates to the theoretical framework. The methodological thought trials involve different perspectives and ways to make sense of the material, which is a way to do rigorous and relevant research through pragmatic, field-based grounding (Carter, 2008b; Flynn, 2008; Mentzer, 2008). Future research and limitations I have suggested that the explorative mystery methodology facilitates SCM practice research because it enables following integration in its making. As an example of a social practice approach the methodology is a contribution to logistics and SCM research (Gammelgaard, 2004); in particular the methodology supports abductive approaches. As such, the mystery methodology is appropriate to use as a critical perspective, which is basically non-existing in SCM and logistics research (Craighead et al., 2007; Sachan and Datta, 2005). I studied the phenomenon of integration and decided to take a practice perspective rather than the prevailing manufacturer perspective (Halldo´rsson and Arlbjørn, 2005). More approaches that are consistent with constructivist/interpretivist assumptions would be fruitful for development of the field. The theories I used explain integration in a complementary way in relation to each other and in a new way in relation to supply chain literature. The mystery construction facilitated the choices and increased transparency and structure in the decision process in the research approach, which is

an important quality criterion of abductive research (Dubois and Gibbert, 2010; Spens and Kova´cs, 2006). The close empirical engagement of occurring phenomena that are scrutinized from different theoretical perspectives not only facilitates research; it also offers reflection and potential freedom from old “truths” together with hints about other possibilities (Czarniawska, 2004; Law, 2004). Action research and ethnography are sense-making methods in use by logistics and SCM researchers (Nilsson, 2005; Na¨slund, 2002), but more methods are possible as long as these are used in consistence with constructivist/interpretivist assumptions (Hatch and Yanow, 2008). The rigour and relevance of mystery-focused research will not lead to exact measurement or to population-based generalization but to richer insights that will develop the field of SCM (Carter, 2008b; Flynn, 2008; Mentzer, 2008), which are limitations of the methodology as well as a potential for future research. References Alvesson, M. (2003), “Beyond neopositivists, romantics, and localists: a reflexive approach to interviews in organisational research”, Academy of Management Review, Vol. 28 No. 1, pp. 13-33. Alvesson, M. and Ka¨rreman, D. (2000), “Taking the linguistic turn in organizational research: challenges, responses, consequences”, Journal of Applied Behavioral Science, Vol. 36 No. 2, pp. 136-58. Alvesson, M. and Ka¨rreman, D. (2007a), “Constructing mystery: empirical matters in theory development”, Academy of Management Review, Vol. 32 No. 4, pp. 1265-81. Alvesson, M. and Ka¨rreman, D. (2007b), “Unraveling HRM: identity, ceremony, and control in a management consulting firm”, Organization Science, Vol. 18 No. 4, pp. 711-23. Alvesson, M. and Ka¨rreman, D. (2008), “On the social nature of explicating mystery construction in theory development: a response to McKinley”, Academy of Management Review, Vol. 33 No. 2, pp. 543-5. Alvesson, M. and Sko¨ldberg, K. (2000), Reflexive Methodology, Sage, London. Araujo, L., Kjellberg, H. and Spencer, R. (2008), “Market practices and forms: introduction to the special issue”, Marketing Theory, Vol. 8 No. 1, pp. 5-14. Asplund, J. (1970), Om undran info¨r samha¨llet (Reflections on Society), Argos, Lund. Bagchi, P.K., Ha, B.C., Skjøtt-Larsen, T. and Soerensen, L.B. (2005), “Supply chain integration: a European survey”, The International Journal of Logistics Management, Vol. 16 No. 2, pp. 275-94. Carroll, B., Levy, L. and Richmond, D. (2008), “Leadership as practice: challenging the competency paradigm”, Leadership, Vol. 4 No. 4, pp. 363-79. Carter, C.R. (2008a), “Introduction to the forum on the gap between research and practice in supply chain management”, Journal of Supply Chain Management, Vol. 44 No. 2, pp. 56-6. Carter, C.R. (2008b), “Knowledge production and knowledge transfer: closing the research-practice gap”, Journal of Supply Chain Management, Vol. 44 No. 2, pp. 78-82. Carter, C., Clegg, S.R. and Kornberger, M. (2008a), “S-A-P zapping the field”, Strategic Organization, Vol. 6 No. 1, pp. 107-12. Carter, C., Clegg, S.R. and Kornberger, M. (2008b), “So!apbox: editorial essays: strategy as practice?”, Strategic Organization, Vol. 6 No. 1, pp. 83-99.

Studying supply chain practice

859

IJPDLM 42,8/9

860

Chia, R. and MacKay, B. (2007), “Post-processual challenges for the emerging strategy-as-practice perspective: discovering strategy in the logic of practice”, Human Relations, Vol. 60 No. 1, pp. 217-42. Cook, S.D.N. and Brown, J.S. (1999), “Bridging epistemologies: the generative dance between organizational knowledge and organisational knowing”, Organization Science, Vol. 10 No. 4, pp. 381-400. Craighead, C.W., Hanna, J.B., Gibson, B.J. and Meredith, J.R. (2007), “Research approaches in logistics: trends and alternative future directions”, International Journal of Logistics Management, Vol. 18 No. 1, pp. 22-40. Czarniawska, B. (2004), “On time, space and action nets”, Organization, Vol. 11 No. 6, pp. 773-91. Dubois, A. and Gadde, L.-E. (2002), “Systematic combining: an abductive approach to case research”, Journal of Business Research, Vol. 55 No. 7, pp. 553-60. Dubois, A. and Gibbert, M. (2010), “From complexity to transparency: managing the interplay between theory, method and empirical phenomena in IMM case studies”, Industrial Marketing Management, Vol. 39 No. 1, pp. 129-36. Fabbe-Costes, N. and Jahre, M. (2007), “Supply chain integration improves performance: the emperor’s new suit?”, International Journal of Physical Distribution & Logistics Management, Vol. 37 No. 10, pp. 835-55. Fawcett, S.E. and Magnan, G.M. (2002), “The rhetoric and reality of supply chain integration”, International Journal of Physical Distribution & Logistics Management, Vol. 32 No. 5, pp. 339-61. Flynn, B.B. (2008), “Having it all: rigor versus relevance in supply chain management research”, Journal of Supply Chain Management, Vol. 44 No. 2, pp. 63-7. Gammelgaard, B. (2004), “Schools in logistics research? A methodological framework for analysis of the discipline”, International Journal of Physical Distribution & Logistics Management, Vol. 34, pp. 479-91. Gherardi, S. (2009), “Introduction: the critical power of the ‘practice lens’”, Management Learning, Vol. 40 No. 2, pp. 115-28. Giddens, A. (1979), Central Problems in Social Theory: Action, Structure, and Contradiction in Social Analysis, University of California Press, Berkeley, CA. Hagberg, J. (2008), “Flytande identitet – NetOnNet och e-handelns a˚terkomst”, doctoral dissertation, Ho¨gskolan i Bora˚s, Bora˚s. ´ . and Arlbjørn, J.S. (2005), “Research methodologies in supply chain management: Halldo´rsson, A what do we know?”, in Kotzab, H., Seuring, S., Mu¨ller, M. and Reiner, G. (Eds), Research Methodologies in Supply Chain Management, Physica-Verlag HD, Heidelberg, pp. 107-22. Hatch, M.J. and Yanow, D. (2008), “Methodology by metaphor: ways of seeing in painting and research”, Organization Studies, Vol. 29 No. 01, pp. 23-44. Jarzabkowski, P. and Spee, A.P. (2009), “Strategy-as-practice: a review and future directions for the field”, International Journal of Management Reviews, Vol. 11 No. 1, pp. 69-95. Johnson, G., Langley, A., Melin, L. and Whittington, R. (2007), Strategy as Practice: Research Directions and Resources, Cambridge University Press, Cambridge. Ka¨rreman, D. and Alvesson, M. (2009), “Resisting resistance: counter-resistance, consent and compliance in a consultancy firm”, Human Relations, Vol. 62 No. 8, pp. 1115-44. Ka¨rreman, D. and Rylander, A. (2008), “Managing meaning through branding; the case of a consulting firm”, Organization Studies, Vol. 29 No. 01, pp. 103-25.

