15 Mixed methods in interpretive research: an application to the study of the self concept Shalini Bahl and George R. Milne
Introduction
Interpretive research by its very nature is multi-method, involving diff erent forms of data collection, multiple studies, the use of triangulation, and so forth. Illustrative of a multimethods approach in interpretive research is the study by Belk, Wallendorf and Sherry (1989), which used a combination of observation, field notes, interviews, photographs and video. Less common, however, in interpretive research, is the use of mixed methods, which we define as a research approach comprising qualitative and quantitative methodologies for both data collection and analysis. The mixed methods approach is not frequently used in interpretive research owing in part to the paradigm wars, which have created artificial barriers that limit the use of methods across paradigms. In marketing, the positivist and interpretive paradigms have been profiled in terms of their ontological, axiological and epistemological assumptions (Hudson and Ozanne, 1988) that have been historically tied to the qualitative versus quantitative camps (Deshpande, 1983). In the marketing literature most articles employing mixed methods have a positivist orientation. The philosophical assumptions guiding positivist research include an objective view of reality, which the research seeks to measure and explain. Consistent with its ontological and axiological assumptions, positivist research seeks the creation of knowledge that is generalizable across di ff erent people, times and situations and is, thus, time- and context-free. While the methods used in positivist research have primarily been quantitative, qualitative methods have been used but only to support the quantitative methods in the development of measures, in the development of theory using grounded theory approaches or clarifying existing quantitative results. Examples of articles that use mixed methods in interpretive research are fewer in marketing. The philosophical underpinnings for interpretive research include the assumption of multiple realities that are socially constructed and the primary goal of understanding reality from the perspective of those experiencing it. In contrast to positivist research that seeks generalizations, interpretive research is typically time- and context-bound. Interestingly, the mixed method articles that ascribe to an interpretive approach in marketing tend to be more pragmatic, in that their primary focus is the research question, without being committed to a particular research approach and philosophy. Despite the trend for most mixed methods research articles in marketing to be either positivist or pragmatic, we believe it is possible for interpretive research to use mixed methods, and remain true to their philosophical assumptions. Mixed methods may be avoided because interpretive researchers feel that the use of quantitative methods requires altering one’s underlying assumptions. However, this is unnecessary since it is not the method but how it is used that needs to be consistent with the philosophical assumptions. 198
Mixed methods in interpretive research 199
That is, the appropriateness and use of a method is determined by the researcher’s orientation and the phenomena being studied (Morgan and Smircich, 1980). The flip side of this is that the methods do not define the type of study that is being conducted. As noted by Geertz (1973), methods do not define the research; rather, intellectual eff ort does. For example, qualitative methods can be and have been used successfully in positive studies. By the same logic, quantitative methods can be used in interpretive studies. In this chapter we demonstrate how quantitative methods can be used eff ectively in interpretive research. Thus the purpose of this chapter is to illustrate the use of mixed methods in interpretive research. To this end, we present a mixed method approach that was used in a study of the self concept. The balance of this chapter is in four sections. We begin by reviewing the mixed methods approach. Next, we discuss the application of a mixed methods approach to study multiple selves. This is followed by an empirical demonstration of this approach. In the last section, we discuss the advantages and shortcomings of mixed methods, and comment on possibilities for other interpretive work to use this approach. Mixed methods approach