Langley, A. (1999), “Strategies for theorizing from process data”, Academy of Management Review, Vol. 24 No. 4, pp. 691-710. Latour, B. (1987), Science in Action: How to Follow Scientists and Engineers Through Society, Harvard University Press, Cambridge, MA. Latour, B. (2005), Reassembling the Social: An Introduction to Actor-Network-Theory, Oxford University Press, New York, NY. Law, J. (2004), After Method: Mess in Social Science Research, Routledge, London. Law, J. (2008), “On sociology and STS”, The Sociological Review, Vol. 56 No. 4, pp. 623-49. Lindblom, C.E. (1987), “Alternatives to validity: some thoughts suggested by Campbell’s guidelines”, Science Communication, Vol. 8 No. 3, pp. 509-20. Llewellyn, S. and Northcott, D. (2007), “The ‘singular view’ in management case studies”, Qualitative Research in Organizations and Management: An International Journal, Vol. 2 No. 3, pp. 194-207. McKinley, W. (2008), “The mystery about mysteries: a commentary on Alvesson and Ka¨rreman”, Academy of Management Review, Vol. 33 No. 2, pp. 541-5. Mentzer, J.T. (2008), “Rigor versus relevance: why would we choose only one?”, Journal of Supply Chain Management, Vol. 44 No. 2, pp. 72-7. Na¨slund, D. (2002), “Logistics needs qualitative research – especially action research”, International Journal of Physical Distribution & Logistics Management, Vol. 32 No. 5, pp. 321-38. Nilsson, F. (2005), “Adaptive logistics: using complexity theory to facilitate increased effectiveness in logistics”, doctoral dissertation, Lund University, Lund. Nilsson, F. (2006), “Logistics management in practice: towards theories of complex logistics”, International Journal of Logistics Management, Vol. 17 No. 1, pp. 38-54. Pratt, M.G. (2008), “Fitting oval pegs into round holes”, Organizational Research Methods, Vol. 11 No. 3, pp. 481-509. Sachan, A. and Datta, S. (2005), “Review of supply chain management and logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 35 No. 9, pp. 664-705. Sandberg, E. (2007), “The role of top management in supply chain management practices”, doctoral dissertation, Linko¨ping University, Linko¨ping. Solem, O. (2003), “Epistemology and logistics: a critical overview”, Systemic Practice and Action Research, Vol. 16 No. 6, pp. 437-54. Spens, K.M. and Kova´cs, G. (2006), “A content analysis of research approaches in logistics research”, International Journal of Physical Distribution & Logistics Management, Vol. 36 No. 5, pp. 374-90. Spicer, A., Alvesson, M. and Ka¨rreman, D. (2009), “Critical performativity: the unfinished business of critical management studies”, Human Relations, Vol. 62 No. 4, pp. 537-60. Storey, J., Emberson, C., Godsell, J. and Harrison, A. (2006), “Supply chain management: theory, practice and future challenges”, International Journal of Operations & Production Management, Vol. 26 No. 7, pp. 754-74. Sutton, R.I. and Staw, B.M. (1995), “What theory is not”, Administrative Science Quaterly, Vol. 40 No. 3, pp. 371-84. Tan, K.C. (2002), “Supply chain management: practices, concerns, and performance issues”, Journal of Supply Chain Management, Vol. 38 No. 1, pp. 42-53.

Studying supply chain practice

861

IJPDLM 42,8/9

862

Van Maanen, J. (1988), Tales of the Field: On Writing Ethnography, The University of Chicago Press, Chicago, IL. Weick, K.E. (1989), “Theory construction as disciplined imagination”, Academy of Management Review, Vol. 14 No. 4, pp. 516-31. Weick, K.E. (1995), “What theory is not, theorizing is”, Administrative Science Quaterly, Vol. 40 No. 3, pp. 385-90. Whittington, R. (2006), “Completing the practice turn in strategy research”, Organization Studies, Vol. 27 No. 5, pp. 613-34. About the author Benedikte Borgstro¨m is a post doc of Chalmers University of Technology and works part time at Jo¨nko¨ping as Co-director of the Centre of Logistics and Supply Chain Management. Her area of expertise is in supply chain management and logistics management, supply chain integration, and business-to-business marketing, especially in strategizing, and the dynamics and complexity of business relationships. Her current research interests are focused on strategies involving suppliers and buyers in coordinating their activities, the organizing of supply chains, methodology and problem-oriented research, such as of logistics implications of global outsourcing or of postponement in retail chains. Benedikte Borgstro¨m can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0960-0035.htm

Grounded theory: an inductive method for supply chain research

GT: an inductive method for research

Wesley S. Randall Department of Marketing and Logistics, College of Business, University of North Texas, Denton, Texas, USA, and

John E. Mello Department of Management and Marketing, Arkansas State University, Jonesboro, Arkansas, USA

863 Received 1 December 2011 Accepted 16 December 2011

Abstract Purpose – Development of theory remains an essential step in the evolution of supply chain management as an integrative business discipline. Supply chain research often involves phenomena possessing complex behavioral dimensions at both the individual and organizational levels. Such complexity can require the utilization of holistic and inductive approaches in order to more fully understand the behaviors associated with the phenomena. This paper aims to provide a step-by-step guide intended to increase researchers’ understanding of the use of grounded theory (GT) methodology in supply chain contexts. Design/methodology/approach – The paper argues for GT as an appropriate method for studying emerging supply chain phenomena using an inductive, holistic approach. Findings – GT is positioned in a holistic framework of research methodologies. Next a step-by-step explanation of the grounded theory process is offered, illustrated by examples from the authors’ own research. Originality/value – This paper links the complex “system of systems” nature characteristic of supply chains to the need for a holistic research approach such as grounded theory. It also provides a guide for researchers, reviewers, and editors to judge sound GT. Moreover, from a practical perspective, the in-vivo nature of GT provides recognizable solutions to managerial problems. Keywords Grounded theory, Supply chain research methods, Theory development, Research methods, Supply chain management Paper type Research paper

Introduction A broad theory capable of explaining and predicting supply chain management (SCM) phenomena is the first step in the evolution of SCM as an integrative business discipline (Lambert and Garcı´a-Dastugue, 2006; Mentzer et al., 2001; Mentzer and Kahn, 1995). Yet supply chain researchers struggle to define core supply chain theory (Defee et al., 2010). This is not surprising as SCM represents a unifying discipline touching on all elements of business. Consider the Council of Supply Chain Management Professionals (CSCMP) definition of SCM: Supply chain management encompasses the planning and management of all activities involved in sourcing and procurement, conversion, and all logistics management activities. Importantly, it also includes coordination and collaboration with channel partners, which can be suppliers, intermediaries, third party service providers, and customers. In essence, supply chain management integrates supply and demand management within and across companies (CSCMP, 2011).

International Journal of Physical Distribution & Logistics Management Vol. 42 No. 8/9, 2012 pp. 863-880 q Emerald Group Publishing Limited 0960-0035 DOI 10.1108/09600031211269794

IJPDLM 42,8/9

864

This definition implies that a wide range of researchable phenomena, including strategies, processes, and functions fall within the domain of SCM. Supply chain research often involves phenomena possessing complex behavioral dimensions at both the individual and organizational levels. Therefore, research into SCM requires a method that understands inter- and intra-organizational systems as a whole. The holistic, organizational, and inter-organizational focus of SCM extends the systems’ theory of the firm (Drucker, 1962) toward a “systems of systems” (SoS) theory of the SCM network (Gorod et al., 2008; Randall et al., 2010). Classically, the systems approach linked internal firm logistics functions (e.g. inbound and outbound freight) to achieve lowest total cost (Poist, 1974, 1989). Modern SCM involves multiple interrelated firm and inter-firm processes of globally complex “SoS” (DiMario et al., 2008; Lambert et al., 2005; Muller-Merbach, 1994; Sauser, 2006; Sousa-Poza and Kovacic, 2008). Modern supply chain complexity requires a research methodology that informs research as to how individuals interact within the whole. Grounded theory (GT) provides such a methodology (Glaser and Strauss, 1967), allowing for deep insights into social phenomena (Mello and Flint, 2009). GT provides researchers an understanding the human side of SCM – the underlying meanings in human experiences, interactions, and relationships that form company strategies and actions in a supply chain context. The GT approach integrates the system and the individual by providing a process that recognizes, categorizes, and relates key individual and organizational variables in a theoretical framework. The ability of GT to address behavioral dimensions at the individual, organizational, and inter-organizational level is important. Supply chains are inherently social by nature; moreover, they are intrinsically interrelated, inter-organizational, inter-firm, and cross-cultural (Isenberg, 2008; Mello and Stank, 2005). This implies a broad behavioral backdrop that influences supply chain decisions and behaviors. Because of this, supply chain researchers must deal with complex and interrelated phenomena such as applications of power/dependence between firms, supplier negotiations, impacts of outsourcing on employees, customer service perceptions, and supply chain innovation as a social process (Mello and Flint, 2009). For the researcher, understanding how such phenomena impacts the system outcomes involves “interpreting” how supply chain managers perceive problems, react to changing business environments, and influence their business environments. Unfortunately, little has been done in the SCM literature to succinctly explain GT methods or relate it to the study of complex system behavior. Few SCM empirical articles provide detailed explanations of the methodology and logic behind GT, and none lay out a standardized approach for applying and assessing GT. At the same time, GT researchers have a difficult time justifying their method to editors and reviewers who often lack the expertise to evaluate GT research. Oftentimes, GT researchers find themselves having to educate reviewers about GT methods and procedures. Part of this issue stems from the fact that there are foundational differences in GT research methodologies as well as different assessment criteria for GT depending on the model the researcher chooses to follow (e.g. Glaser, Strauss and Corbin, or Charmaz). To address this gap, this paper provides SCM researchers with a guide for employing and assessing GT as a SCM research method. First, the paper offers an overview of GT success in supply chain research. Second, the paper provides a rational for GT as an appropriate method for studying complex SoS supply chain phenomena.

Third, the paper defines GT and outlines the GT process. Fourth, the paper provides a framework for authors, reviewers and editors to evaluate GT research. Fifth, GT concepts are explained and illustrated using examples from current research. Finally, a method for assessing the rigor and trustworthiness of GT research is provided and some of the limitations of GT are explained.

GT: an inductive method for research

GT success in SCM research GT has been used to develop foundational theory across a broad swath of social science disciplines (Bowen, 2008; Charmaz, 2006; Glaser and Strauss, 1967; Goulding, 1998). GT provides a method for emerging and relating key variables by uncovering the environmental and organizational structures and processes upon which organizations act (Bagozzi, 1984; Hunt, 1992; Locke, 2007). The success of GT has led to an increase in the number of SCM studies utilizing GT or GT analysis methods to explain SCM phenomena. For example, Pfohl and Buse (2000) used GT analysis to investigate inter-firm networks, while Pappu and Mundy (2002) used GT methods to explore buyer/seller relationships in transportation from an organizational learning perspective. Haytko et al. (2007) combined phenomenological interviews with GT analysis methods to investigate issues facing the maquiladora industry in Mexico and the cross-border supply chain. Even more recent application of GT includes Manuj and Mentzer (2008), who used GT to explore the phenomenon of risk management and risk management strategies in global supply chains. Also, Thomas (2008) explored the phenomenon of time pressure in supply chain relationships using GT. Randall, Pohlen, and Hanna used GT to create a performance based logistics (PBL) theory grounded in service dominant logic (SDL). Lastly, Russell et al. (2009) applied GT analysis methods in their investigation of the integration of ethanol in the petroleum supply chain. GT is successful because it allows researchers to access to social and structural aspects of SCM that other methodologies may have difficulty uncovering. To explain how this occurs, the next section provides a detailed explanation of GT and why it works.

865

Rationale for GT in supply chain research Science and logic suggest that understanding the whole requires rigorous questioning of its parts via the inductive method (Holland, 1992; Locke, 2007). GT is an appropriate and vetted method of inductive research. In SCM, rapid structural changes drive large-scale shifts. Such shifts include outsourcing of transportation, the increasing use of 3/4PLs and wholesale adoption of enterprise resource planning systems, along with internet-based, common-exchange formats and low-cost bandwidth (Koh et al., 2006; Lieb, 2008; Maltz and Ellram, 1997). SCM is based in the systems theory of the firm (Drucker, 1962). Modern supply chains, global supply chains have increased in complexity and can only be understood from a SoS perspective (Gorod et al., 2008). GT provides a research methodology to understand phenomena against the backdrop of the larger system (Sousa-Poza and Kovacic, 2008). This holistic SoS view sees SCM as a complex and adaptive discipline involving systems within systems that span multiple firms, economies, and cultures (DiMario et al., 2008; Lambert et al., 2005; Meixell and Gargeya, 2005; Muller-Merbach, 1994; Sauser, 2006; Sousa-Poza and Kovacic, 2008). The macro and interrelated nature of SCM presents methodological issues that enable comprehension of the whole through study of its parts difficult:

IJPDLM 42,8/9

866

All complex adaptive systems involve large number of parts undergoing a kaleidoscopic array of simultaneous nonlinear interactions. Because of the nonlinear interactions, the behavior of the whole system is not even to an approximation a simple sum of the behavior of its parts. The unusual mathematical techniques of linear approximation – linear regression, normal coordinates [. . .] and the like – make little progress in the analysis of complex adaptive systems [. . .] (Holland, 1992, pp. 184-5).