Mixed methods is becoming a viable research approach that bridges the gap between quantitative and qualitative methodologies (Brewer and Hunter, 1989; Creswell, 2003; Tashakkori and Teddlie, 1998). Mixed methods are often seen as eclectic, pluralistic and rejecting traditional dualisms. Moreover they have been argued to be more pragmatic and driven by the research question rather than being constrained by paradigmatic assumptions (Johnson and Onwuegbuzie, 2004). Tashakkori and Teddlie (1998) outline the characteristics of pragmatism, where they feel that most mixed methods studies fit. The methods used are both quantitative and qualitative; the logic is both deductive and inductive. Moreover pragmatic research does not align itself with a single system of philosophy and reality (Creswell, 2003). In marketing, a wide range of mixed methods studies can be found in the continuum between pure positivist and pure interpretive studies. Some studies tend to be closer to the positivist end of the continuum, in that they use qualitative methods to support quantitative methods that are used to measure and explain phenomena. Others tend to be more interpretive, in that both qualitative and quantitative methods are used to understand the meaning of an observed phenomenon from the perspective of the consumer. Interestingly, many of the mixed method studies do not explicitly state their philosophical assumptions and, based on criteria outlined by Tashakkori and Teddlie (1993) and Creswell (2003), these studies could be viewed as pragmatic (e.g. Price, Arnould and Tierney, 1995). The classification of mixed method research has been undertaken by both Creswell (2003) and Tashakkori and Teddlie (1998). Creswell notes that mixed method research can be classified as having the quantitative and qualitative portions of the study, which can be conducted sequentially or simultaneously. Tashakkori and Teddlie (1998) emphasize that mixed methods influence multiple stages of the research, including type of inquiry, data collection/operations and analysis/inferences. In Table 15.1 we present a 2 3 matrix to show the range of mixed methods approaches that are possible. The columns in the matrix represent the orientation of the research article (positivist or interpretive) and the rows represent the mixed methods strategy used in the article (independent, overlapping and fully integrated). We classify articles as having a positivist or interpretive orientation according to the philosophical underpinnings implicit
Mixed methods in interpretive research 201
O’Guinn and Faber (1989) combine quantitative and qualitative data in a more interpretive manner. The purpose of their research was to provide a descriptive account of compulsive buyers. In their research they used in-depth qualitative interviews to contextualize descriptive survey data. This method included observing group therapy sessions, group and individual interviews and reading a thousand letters from compulsive buyers prior to fielding their survey. This was used to frame the study and form a sample for the mailing. In the study itself they used mixed methods. First, they used a mail survey to compare compulsive shoppers with other shoppers; second, they conducted five in-depth qualitative interviews with individual compulsive buyers to contextualize the quantitative findings. In this example of the overlapping strategy, data from both methods were analyzed and interpreted jointly. However, this article was not considered fully integrated because data from one method did not serve as input to the other method. The last strategy is fully integrated, in which data from one method serve as input for the other, and the data from both methods are analyzed together. McAlexander, Schouten and Koenig’s (2002) study of the jeep brand community and customer relationships within the community is an example of a fully integrated mixed method strategy with a positivist orientation. This research strategy starts with an ethnography, which was used to develop themes. A pre-test/post-test quantitative survey tested conceptual relationships using structural modeling. The hypothesis tested and the constructs used in the model were derived from ethnographic data. The two sets of findings were then put together and the findings ‘harmonized’. Since the ethnographic data were used to develop the survey and hypotheses and data from both methods were analyzed together, this article exemplifies the fully integrated research strategy. Arnould and Price’s (1993) article on service encounters provides another excellent example of fully integrated mixed methods strategy. The Arnould and Price (1993) ‘River Magic’ article is an interpretive study that has data collection over two years, where data are gathered from multiple participants. Focus group data served as input to surveys conducted at two diff erent time periods. Results from onsite depth interviews, participant observation (including field notes and photographs) and surveys (mail and drop-o ff ) were analyzed together to understand the relationships between client expectations and satisfaction.