SCM represents a complex and adaptive organizational system. By understanding how organizations adapt, researchers can understand the rules and environmental forces shaping that adaptation (Holland, 1992). As business environments change (e.g. September 11th, volatile fuel prices, economic downturns and upturns), human systems adapt to those changes. GT uses an aggregate, inductive, and pattern-seeking process to relate behavior to higher-level concepts (categories and properties) within a theory building framework (Charmaz, 2006; Glaser and Strauss, 1967). Thus, GT goes beyond description to develop categories with dimensions (e.g. high or low knowledge awareness, high or low adoption of innovation) and then posit relationships between these categories (e.g. knowledge awareness influences the degree to which innovation is adopted). By proposing such relationships, GT builds theory that can be tested using quantitative methods, leading toward broader generalization. In GT, researchers conduct research without a priori constructs, allowing them to let concepts emerge from the data rather than being constrained by previous theory. GT uncovers the behavioral dimensions in SCM research as GT provides researchers with a method to collect and interpret data from personnel at various levels and functions within and across firms. This direct contact gets at core processes underlying SCM in practice. GT works because it uses a holistic and process-oriented method to determine the rules, processes, and strategies upon which supply chains operate (Glaser, 1978; Holland, 1992; Mello and Flint, 2009). The following are examples of research questions that GT may be particularly well suited to answer: RQ1. What are the appropriate managerial mindsets in order for collaborative relationships between supply chain members to be successful? RQ2. What are the issues that must be overcome in order for firms to enact inter-firm information sharing? RQ3. How is corporate culture related to SCM practices? RQ4. How do firms cope with disparities of technical, financial, or human resource capability between themselves and other supply chain members? RQ5. What are the impacts of mergers and acquisitions on the SCM practices of involved companies? RQ6. In what ways does a company’s supply chain orientation affect its boundary-spanning employees? While by no means all-inclusive, this list illustrates the type of phenomena suited to a GT study.

What is GT? GT is an exploratory, observational research technique used to understand phenomena about which there exists little theoretical knowledge (Locke, 2007; Suddaby, 2006). GT creates a holistic and inductive understanding of phenomena based upon the view of the participants (Charmaz, 2006) and an interpretive analysis of data obtained in the field (Goulding, 1998). This methodology differs from quantitative approaches that deductively test theory based upon a priori assumptions. The analytical engine of GT is a process known as “constant comparison” (Glaser, 1978). Constant comparison seeks to identify theoretical similarities and differences between the perceptions of participating individuals. These contrasting perspectives allow the development of higher order concepts that explain behavior (Gephardt, 2004; Glaser, 1978). The key to constant comparison is “theoretical sampling”, which is a method for gaining multiple perspectives of phenomena. Grounded theorists select a sample that is rich in the phenomena under investigation (Charmaz, 2006). This sample provides data from which to reveal concepts (categories and their properties) in support of inductive theory development. This sampling approach contrasts with quantitative methodology where the sampling goal is to support statistical generalization. Grounded theorists validate emerging categories and relationships by comparing findings from one sample to a new sample selected specifically to confirm or disconfirm the previous propositions. Categories are abstract, higher order concepts that represent groups of underlying codes sharing conceptual content. Theoretical sampling may involve: . comparison of concepts articulated by study participants based upon a single interview; . comparison of interviews with multiple participants; . comparison and review of field memos; . comparison with organizational artifacts such as meeting minutes, awards, and appraisals; and . comparison with relevant literature (Charmaz, 2006; Glaser, 1978). Good GT is therefore a “set of well-developed concepts related through statements of relationship, which together constitute an integrated framework that can be used to explain or predict phenomena” (Strauss and Corbin, 1998, p. 22). Using the constant comparison process, the researcher develops a theoretical framework built upon categories, properties and relationships (Cho and Trent, 2006). This framework explains and predicts the majority of the behavior in a phenomenon. These theories can be verified through quantitative hypothesis testing. Placing GT in perspective with other approaches To understand GT it is useful to understand where GT resides in perspective to other research methodologies. Burrell and Morgan (1979) and Meredith et al. (1989) describe academic research paradigms as continuums. These continuums represent the philosophical paradigms of the researcher and discipline (Burrell and Morgan, 1979; Hunt, 1992). They can range from positivist on one extreme to relativist/interpretivist on the other (Figure 1). Positivists stress objectivity, generality, replicability, and casualty through falsification (Hunt, 1990, 2002; Popper, 1959). The positivist perspective emphasizes

GT: an inductive method for research 867

IJPDLM 42,8/9

868

Figure 1. The continuum of research paradigms

Source: This figure is a derivation of a previous work published in Randall et al. (2010)

precise, probabilistic theories aimed at explanation and prediction (Hunt, 1992). This paradigm includes the realist and scientific realists (Hunt, 1992) along with the meta-theory developmentalist (Bartels, 1970). Examples include falsification (Popper, 1959), experimentation (Campbell and Stanley, 1963), mathematical modeling and linear programming (Bender, 1981; Kuehn and Hamburger, 1963), simulation (Shycon and Maffei, 1960), and statistical approaches (Bagozzi and Yi, 1988; Hair et al., 2010). At the other end of the continuum reside the relativist and interpretivist. These researchers, such as Anderson (1983) and Hirschman (1986), emphasize description rather than prediction. Interpretivists rely on qualitative methods. The relativist/interpretive perspectives include ethnographic (Belk et al., 1988; Mcgrath et al., 1993), semiotic (Holbrook and Grayson, 1986; Mick, 1986), case-based (Eisenhardt, 1989), and phenomenological approaches (Thompson et al., 1990). In general, the relativist/interpretivist seeks to understand the environment from the perspective of the individual. These researchers accept that individuals shape their environments, and individual perceptions influence the causal structure of their environment. Between the positivists and the relativist/interpretivist lie the GT objectivists and constructivists. Grounded theorists are more positivistic in their qualitative approach with a focus on explanation and prediction. For Glaser and Strauss (1967), a theory is “an integrated framework that can be used to explain or predict phenomena”. Charmaz (2006, p. 131) terms this approach “objectivist grounded theory”, one that “assumes an external reality awaiting discovery and an unbiased observer who records facts about it”. The GT constructivists reside between objectivists and relativists; they assume that: . participants construct meanings specific to contexts of time, place and culture; and . that researchers themselves may interpret meanings based on their own values (Charmaz, 2006). Constructivist GT researchers raise categories to high-level theoretical concepts. However, for objectivists these concepts become core variables that “serve as interpretive frames and offer an abstract understanding of relationships” (Charmaz, 2006, pp. 139-40).

This movement towards abstract understanding places this approach on the relativist/interpretivist side of the continuum (for a more detailed discussion of the differences, see Mello and Flint (2009, pp. 115-21). Like most qualitative analysis methods, GT resides toward the relativist/interpretivist end of the continuum. Yet GT’s pattern-seeking, theory building focus set it apart from other qualitative methods. GT researchers eschew simple description for its lack of theory development, contrasted with ethnographers, who rely heavily on description to convey the essences of cultures, existential phenomenologists, who produce clear and accurate descriptions of a particular aspect of human experience (Thompson et al., 1990), or semioticits who analyze the structures of meaning-producing events, both verbal and non-verbal, but do not seek to develop theory in the manner of GT (Holbrook and Grayson, 1986; Mick, 1986). GT is close to case studies in that both GT and case study researchers typically combine a variety of data collection methods gathered in a field setting to yield theory (Eisenhardt, 1989). While GT generally relies on qualitative data and data analysis techniques, its focus on perception, repeatability and follow-on hypothesis testing suggests that GT straddles multiple research paradigms.

GT: an inductive method for research 869

The GT process This section provides a step-by-step overview of the GT process. Figure 2 shows the GT process as a two-dimensional flow that moves back and forth along this process to compare concepts articulated by study participants. The iterative, overlapping and dynamic nature of GT is shown using the arrows in Figure 2. The research problem, initial question, and sample Step 1 involves defining the research problem, developing the initial research questions, and identifying the sample. During this step the literature should act like a lens that focuses but does not constrain the research (Charmaz, 2006; Glaser, 1978, 1992). For example, in an investigation into supply chain innovation and culture, one of the authors delved into inter-company relationship literature to gain theoretical sensitivity on inter-firm governance procedures. In another study, one of the authors considered how SDL, a new marketing framework, might provide insight into the success of PBL. Once the investigation began, the researchers set the literature aside.

Source: Adapted from Charmaz (2006)

Figure 2. GT process

IJPDLM 42,8/9

870

This allows concepts and relationships to emerge from the field data unconstrained by the literature (Charmaz, 2006). The second element of step 1 involves having industry personnel and academic experts clarify the research questions and validate the initial sample. During this phase the sample is critiqued for: . appropriateness; and . as a foundation for theory development (Charmaz, 2006). Next the researcher begins to code and uncover variables. These variables may be familiar from a research perspective. The strength in the grounded approach is that these variables are often related in new ways or to new variables (Glaser, 1978). Thus, reviewers and authors may recognize these variables and must caution themselves not to assume that nothing new has been found. This process differs from the use of literature develop a priori constructs to be tested. Open and focused coding At steps 2 and 4, interviews and other forms of data are collected. GT requires dimensionality; this means that phenomena need to be seen from various aspects. The sample in GT requires perceptions from personnel in different organizations at varying management levels. The objective of this sampling procedure is to gain varying perspectives on the phenomena under study. For instance, the idea of access from an information technology perspective in an SCM setting can only be assessed if the researcher interviews multiple SCM partners from multiple locations and multiple tiers of the supply chain. Typically, interviews are recorded and transcribed verbatim. This process ensures that nuances missed by less exacting methods, such as tone and certainty, are maintained and incorporated into the findings. The interview words, sentences and paragraphs are coded using open and focused coding (steps 2 and 4). Open coding (sometimes called initial coding) breaks down data into meaningful units that are later reassembled into higher-level concepts (Strauss, 1987). The researcher analyzes the transcript line-by-line to identify “key words or phrases which connect the informant’s account to the experience under investigation” (Goulding, 2002, p. 76). During open coding, similarities and differences are compared and grouped together into abstract concepts or categories (Glaser, 2001). Focused codes are used to distill and aggregate emerging concepts into categories and properties during step 7. To illustrate, consider a sentence transcribed during one of the author’s research projects (Randall et al., 2010): “Improving performance requires knowledge, and that means we have to have information systems to transfer that knowledge.” The participant considers knowledge to be an element of improved performance, and improved performance and knowledge are related to information systems. That sentence may therefore be coded as “performance”, “knowledge” and “IT system”. In addition, theoretical codes and memos may be taken to remind the researcher to determine if there is a pattern of relationship linking “knowledge”, “information systems” and “performance” in subsequent interviews. Memos Documenting the logic behind codes and emerging categories is central to establishing validity in GT. GT uses a process called memoing (steps 3, 5, and 7) to capture the

thoughts and decisions that lead to category emergence. Memos capture hypothesized relationships, provide a record of how these relationships developed and create the logic for subsequent interviews. Memos provide the logic and process through which the GT emerges (Charmaz, 2006). Memos help researchers:

GT: an inductive method for research

[. . .] a) to grapple with ideas about the data, b) to set an analytic course, c) to refine categories, d) to define the relationships among various categories, and e) to gain a sense of confidence and competence in their ability to analyze data (Charmaz, 1998, pp. 517-18).

871

Memos are also used to capture ideas concerning what directions the researcher should take next (Charmaz, 2006; Glaser, 1978; Strauss and Corbin, 1998) and to trace the development in thought processes regarding categories over the course of the study. Memos serve as an audit trail, providing validation of how researchers arrive at their conclusions. For example, one of the authors developed a figure depicting the evolution of concept development from open coding through final theory based on memos dealing with theoretical sampling and category development (Mello and Hunt, 2009). These memos, written during the data analysis phase, contained thoughts concerning what types of companies and which individuals within those companies could further the theoretical sampling process, potential categories and properties, and the linkages between category/property development and existing theory in the literature. Theoretical sampling In step 6, GT researchers engage in a process called theoretical sampling. Theory sampling involves looking for statements, events or cases that illuminate dimensionality as well as positive and negative instances of categories (Charmaz, 2006). For example, in a study of control dealing with third-party logistics providers, one of the authors sought samples that provided different power/dependence relationships to understand the dimensional ranges of control. This sampling approach illuminates the various aspects of the emerging variables (e.g. low/high dependence) and their theoretical relationships (Charmaz, 2006). By seeking theoretical similarities, differences and relationships to other concepts, theoretical sampling evolves and relates concepts into higher order categories that explain and predict behavior (1978). Constant comparison New interviews lead to new insights that often themselves suggest a need to reinterpret previous codes and categories. The process of moving back and forth between the activities shown in Figure 2 and comparing emerging concepts with those already uncovered forms the analytical engine of GT. This process is called constant comparison. In GT the researcher relies on interviews and other forms of data collection (e.g. observation, company records) to provide comparisons between different individuals, companies and events. Participants are selected based on their ability to inform emerging categories and properties. Constant comparison relies on theoretical sampling as researchers pursue subsequent interviews, reexamine previous interviews, and pose and answer testable hypotheses. Specifically, the researcher proposes a relationship based upon one set of interviews, observations and other forms of data, and then tests that relationship through follow-on interviews and data collection activities (Charmaz, 2006; Glaser and Strauss, 1967).

IJPDLM 42,8/9

Constant comparison provides a rigorous, process-oriented approach that illuminates theoretical structure (Cho and Trent, 2006; Glaser, 1992). Using constant comparison, the GT researcher uncovers repeatable patterns and tests those relationships using theoretical sampling. The repeatability of this process, albeit without the use of statistical methods, sets GT apart from other qualitative methodologies.

872

Sorting and adoption of categories As shown in step 7, new interviews fill out categories and lead to increasingly focused follow-on interviews and data collection. Categories, with their explanatory memos, are sorted into theoretical structures (Glaser, 2001). The focus is now on codes that relate one category to another and on finalizing the adoption of categories (see Randall et al. (2010) for an overview of this process). Saturation and diagramming In step 8, categories are refined, and theoretical relationships between the categories are made. Often GT researchers build a framework at this point of the process. Such frameworks relate antecedent categories (the left side of a diagram) to processes (the middle) and outcomes (the right) in a logical fashion. However, not all GT diagrams are constructed in this manner. Some GT diagrams show the interrelationships of concepts without placing them within a linear framework, since not phenomena proceed in a linear fashion. The process of constant comparison provides crisp categorical definition and relationships as the theory emerges (step 9). As the research progresses, the theory becomes more dense and complex, and the categories and their properties become conceptually saturated. Theoretical predictions are validated within the study using theoretical sampling, which eventually leads to theoretical saturation (Charmaz, 2006). Theoretical saturation occurs when follow-on interviews fail to add to existing categories, properties, processes, and relationships (step 10). Theory emergence The end result of theoretical saturation is the theory. The theory depict relationships between categories and the phenomenon (e.g. the effect of the category on organizational or individual behavior), macro conditions affecting the phenomenon (e.g. government regulations), structural conditions within the phenomenon (e.g. channel structures), and other important aspects of the theory. Diagrams are useful to the researcher because they: . integrate all of the emerging concepts from analysis of the data into a formulated theory; and . present the findings of the research in a concise manner. The goal of GT is to present an integrated model explaining most or all of the behavior associated with the phenomenon under investigation (Glaser, 1978). Assessing GT research GT has established methods of assessing rigor and trustworthiness. However, unlike other methods such as regression analysis and its associated r-value, GT lacks a concise measure of validity. To assist authors defending their work and reviewers/editors