In the next section we describe an interpretive study (Bahl, 2005), that uses a fully integrated mixed methods strategy for exploring multiple selves. In contrast to the interpretive studies that use mixed method strategies to explore relationships across aggregates of people, this study combines narratives and di ff erent multivariate techniques in a more integrated manner to help understand the qualitative data at the individual level. Typically, in mixed methods studies, qualitative data are used to inform quantitative methods. However this study is novel in that the quantitative results are used to inform the qualitative methods. And because data from the mixed methods are also analyzed in an integrated manner, this study is an example of the fully integrated strategy. Mixed methods approach for studying self concept
The self concept is a complex subject that has been studied using diff erent paradigms and methodologies. Recent studies recognize the multiplicity of the self concept. As such, it has been studied as situational self (Hogg and Savolainen, 1998; Schenk and Holman, 1980), multiple self-concepts – actual, ideal and social (Belch and Landon, 1977; Sirgy, 1982), malleable self (Aaker, 1999), fragmented self (Firat and Shultz, 1997, 2001;
Handbo Handbook ok of Qualit Qualita ati tiv ve Rese Resear arch ch Methods in Marketing
Edited by
Russell W. W. Belk Kraft Kraft Foods Foods Canada Canada Chair Chair of Market Marketing ing,, Schulic Schulichh School School of Business Business,, York University University,, Toronto oronto, Canada
Edward Elgar Chelte Cheltenha nham, m, UK • North Northamp ampton ton,, MA, USA
© Russel Russelll W. Belk, 2006 All rights reserv reserved. ed. No part of this publica publication tion may may be reproduced, reproduced, stored stored in a retrieval retrieval system or transmitted transmitted in any any form or by any any means, means, electronic, electronic, mechanical mechanical or photocopyi photocopying, ng, recording, recording, or otherwise without the prior permission of the publisher. publisher. Published by Edward Edward Elgar Publishing Limited Glensanda House Montpellier Parade Cheltenham Glos GL50 1UA UK Edward Edward Elgar Publishing, Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA
A catalogue record for this book is available from the British Library
Library of Congress Cataloguing in Publication Data Handbook of qualitative research methods methods in marketing/[edited marketing/[edited by] by] Russell W. W. Belk. p. cm. Includes bibliographical references and index. 1. Marketing Marketing research research—Metho —Methodolo dology gy.. 2. Consumers— Consumers— Resear Research—Met ch—Methodol hodology ogy.. 3. Qualitative Qualitative research— research—Metho Methodolo dology gy.. I. Belk, Russell ussell W. HF5415.2.H288 20 0 6 658.8’3—dc22 2006004283
ISBN-13: ISBN-13: 978 1 84542 100 100 7 (cased) (cased) ISBN-10: ISBN-10: 1 84542 100 0 (cased) (cased) Printed Printed and bound in Grea Greatt Britain by MPG Books Ltd, Ltd, Bodmin, Bodmin, Cornwall Cornwall
Contents List List of contrib contributo utors rs PAR ART TI
ix
HIST HISTOR ORY Y AND AND SCOP SCOPE E
1. History of qualitative research methods in marketing
3
Sidney Sidney J. J. Levy PAR ART T II
PARA ARADIG DIGMA MATIC TIC PERSPE PERSPECTI CTIVES VES
2. Breaking Breaking new ground: ground: developi developing ng grounded grounded theories theories in in mark marketing eting and consumer behavior
19
Eileen Eileen Fischer Fischer and Cele C. Otnes 3. The semiotic paradigm on meaning in the marketplace
31
David Glen Mick and Laura R. Oswald 4. Rethinking the critical imagination Je ff B. Murray Murray and Juli Juliee L. Ozanne Ozanne PAR ART T III III
46
RESE RESEAR ARCH CH CONT CONTEX EXTS TS
5. Qualitative research in advertising: twenty years in revolution
59
Linda Linda M. M. Scott Scott 6. Qualitative historical research in marketing
70
Terrence H. Witkowski and D.G D.G.. Brian Jones Jones 7. Researching the cultures of brands
83
Anders Bengtsson and Jacob Ostberg 8. Researching esearching brands ethnograp ethnographical hically: ly: an interpr interpretiv etivee commu community nity approach
94
Steven Steven M. M. Kates Kates 9. Maki Making ng cont conteexts xts ma matter tter:: sele select ctin ing g res resea earrch cont contex exts ts for theo theorretic etical al insi insigh ghts ts
106 10 6
Eric Arnould, Arnould, Linda Price Price and Risto Moisio PAR ART T IV
DATA COLLE COLLECTI CTION ON METH METHODS ODS
10. Netnography 2.0
12 9
Robert V. Kozinets 11. Let’s pretend: projective methods reconsidered
143
Dennis Dennis W. W. Rook 12. Stories: how they are used and produced in market(ing) research
1 56
Gillian Gillian C. Hopkinson Hopkinson and Margar Margaret et K. Hogg Hogg 13. The extended case method in consumer research
Steven Steven M. M. Kates Kates v
1 75