evaluating such work, we outline assessment criteria common to good GT research. Glaser (1992) maintains that there are six key criteria for evaluating GT studies. Those six criteria are: fit, workability, relevance, modifiability, parsimony, and scope. Strauss and Corbin (1990) establish criteria for judging the adequacy of the research process and the empirical grounding. They suggest a study should include fit, understanding, generality, and control. Charmaz (1998) identifies credibility, originality, resonance, and usefulness as potential criteria for judging the effectiveness of the GT research. Table I organizes these criteria into general headings and suggests ways by which reviewers, editors and readers can analyze the quality of the study. Several of the criteria are essentially the same in concept but have been identified in the literature using different terminology. It is our position that regardless of the approach a GT researcher takes (e.g. Glaserian, Straussian, or Constructivist methods), the same set of criteria can be used to judge GT research. These criteria include rigor, trustworthiness, contribution to practice, and contribution to theory. The problematic side of GT As with any research methodology, GT has downsides, and we would be remiss if we did not include their discussion. GT greatest drawback is its interpretive nature. Potential bias is always a threat to the validity of interpretive findings, and therefore such findings can be subject to criticisms of researcher bias. For example, it could be argued that other interpretations of the data may better explain the phenomena. Also, because interviews are often open ended, participants have more control over the process than in quantitative methods. This openness means that if the researcher is not careful, participants may take interviews in directions of their own choosing. In addition, researchers have limitations in their ability to interpret data. An important issue could be overlooked if the researcher does not have theoretical sensitivity in that area (Glaser, 1978). Another issue relates to the theory-building, rather than theory-testing, nature of GT. A generalized validation of theoretical concepts and relationships cannot be claimed by the GT researcher. Such generalization requires empirical investigation using methodologies different from GT. Due to the relatively small sample size and selection of participants typically found in a GT study, findings cannot be easily generalized beyond the contexts and firms within the study. A third issue revolves around the fact that reviewers are often unfamiliar with GT processes and terminology. This does not necessarily mean that such reviewers are hostile to GT, only that they cannot easily determine how well the research was conducted. This puts the burden on the researcher to explain how the method works, what process was followed, and what steps were taken to ensure that the research was conducted with rigor. Additionally, from a research practicality standpoint, GT can be problematic. First, since the data are obtained in field settings, the researcher’s physical presence is required to gather most, if not all, of the data. This often requires the researcher to spend resources traveling to participating companies. For example, one of the authors spent a full month in a foreign country conducting interviews with participants. The other author spent weeks at various domestic manufacturing plants conducting interviews, sitting in on meetings and reviewing archival data. Second, the data analysis methods used in GT are time consuming. Open coding of interviews and the constant comparison of emerging concepts requires large amounts of

GT: an inductive method for research 873

Table I. Assessing GT Definition

Evidence of compliance

874

Rigor Fit (Glaser, 1992; Strauss and How well the categories and their relationships correspond (1) Content experts independently analyzed the researcher’s Corbin, 1990) to the reality of the area being investigated interpretations, processes, and memos. (2) Participants were asked to review these interpretations and comment as to whether or not the summaries accurately reflect the participants’ meaning. Interpretations from these sources were discussed and reconciled with researchers’ interpretations Modifiability (Glaser, 1992) Both positive and negative incidents are handled within the Researchers kept an audit trail that shows how and why the theory was modified. This audit trail can be maintained theory. Modification of the theory is required in order to through theoretical memos, figures detailing the evolution of accommodate new concepts discovered by comparing the theory, memos embedded in coding software, and other findings across people, places, and time types of records Resonance (Charmaz, 2006) How well the categories reflect the fullness of the studied Participants were asked to determine the extent to which the experience resulting theory makes sense and offers them deeper insights about the phenomenon. Researchers took feedback into account in findings Trustworthiness Credibility (Charmaz, 2006) Strong logical links exist between the data, its analysis, and (1) There is evidence of the use of theoretical sampling to the resulting theory; findings appear to represent the data select participating organizations and individuals. (2) There is evidence that the researcher made systematic well comparisons between empirical observations and between categories during their analysis. (3) The research provides enough evidence for the reader to form an independent assessment that agrees with the researcher’s claims Understanding (Strauss and The extent to which participants accept results as (1) Participants were asked to review summaries of research Corbin, 1990) representations of their worlds findings and verify that they reflected their stories. Discrepancies were reconciled. (2) Colleagues and practitioners were asked to review a summary of findings to verify that they buy into the findings (continued)

Heading

IJPDLM 42,8/9

Definition

Evidence of compliance

The theory provides an actionable framework with practical Participants and practitioners were asked to review implications summaries of research findings and agreed that the findings provide insights into the phenomenon that are meaningful to practitioners Usefulness (Charmaz, 2006) The analysis offers solutions to problems that people can The analysis offers useful interpretations that lead to better use in their work lives understanding or processes that can be applied to real world problems or situations Control (Strauss and Corbin, The extent to which organizations have influence over some The analysis offers variables within the theory over which 1990) aspects of the theory participants in the phenomenon have some degree of control and can therefore influence outcomes within the phenomenon Contribution to theory Participants were invited to comment on how well the Relevance (Glaser, 1992); The theory provides new or alternative explanations for research addressed the core issues of the phenomenon. This Originality (Charmaz, 2006) behavior that go beyond that offered in the literature. Relevance measures the weight of the theory’s contribution can be accomplished by providing participants with a summary of the research results after the researcher writes as a new or alternative explanation up the initial draft of the results. The focus of this process is on identifying core issues that have relevance for research (1) Evidence is presented that the researchers used The extent to which the research investigates, and the Generality (Strauss and theoretical sampling to select participating organizations Corbin, 1990); Scope (Glaser, resulting findings clarify, a broad variety of situations and individual participants. (2) Interviews and other data within and across multiple aspects of the phenomenon 1992) collection procedures were of sufficient detail and openness to provide a multi-faceted understanding of the phenomenon Parsimony (Glaser, 1992) To achieve parsimony, categories and their properties are The theory utilizes the minimum number of categories and limited to those that best help explain behavior properties that explain the majority of the behavior. Member checking sessions were used to assess how well the core category succinctly explains activities within the phenomenon without bogging down the theory with details of marginal value. Adjustments were made based upon participant feedback

Contribution to practice Workability (Glaser, 1978)

Heading

GT: an inductive method for research 875

Table I.

IJPDLM 42,8/9

876

time and considerable patience. As the sample size is based on reaching theoretical saturation the length of time required for data collection is unknown at the start of the research. It can often be difficult to determine whether theoretical saturation has been reached. The researcher may need to spend significant time sampling organizations and individuals before he or she is sure that no new theoretical concepts are emerging from the research. Third, the researcher must be skilled in interviewing, participant observation and qualitative analysis. Such skills must be learned and practiced. While qualitative analysis tools are available to assist the researcher, they will not perform important functions such as the determination of the most appropriate open code for a meaning unit or selection of core categories. These tasks are left to the skill of the researcher. Lastly, GT can be somewhat intrusive. Organizations may be reluctant to allow employees to participate in confidential interviews, may not want their employees to take time away from their jobs, may be uncomfortable allowing a researcher access for the purposes of observing organizational activities, or may have to refuse participation due to objections from their legal department. This often makes finding suitable firms willing to assist very difficult. Summary GT provides a pattern searching method that enables researchers to understand the underlying environmental and organizational structures and conditions that firms act within and upon. GT provides research results that are understandable to practitioners because the findings are grounded in the practitioners’ experience. These results are also timely as they provide salient issues of concern to practitioners. Lastly, GT offers results with theoretical significance in an area not previously researched or lacking strong theory. This paper demonstrates how GT, an inductive method, can be an appropriate tool for creating theory in supply chain research. In doing so, we have provided an overview of the GT technique and given a basic understanding of why GT works and where it might be applied. With that foundation, we provide a step-by-step guide of how the GT technique should be employed using examples from our own research. This guide can serve as a basis for understanding GT techniques in supply chain research. Finally, a method for assessing GT research, along with a table for researchers and reviewers to judge the merit of particular research, has also been provided. This paper distills the essential elements of “good” GT. We believe GT provides a holistic approach for furthering the development of SCM theory and hope that our work – creating a framework for GT research in SCM research – provides the tools that researchers need in seeking to discover and develop SCM theory. References Anderson, P.F. (1983), “Marketing, scientific progress and scientific method”, Journal of Marketing (pre-1986), Vol. 47 No. 4, p. 18. Bagozzi, R.P. (1984), “A prospectus for theory construction in marketing”, Journal of Marketing, Vol. 48 No. 1, pp. 11-29. Bagozzi, R.P. and Yi, Y. (1988), “On the evaluation of structural equation models”, Journal of the Academy of Marketing Science, Vol. 16 No. 2, pp. 79-94.

Bartels, R. (1970), Marketing Theory and Metatheory, R.D. Irwin, Homewood, IL. Belk, R.W., Sherry, J.F. and Wallendorf, M. (1988), “A naturalistic inquiry into buyer and seller behavior at a swap meet”, Journal of Consumer Research, Vol. 20 No. 3, pp. 34-63. Bender, P.S. (1981), “Mathematical modeling of the 20/80 rule: theory and practice”, Journal of Business Logistics, Vol. 2 No. 2, pp. 139-57. Bowen, G.A. (2008), “Naturalistic inquiry and the saturation concept: a research note”, Qualitative Research, Vol. 8 No. 1, pp. 137-52. Burrell, G. and Morgan, G. (1979), Sociological Paradigms and Organisational Analysis, Ashgate, Aldershot. Campbell, D.T. and Stanley, J.C. (1963), Experimental and Quasi-experimental Designs for Research, Rand McNally, Chicago, IL. Charmaz, K. (1998), “Grounded theory: objectivist and constructivist methods”, in Dezin, N. and Lincoln, Y. (Eds), Handbook of Qualitative Research, Sage, Thousand Oaks, CA. Charmaz, K. (2006), Constructing Grounded Theory: A Practical Guide Through Qualitative Analysis, Sage, Thousand Oaks, CA. Cho, J. and Trent, A. (2006), “Validity in qualitative research revisited”, Qualitative Research, Vol. 6 No. 3, pp. 319-40. CSCMP (2011), CSCMP Supply Chain Management Definitions, Vol. 2011, Council of Supply Chain Management Professionals, Lombard, IL. Defee, C.C., Williams, B., Randall, W.S. and Thomas, R. (2010), “An inventory of theory in logistics and SCM research”, International Journal of Logistics Management, Vol. 21 No. 3, pp. 404-89. DiMario, M., Cloutier, R. and Verma, D. (2008), “Applying frameworks to manage SoS architecture”, Engineering Management Journal, Vol. 20 No. 4, pp. 18-25. Drucker, P.F. (1962), “The economy’s dark continent”, Fortune, Vol. 5 No. 4, pp. 265-70. Eisenhardt, K.M. (1989), “Building theories from case study research”, Academy of Management Review, Vol. 14 No. 4, pp. 532-50. Gephardt, G.F. (2004), “Qualitative research and the Academy of Management Journal”, Academy of Management Journal, Vol. 47 No. 4, pp. 445-562. Glaser, B.G. (1978), Theoretical Sensitivity, Sociology Press, Mill Valley, CA. Glaser, B.G. (1992), Basics of Grounded Theory Analysis, Sociology Press, Mill Valley, CA. Glaser, B.G. (2001), Grounded Theory Perspective: Conceptualization Contrasted with Description, Sociology Press, Mill Valley, CA. Glaser, B.G. and Strauss, A.L. (1967), Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine de Gruyter, New York, NY. Gorod, A., Sauser, B. and Boardman, J. (2008), “System-of-systems engineering management: a review of modern history and a path forward”, Systems Journal, IEEE, Vol. 2 No. 4, pp. 484-99. Goulding, C. (1998), “Grounded theory: the missing methodology on the interpretivist agenda”, Qualitative Market Research, Vol. 1 No. 1, pp. 50-7. Goulding, C. (2002), Grounded Theory: A Practical Guide for Management Business and Market Researchers, Sage, Thousand Oaks, CA. Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E. and Tatham, R.L. (2010), Multivariate Data Analysis, 7th ed., Prentice-Hall, Upper Saddle River, NJ.

GT: an inductive method for research 877

IJPDLM 42,8/9

878

Haytko, D.L., Kent, J.L. and Hausman, A. (2007), “Mexican maquiladoras: helping or hurting the US/Mexico cross-border supply chain?”, International Journal of Logistics Management, Vol. 18 No. 3, pp. 347-63. Hirschman, E.C. (1986), “Humanistic inquiry in marketing research: philosophy, method, and criteria”, Journal of Marketing Research, Vol. 23 No. 3, pp. 237-49. Holbrook, M.B. and Grayson, M.W. (1986), “The semiology of cinematic consumption: symbolic consumer behavior in out of Africa”, Journal of Consumer Research, Vol. 13 No. 3, pp. 374-81. Holland, J.H. (1992), Adaptation in Natural and Artificial System, MIT Press, Cambridge, MA. Hunt, S.D. (1990), “Truth in marketing theory and research”, Journal of Marketing, Vol. 54 No. 3, pp. 1-15. Hunt, S.D. (1992), “For reason and realism in marketing”, Journal of Marketing, Vol. 56 No. 2, pp. 89-102. Hunt, S.D. (2002), Foundations of Marketing Theory: Toward a General Theory of Marketing, M.E. Sharp, Armonk, NY. Isenberg, D.J. (2008), “The global entrepreneur”, Harvard Business Review, Vol. 86 No. 12, pp. 107-11. Koh, S.C.L., Saad, S. and Arunachalam, S. (2006), “Competing in the 21st century supply chain through supply chain management and enterprise resource planning integration”, International Journal of Physical Distribution & Logistics Management, Vol. 36 No. 6, pp. 455-65. Kuehn, A.A. and Hamburger, M.J. (1963), “A heuristic program for locating warehouses”, Management Science, Vol. 9 No. 4, p. 643. Lambert, D.M. and Garcı´a-Dastugue, S.J. (2006), “Cross functional business processes for the implementation of service-dominant logic”, in Lusch, R.F. and Vargo, S.L. (Eds), The Service Dominant Logic of Marketing: Dialogue, Debate and Directions, M.E. Sharpe, Armonk, NY. Lambert, D.M., Garcia-Dastugue, S.J. and Croxton, K.L. (2005), “An evaluation of process-oriented supply chain management frameworks”, Journal of Business Logistics, Vol. 26 No. 1, pp. 25-51. Lieb, R. (2008), “The North American third-party logistics industry in 2007: the provider CEO perspective”, Transportation Journal, Vol. 47 No. 2, pp. 39-54. Locke, E.A. (2007), “The case for inductive theory building”, Journal of Management, Vol. 33 No. 6, pp. 867-90. McGrath, M. A., Sherry, J.F. and Heisley, D.D. (1993), “An ethnographic research of an urban periodic marketplace: lessons from the Midville farmer’s market”, Journal of Retailing, Vol. 69, Fall, pp. 280-319. Maltz, A.B. and Ellram, L.M. (1997), “Total cost of relationship: an analytical framework for the logistics outsourcing decision”, Journal of Business Logistics, Vol. 18 No. 1, pp. 45-66. Manuj, I. and Mentzer, J.T. (2008), “Global supply chain risk management strategies”, International Journal of Physical Distribution & Logistics Management, Vol. 38 No. 3, pp. 192-223. Meixell, M.J. and Gargeya, V.B. (2005), “Global supply chain design: a literature review and critique”, Transportation Research Part E: Logistics & Transportation Review, Vol. 41E No. 6, pp. 531-50. Mello, J.E. and Flint, D. (2009), “A refined view of grounded theory and its application to logistics research”, Journal of Business Logistics, Vol. 30 No. 1, pp. 107-26.

Mello, J.E. and Hunt, C.S. (2009), “Developing a theoretical framework for research into driver control practices in the trucking industry”, Transportation Journal, Vol. 48 No. 4, pp. 20-39. Mello, J.E. and Stank, T.P. (2005), “Linking firm culture and orientation to supply chain success”, International Journal of Physical Distribution & Logistics Management, Vol. 35 Nos 7/8, pp. 542-54. Mentzer, J.T. and Kahn, K.B. (1995), “A framework of logistics research”, Journal of Business Logistics, Vol. 16 No. 1, pp. 231-50. Mentzer, J.T., DeWitt, W., Keebler, J.S., Soonhoong, M., Nix, N.W., Smith, C.D. and Zacharia, Z.G. (2001), “Defining supply chain management”, Journal of Business Logistics, Vol. 22 No. 2, pp. 1-25. Meredith, J.R., Raturi, A., Amoako-Gyampah, K. and Kaplan, B. (1989), “Alternative research paradigms”, Journal of Operations Management, Vol. 8 No. 4, pp. 297-326. Mick, D.G. (1986), “Consumer research and semiotics: exploring the morphology of signs, symbols, and significance”, Journal of Consumer Research, Vol. 13 No. 2, pp. 196-213. Muller-Merbach, H. (1994), “A system of systems approaches”, Interfaces, Vol. 24 No. 4, p. 16. Pappu, M. and Mundy, R.A. (2002), “Understanding strategic transportation buyer-seller relationships from an organizational learning perspective: a grounded theory approach”, Transportation Journal, Vol. 41 No. 4, pp. 36-50. Pfohl, H.-C. and Buse, H.P. (2000), “Inter-organizational logistics systems in flexible production networks”, International Journal of Physical Distribution & Logistics Management, Vol. 30 No. 5, p. 388. Poist, R.F. (1974), “The total cost vs total profit approach to logistics systems design”, Transportation Journal, Vol. 14 No. 1, pp. 13-24. Poist, R.F. (1989), “Evolution of conceptual approaches to the design of logistics systems: a sequel”, Transportation Journal, Vol. 28 No. 3, pp. 35-9. Popper, K.R. (1959), The Logic of Scientific Discovery, Basic Books, New York, NY. Randall, W.S., Pohlen, T.L. and Hanna, J.B. (2010), “Evolving a theory of performance based logistics using insights from service dominant logic”, Journal of Business Logistics, Vol. 31 No. 2, pp. 35-62, available at: http://onlinelibrary.wiley.com/doi/10.1002/j.2158-1592.2010. tb00142.x/abstract Russell, D.M., Ruamsook, K. and Thomchick, E.A. (2009), “Ethanol and the petroleum supply chain of the future: five strategic priorities of integration”, Transportation Journal, Vol. 48 No. 1, pp. 5-22. Sauser, B. (2006), “Toward mission assurance: a framework for systems engineering management”, Systems Engineering, Vol. 9 No. 3, pp. 213-27. Shycon, H.N. and Maffei, R.B. (1960), “Simulation-tool for better distribution”, Harvard Business Review, Vol. 38 No. 6, pp. 65-75. Sousa-Poza, A. and Kovacic, S. (2008), “Research agenda for complex situations”, Engineering Management Journal, Vol. 20 No. 4, pp. 32-9. Strauss, A.L. (1987), Qualitative Analysis for Social Scientists, 1st ed., The Press Syndicate of the University of Cambridge, Cambridge. Strauss, A.L. and Corbin, J. (1990), Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, Sage, Thousand Oaks, CA. Strauss, A.L. and Corbin, J. (1998), Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, 2nd ed., Sage, Thousand Oaks, CA.

GT: an inductive method for research 879

IJPDLM 42,8/9

880

Suddaby, R. (2006), “What grounded theory is not”, Academy of Management Journal, Vol. 49 No. 4, pp. 633-42. Thomas, R. (2008), “Exploring relational aspects of time-based competition”, International Journal of Physical Distribution & Logistics Management, Vol. 38 No. 7, pp. 540-50. Thompson, C.J., Locander, W.B. and Pollio, H.R. (1990), “The lived meaning of free choice: an existential-phenomenological description of everyday consumer experiences of contemporary married women”, Journal of Consumer Research, Vol. 17, December, pp. 346-61. Further reading Anderson, P.F. (1986), “On method in consumer research: a critical relativist perspective”, Journal of Consumer Research (1986-1998), Vol. 13 No. 2, p. 155. Eisenhardt, K.M. (1991), “Better stories and better constructs: the case for rigor and comparative logic”, Academy of Management Review, Vol. 16 No. 3, pp. 620-7. Henwood, K. and Pidgeon, N. (2003), “Grounded theory in psychological research”, Qualitative Research in Psychology: Expanding Perspectives in Methodology and Design, American Psychological Association, Washington, DC. Lamma, E. and Mello, P. (1999), “Integrating induction and abduction in logic programming”, Information Sciences, Vol. 116 No. 1, pp. 25-57. About the authors Wesley S. Randall, PhD, is an Assistant Professor of Logistics at the University of North Texas. His research interests include supply chain performance, research methods, and supply chain finance. Dr Randall has secured over $620,000 in research grants in areas such as performance-based contracting and unmanned air system risk modeling. His research has been published in the International Journal of Physical Distribution & Logistics Management, Journal of Business Logistics, Journal of Supply Chain Management, Journal of Transportation Management, Transportation Journal, Defense Acquisition Review, and International Journal of Logistics Management. Wesley S. Randall is the corresponding author and can be contacted at: [email protected] John E. Mello, PhD (Tennessee) is an Associate Professor of Marketing at Arkansas State University. His research interests are in the areas of corporate culture, carrier management, and sales forecasting. Dr Mello has published in International Journal of Physical Distribution & Logistics Management, Journal of Business Logistics, Transportation Journal, Journal of Business Forecasting, and Foresight: The International Journal of Applied Forecasting.

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints