The state of research in marketing

ejm cover (i).qxd

3/9/05

11:24 AM

Page 1

ISBN 1-84544-098-6

ISSN 0309-0566

Volume 39 Number 3/4 2005

European Journal of Marketing The state of research in marketing Guest Editors: Nick Lee and John Saunders

www.emeraldinsight.com

European Journal of Marketing

ISSN 0309-0566 Volume 39 Number 3/4 2005

The state of research in marketing Guest Editors Nick Lee and John Saunders

Access this journal online _________________________

243

Editorial advisory board __________________________

244

GUEST EDITORIAL Whither research in marketing? John Saunders and Nick Lee _____________________________________

245

COMMENTARY Where are we and where are we going? The status and future of research in marketing Dale Littler and Caroline Tynan __________________________________

261

Beyond incommensurability? Empirical expansion on diversity in research Andrea Davies and James A. Fitchett ______________________________

272

Grounded theory, ethnography and phenomenology: a comparative analysis of three qualitative strategies for marketing research Christina Goulding _____________________________________________

294

Qualitative research in marketing: road-map for a wilderness of complexity and unpredictability Evert Gummesson______________________________________________

Access this journal electronically The current and past volumes of this journal are available at:

www.emeraldinsight.com/0309-0566.htm You can also search more than 100 additional Emerald journals in Emerald Fulltext (www.emeraldinsight.com/ft) and Emerald Management Xtra (www.emeraldinsight.com/emx) See page following contents for full details of what your access includes.

309

CONTENTS

CONTENTS

The rise and fall of the Latin Square in marketing: a cautionary tale

continued

Robert P. Hamlin ______________________________________________

328

Equivalence of survey data: relevance for international marketing Hester van Herk, Ype H. Poortinga and Theo M.M. Verhallen __________

351

The evolution of ‘‘classical mythology’’ within marketing measure development Nick Lee and Graham Hooley ____________________________________

365

Long life to marketing research: a postmodern view Michela Addis and Stefano Podesta`________________________________

386

About the authors ________________________________

413

Call for papers ___________________________________

416

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

Additional complimentary services available

How to access this journal electronically

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

To benefit from electronic access to this journal you first need to register via the internet. Registration is simple and full instructions are available online at www.emeraldinsight.com/admin 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/0309-0566.htm More information about the journal is also available at www.emeraldinsight.com/ ejm.htm 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.

Key features of Emerald electronic journals Automatic permission to make up to 25 copies of individual articles This facility can be used for training purposes, course notes, seminars etc. 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 Fulltext and Emerald Management Xtra. 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. 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 to relevant and interesting articles to another computer for later use, reference or printing purposes. Emerald structured abstracts New for 2005, Emerald structured abstracts provide consistent, clear and informative summaries of the content of the articles, allowing faster evaluation of papers.

Your access includes a variety of features that add to the functionality and value of your journal subscription:

Research register A web-based research forum that provides insider information on research activity world-wide located at www.emeraldinsight.com/researchregister You can also register your research activity here. User services Comprehensive librarian and user toolkits have been created to help you get the most from your journal subscription. For further information about what is available visit www.emeraldinsight.com/usagetoolkit

Choice of access 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 Huber E-Journals e-journals.hanshuber.com/english/index.htm 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 TDnet www.tdnet.com

Emerald Customer Support For customer support and technical help contact: E-mail [email protected] Web www.emeraldinsight.com/customercharter Tel +44 (0) 1274 785278 Fax +44 (0) 1274 785204

EJM 39,3/4

EDITORIAL REVIEW BOARD

Professor Nicholas Alexander University of Ulster, Northern Ireland Dr Ali Bin Al-Khalifa University of Bahrain, State of Bahrain Professor George Avlonitis Athens University of Economics & Business, Greece Professor Michael Baker Westburn Publishers Ltd, UK Dr Susan Baker Cranfield University, UK Professor John Balmer Bradford Management Centre, UK Professor Jim Bell University of Ulster, Magee College, Northern Ireland Professor Bjo¨rn Bjerke Malmo¨ University, Sweden Professor Stephen Brown University of Ulster at Jordanstown, Northern Ireland Professor Francis Buttle Macquarie University, Australia Professor Tamar Cavusgil Michigan State University, USA Professor Bill Clarke University of Ulster, Northern Ireland Professor Brett Collins Auckland University of Technology, New Zealand Professor Nicole Coviello University of Auckland, New Zealand Professor David Cravens Texas Christian University, USA Professor Anthony Cunningham Co. Dublin, Ireland Professor Tevfik Dalgic University of Texas at Dallas, USA Dr Ken Deans University of Otago, New Zealand Professor Adamantios Diamantopoulos Loughborough University, UK Dr John Egan Middlesex University Business School, UK Professor John Fahy University of Limerick, Ireland Dr Kim Fam City University of Hong Kong, Hong Kong Professor Gordon R. Foxall Cardiff University, UK Professor Pervez Ghauri Manchester School of Management, UMIST, UK Professor Christina Goulding University of Wolverhampton, UK Ken Grant Monash University, Australia Professor Gordon Greenley Aston Business School, UK Professor Kjell Grønhaug Norges Handelshoyskole, Norway Professor Lloyd Harris Cardiff Business School, UK Dr Phil Harris Manchester Metropolitan University, UK Professor Roy Hayhurst University of Limerick, Ireland Professor Graham J. Hooley Aston Business School, UK European Journal of Marketing, Dr Gillian Hopkinson Vol. 39 No. 3/4, 2005 Lancaster University Management School, UK p. 244 Professor Ga´bor Hova´nyi # Emerald Group Publishing Limited Panno´nia UTCA, Hungary

244

0309-0566

Professor Claes Hultman ¨ rebro University, Sweden O Professor Mark Jenkins Nottingham University Business School, UK Professor David Jobber University of Bradford, UK Dr La´szlo´ Ka´rpa´ti University of Debrecen, Hungary Professor Hans Kasper University of Maastricht, The Netherlands Professor Erdener Kaynak Pennstate Harrisburg, Pennsylvania, USA Professor David Kirby University of Surrey, UK Professor Philip Kitchen The University of Hull, UK Professor Simon Knox Cranfield University, UK Professor Raymond LaForge University of Louisville, USA Professor Uolevi Lehtinen University of Tampere, Finland Professor Barbara Lewis Manchester School of Management, UMIST, UK Professor Veronica Liljander Swedish School of Economics and Business Administration, Finland Professor Andrew McAuley University of Stirling, UK Professor Jan Mattsson Roskilde University, Denmark Professor Bill Merrilees Griffith University, Australia Professor Morgan Miles Georgia Southern University, USA Professor dr Carla Millar TSM Business School, The Netherlands Professor Luiz Moutinho University of Glasgow Business School, UK Professor Patrick Murphy University of Notre Dame, USA Professor Aron O’Cass The University of Newcastle, Australia Professor Adrian Palmer University of Gloucestershire, UK Professor Paul Patterson University of New South Wales, Australia Professor Chad Perry The Gap, Queensland, Australia Professor Nigel Piercy University of Warwick, UK Professor David Shipley University of Dublin, Ireland Dr Wai-sum Siu Hong Kong Baptist University, Hong Kong Professor Richard Speed Melbourne Business School, Australia Professor Peter Turnbull University of Birmingham, UK Professor Caroline Tynan Nottingham University Business School, UK Professor Eduard Urban University of Economics, Czechoslovakia Dr Cleopatra Veloutsou University of Glasgow, Scotland, UK Professor Salvatore Vicari Bocconi University, Milan, Italy Professor Martin Wetzels Eindhoven University of Technology, The Netherlands Professor Len Tiu Wright De Montfort University, UK

GUEST EDITORIAL

Guest editorial

Whither research in marketing? John Saunders and Nick Lee Aston Business School, Aston University, Birmingham, UK

245

Abstract Purpose – To introduce the contents of this special issue on research in marketing and comment on the development of the discipline in UK universities. Design/methodology/approach – Relates each paper to a taxonomy of academic research and comments on their content. Examines major trends in higher education and relates them to the fortunes of marketing educators. Findings – There are reasons to be cheerful about academic marketing in the UK: there are clearly opportunities to publish in the world’s leading academic journals, increased funding for the discipline, the acceptability of a wide range of methodologies and the increasing influence of marketing. Less encouraging is the naı¨ve and destructive competition between universities and the consequent destruction of academic development. Research limitations/implications – This is a UK perspective that depends on limited knowledge of other than a few other countries. Practical implications – There are good reasons to be positive about an academic career in marketing, but also a desperate need to tackle the naı¨ve strategies of universities and to intervene to mend the gaps in the development of academic marketers. Originality/value – Gives an insight in to the range of research in marketing, and an insight into the opportunities and pitfalls of a career in academic marketing research. Keywords Market research, Universities, United Kingdom

About the Guest Editors John Saunders is Professor of Marketing, Head of Aston Business School and Pro-Vice Chancellor of Aston University. Besides being Dean of the Senate of the CIM, he has senior responsibilities within the EFMD and AACSB International. In addition, he has served on the research selectivity exercises of the UK, The Netherlands and New Zealand and on accreditation panels for EQUIS, AACSB International and the QAA. His research has covered market modelling and marketing strategy but focuses most recently on managerial incompetence. His articles have appeared in Journal of Marketing Research, Journal of Marketing, Marketing Science, International Journal of Research in Marketing, Journal of International Business Studies, Journal of Business Research, European Journal of Marketing, and many others. Nick Lee is a Lecturer in Marketing at Aston Business School. His research interests include sales management, electronic commerce, and marketing research philosophy and methodology. He has presented his work at international conferences around the world, and was the recipient of the 2002 EMAC Conference award for best paper based on doctoral work, as well as a 2002 UK Chartered Institute of Marketing Research Initiative Award. He lectures in marketing research at undergraduate and postgraduate level.

Reasons to be cheerful: part 1 Dale Littler and Caroline Tynan’s powerful commentary in this special issue suggests that we have reason to despair, but we also have reasons to cheerful about the state of

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 245-260 q Emerald Group Publishing Limited 0309-0566 DOI 10.1108/03090560510581755

EJM 39,3/4

246

research in marketing. Our subject is blossoming. The long established marketing academies, including AM, AMA, EMAC and MSI, are thriving to such an extent that EMAC is forever concerned about its annual conference getting too large and inclusive. As our discipline globalises new associations, such as ANZMAC, are flourishing and spawning new journals. Much of this energy comes from demand for our subject from undergraduate and graduate students who pay most of our wages and justify our existence. MBA and other graduate programmes, that have always been associated with the strongest research tradition, account for most of the increased demand. So strong is the growth in this market that Britain is producing more MBAs per head of population than the USA[1]. This burgeoning market contrasts with other parts of higher education where demand is slipping. Across the world, students are spurning science and technology for the green fields and blue skies of marketing, psychology and other social sciences disciplines. Marketers should have some concern about this trend because someone still has to design and make what they sell. Even though academic salaries are notionally standardised across disciplines and across institutions in state universities, markets are at work and demand is pushing up salaries and promotion prospects for marketing academics. Americans invented academic marketing and they still dominate the subject. Some have bemoaned the dominance of “American journals” which are biased against non-American research. Many still hold this view though the evidence is otherwise. The Benelux countries have a string of academics that are top of the food chain in the very journals that are supposedly “biased”, and their data sets and first language are certainly not American. Among these are colleagues at Tilburg, Groningen and Louven[2]. Within the UK, the successes of Stephen Brown (Brown et al., 2003), and Adamantios Diamantopoulos (Diamantopoulos and Winklhofer, 2001) show a diversity of European research that has overcome this supposed prejudice. The circumvention of once impregnable journals is not restricted to the odd individual, the majority of current output of at least three marketing groups in the UK is within “fortress USA”. The point here is not to elevate US journals over those published elsewhere in the world, but to celebrate their lack of geographical exclusivity. Reasons to be cheerful: part 2 Taking a more parochial view, the picture in Britain is particularly rosy. British business schools take up solid share of the world’s leading MBA programmes, although the top slots are still held by America’s hugely endowed private schools (The Economist, 2003). British business schools also have a much more embedded research culture than in the majority of European counties, particularly the larger ones. This often becomes apparent in EQUIS accreditations that uncover research outputs and research cultures that are undeveloped compared with those in English-speaking countries. Realistically European business schools will never compete with their US competitors until university vice-chancellors and business school deans focus on raising money rather than spending it. The Oxford Centre for Higher Education Funding (Palfreyman, 2002) squeals that expenditure per undergraduate at relatively wealthy Oxford is one-third of the levels of Harvard and Princeton, the teaching load per academic is twice as high and salaries are a third to a half lower. Notwithstanding

their academic source, their figures are misleading. The OxCHEPS calculations add all of the university’s endowment income to undergraduates then forgets that Oxford also receive £80 million from the Higher Education Funding Council for England (HEFCE) for research and staff development. When calculating funding it is as reasonable to add this to the income per student as it is to allocate the whole of endowment income. Although HEFCE research income is not for teaching it does pay more than half of some academic salary bills, increases the number of staff a university can employ to teach the same number of students and therefore brings down teaching loads. When the fee income per students rises from £1,050 to £3,000, Oxford’s total fee and state funding per undergraduate will be higher than Harvard. However, the fees increase will do little to overcome Oxford’s complaints. The gap in income is the huge endowment income of the US school that, based on OxCHEPS’ figures, account for 82 per cent of the cost of educating Harvard’s undergraduates. Adding Birmingham, a Russell group university, and UCE, a new university shows the funding gap per students between Oxford and Birmingham is as wide as that between Oxford and Harvard. The reason why Oxford students are so well funded compared with Birmingham’s students is Oxford’s superior endowments and the government’s extra funding of £2,000 per year given to the colleges. Even though Birmingham’s endowments and research income are hugely greater than UCE’s, their overall funding per student is surprisingly close. Whatever the figures, for the moment many British business school deans are off the hook. The HEFCE has just increased the “band D” funds for most marketing students by over 20 per cent starting 2004/2005. In addition, marketing can also access some higher level “band C” funding that is available for cost centre 7 (psychology and behavioural sciences) and 25 (IT and system sciences) (HEFCE, 2003a). The money is there, but anticipate that many heads of business school and heads of marketing will have to spend the next few years getting their hands on it. Then Tony Blair risked his Government over increased student fees. The Higher Education Bill is still in the committee stage at the time of writing but it is likely that fees will rise from £1,050 to £3,000 for most business students within a few years (Grice, 2004). The impact of England’s proposed increase in student fees also shows why the super-e´lite colleges are shouting about wanting to set fees higher than £3,000 (see Table I). Because Oxford has so much other income, the increased fee only increases their income by 9 per cent. In contrast, Birmingham University can hope to receive 27 per cent more per student and UCE 38 per cent. That is, if UCE have the pulling power to charge the maximum fee allowed. Few academic marketers across the world have as rosy funding prospects as the British. In the USA many state colleges are cutting their student intakes to keep within ever-declining state budgets. In most of Europe and many other developed economies, academics suffer in regimes where funds are low and where universities have to accept all students who matriculate. The result is huge first year classes with little student support and where most students drop out or take many years to graduate (The Economist, 2004). The opening address of one of Sorbonne’s professors to his class of 80 students gives an insight: Too many. Half of you will have to leave. When we are down to 40 I’ll start teaching. Foreigners will go first.

Guest editorial

247

EJM 39,3/4

Oxford

Harvard

2,000 80

9,000 360

1,000 40

57 2

1 0

59 80 139

0 0 0

0 0 0

67 34 101

41 1

10,500 5,500 16,000

6,680 10,820 17,500

6,000 5,750 11,750

16,765 7,923 24,688

12,873 2,290 15,163

State research funding Per student Per graduate

5,000 7,619

0 0

0 0

1,377 2,028

66 78

Endowment income Per student Per graduate Undergraduate fee

5,000 7,619 1,050

20,571 53,892 15,940

3,404 6,667 15,620

92 136 1,050

3 3 1,050

State funding Per undergraduate Undergraduate financial aid

6,000 0

0 23,990

0 22,190

4,000 0

4,000 0

Income per undergraduate Teaching plus endowment Teaching, endowment and other state funding Expressed as percentage of Oxford

14,669 22,288 100

65,842 65,842 295

20,097 20,097 90

5186 7214 32

5,053 5,131 23

Impact of increased fees Income per student Increased income

24,238 9

9,164 27

7,081 38

Endowment capital: £ millions Endowment yield: £ millions (4 per cent)

248

State funding (2002) Teaching: £ millions Research: £ millions Total: £ millions Students Undergraduate Postgraduate Total

Table I. Poor little rich colleges

Duke

Birmingham

UCE

No wonder that wealthy French students study at the fiercely selective grande e´coles. It is also clear why the OECD sees the university funding model that caused Tony Blair such difficulties as a “role model for Europe” (Green, 2004). Reasons to be cheerful: part 3 Part of the joy of being a researcher in marketing is its sheer variety and eclecticism. Figure 1 tries to capture some of the range, but is a dendrogram that will always grow, and whose choice of dendrites will always be personal. The papers in this special issue show this profusion of approaches and cover many parts of the framework. In the first paper[3], Davies and Fitchett study a divide that often occurs across between the qualitative, quantitative and modelling traditions in our dendogram, provoking a rethink of what appears to have become a key issue within marketing research – both in an academic and social sense (as attendance at many conferences will demonstrate!). Specifically, while it often seems to many of us that there is an unbridgeable divide between the epistemological “camps” of positivism and interpretivism which have developed in marketing over the last few decades, Davies

Guest editorial

249

Figure 1. Dendrogram of research in marketing

EJM 39,3/4

250

and Fitchett argue that this divisive approach is unnecessary and potentially harmful. Beginning with a discussion of how Kuhn’s (1970) paradigmatic incommensurability thesis has been misunderstood within marketing research, they go on to show that one should consider the difference between how we think about research, and how we actually go about doing research, leading to a re-examination of the very concept of a “paradigm”. They illustrate their argument with a discussion of two studies on the same topic, each conducted independently under different epistemological assumptions. Ultimately, adherence to ideas of paradigm incommensurability is shown to be an unhelpful and artificial barrier to greater understanding, leading to marginalisation of potentially useful findings and the irrelevance of marketing research to all but those inside their narrow paradigmatic bunkers. The second paper, by Goulding, fits rather more neatly into the qualitative branch in Figure 1, as a comparison and review of three different qualitative strategies. Here, Goulding provides an interesting counterpoint to Davies and Fitchett, by touching on the paradigmatic debate described above, but considering it to be slowing down rather than still raging. Goulding discusses the methodologies of Grounded Theory, ethnography, and phenomenology, giving an overview of the philosophical underpinnings of each, as well as an appreciation of the key authors and topics relevant to each approach within marketing research. Furthermore, the paper provides an excellent introduction to the process of conducting research using each of the three approaches. One of the key contributions of this paper is its clear exposition of exactly what these three methodologies actually entail. The need for such an appreciation is clear since, as Goulding argues, for too long have marketing researchers rather carelessly used terms such as “grounded theory”, “ethnography”, and “phenomenology”, often to indicate simply the use of a qualitative approach. Goulding provides a valuable overview of the key differences between such methods, and their characteristics. She also argues that these methods deserve wider application within marketing research, rather than their current marginalisation within consumer-related research, which will hopefully inspire readers who may have previously ignored these methodologies to investigate them further. Gummesson’s paper is also squarely located within the “qualitative” branch of our dendogram. While his paper is a very personal view, it is clear that it also touches on many of the themes that run throughout the content of this special issue. In particular, Gummesson appears to concur with Davies and Fitchett’s argument that qualitative and quantitative approaches should not be seen as fundamentally opposed, but instead as complementary. Gummesson’s paper is also relevant to Goulding’s work, with its focus on the actual process of doing research. In particular, Gummesson argues that we are doing our students, and the readers of our research, a disservice by focussing on the data collection methodology of our studies rather than the analysis and interpretation of the qualitative data themselves. He makes a large contribution by integrating a huge amount of existing theoretical knowledge with his own considerable experience to create a set of useful guidelines for researchers. Following this, Gummesson concentrates on the generation of theory itself – and again Kuhn (1970) makes an appearance. In this section Gummesson discusses a number of viewpoints which should stimulate considerable debate amongst marketing academia – such as that we should not solely consider theory generation as a cumulative process, nor accept unquestioningly everything we read simply because it is in a certain journal or by a

certain author. Finally, similar to Addis and Podesta below, Gummesson recommends that each of us follows his/her own path to generate knowledge within a pluralistic framework, rather than blindly participating in what he calls the “power games” of the mainstream. As an example he then discusses his own method of choice – interactive research. Whether or not you agree with his thesis, Gummesson in this paper does what all good research should do, makes us think. The next three papers provide something of a break from the qualitative or integrative approaches presented so far, being as they are each about a key quantitative concept that has had a long history within marketing research. First, directly referencing branch 2.2.4 of our tree diagram, Hamlin provides a very interesting historical analysis of the Latin Square experimental design in marketing research – often using a metaphor of the “Titanic” to illustrate his points, although unfortunately without visual reference to Kate Winslet. The narrow focus of Hamlin’s paper provides two key contributions. First, he is able to provide an excellent introduction to the Latin Square (and other experimental designs in general), which should be of particular interest to our younger readers raised on the hegemony of survey research. Second, and perfectly in keeping with the aims of this special issue, Hamlin is able to draw implications for academic research in marketing with his cautionary tale of how the Latin Square design was “done to death” by errors in its application and interpretation. This case study has important implications for marketing researchers – some of which are also touched on by Lee and Hooley in their paper. Specifically, poor application of analysis techniques does not mean that the techniques are poor, merely that the researcher is poor. Furthermore, we as researchers, teachers and reviewers have a duty to understand fully the techniques we use, rather than simply quoting precedence, which can allow serious errors to compound one another – eventually leading to invalid results and the death of rigorous research. All of us would do well to take note of Hamlin’s work here, and its relevance to our own. The next paper is by van Herk, Poortinga and Verhallen, and it deals with a topic that has generated considerable interest among marketing researchers recently – that of the cross-cultural equivalence of survey data – thus relating to branch 2.1 of the dendogram. Van Herk, Poortinga and Verhallen contribute to this area by integrating what may have become two rather disparate approaches – that of the statistical assessment of measurement invariance, and the more general assessment of the research design itself. Ignorance of the latter is likely to introduce bias into research designs, which is compounded by ignorance of the former, meaning one does not realise the biases inherent in one’s data. Ultimately, not taking into account either of these approaches may lead to erroneous conclusions – and worse, one may never be aware of the errors. Van Herk, Poortinga and Verhallen’s paper demystifies and synthesises the cross-cultural equivalence issues, which can appear rather arcane to the uninitiated researcher, and they also provide some extremely useful guidelines on how to interpret the results of equivalence tests. The paper, like Hamlin’s, also has considerable relevance to commercial market researchers (and users of commercial research) with its integration of an interesting commercial study as an example and specific reference to areas where commercial research needs to improve its methods. Lee and Hooley’s paper is squarely located in section 2.1.4 of our dendogram, that of psychometrics. However in a similar, although less overt, way to Hamlin they also attempt to draw some wider implications for marketing research from their discussion

Guest editorial

251

EJM 39,3/4

252

of one specific technique. Psychometrics has become one of the driving forces of marketing research in the last 25 years, and Lee and Hooley attempt to elucidate some of the key issues regarding the evolution of these techniques in marketing research. Specifically, they discuss what have become unquestioned assumptions in psychometric measure development in marketing – which they term “classical mythology”. In particular, Lee and Hooley deal with the statistical techniques of measure purification, such as exploratory factor analysis and internal consistency, which have been the mainstays of measure development within marketing for over 20 years, and are still widely used today. Lee and Hooley link the underlying theory of psychometrics to our accepted assumptions, such as the widely quoted cut-off point of 0.7 for coefficient alpha, to show that – rather than being absolute rules – these assumptions are often somewhat groundless and should be viewed as such. As a result they recommend that researchers should learn to interpret their results rather than merely report them. In fact, they suggest that the dogmatic adherence to psychometric “rules” could have actually detracted from the validity of our findings rather than enhanced it, a situation which may be seen in other branches of marketing research as well. Thus, within this paper are some valuable conceptual and technical tools for interpreting existing work, as well as to guide research in the future. Our final paper is by Addis and Podesta`, and deals with an issue which has been the source of much debate within marketing research over the past decade – both within the hallowed halls and journals of marketing academia, and the smoky bars many of us also frequent – namely the postmodern “turn”. We might consider this piece of research as slotting into branch 1.1.3 of our tree, although doubtless others would argue it cannot be so atomistically classified. In their paper Addis and Podesta` consider that while the establishment of postmodern thought within marketing would necessitate a serious rethink of the discipline, this “revolution” could have major benefits. Essentially, the rejection of the scientific metanarrative and the removal of objectivity from our frames of reference means we would operate without the “safety net” of scientific rationality – which they contend could also constitute a cage. Thus, similar to Gummesson’s work, they advocate freedom for marketing researchers – and in this case freedom even from the concepts of true and false. Research is judged herein on a single criterion, does the knowledge it creates enrich the individual and/or the community? To illuminate some of these concepts, Addis and Podesta` trace the history of thought in marketing research, beginning from the 1950s, arguing that it has become stagnant and trapped inside the boundaries of narrow specialisms and methodologies. They then move to a discussion of relationship and experiential marketing, two approaches that challenged the existing direction of marketing research. Addis and Podesta` argue that, although both approaches criticised the application of marketing knowledge, neither actually attacked the knowledge generation process itself. By contrast, Addis and Podesta` provide a fascinating deconstruction of the scientific process, leading to some thoughts on the possible postmodern future of marketing research. Ultimately, Addis and Podesta` strongly deny the charge of nihilism which has sometimes been levelled against postmodernism, and point to a future familiar to all of us who watched the “Matrix” trilogy – where we are freed from “the machine”, and are responsible for our own knowledge generation. Free to conduct research, as Neo might say “because we choose to”.

Reasons to be doleful: part 4 But enough of optimism! There is plenty of reason to be doleful about research in marketing. Success is damaging the discipline. Marketers would suggest that in a crowded market, like that for university education, competitors should specialise. Unfortunately university vice-chancellors or principles have not read the book, so are careering towards lemming-like mutual destruction and pain (Saunders et al., 2000). Rather than focussing on their strengths, vice-chancellors are responding to student demand and making universities more alike. Marketing aside, university after university is closing down languages, science and engineering departments while diving into anything that is likely to increase scale. Media studies was the “joke” of the 1990s that teaching quality and research assessments have shown to be not so funny. The explosion in student numbers means psychology is rapidly becoming “the joke” of the new millennium, although it will also be fun watching how troubled Boots absorbs the huge supply of pharmacists that new programmes will soon be tipping onto an oversubscribed market. As part of business studies, marketing is nearing the tail end of a headlong dash into a mature and crowded market. For the last two decades, every so often, a lethargic university wakes up and says: “business school, we’ll have one of those too”, Depending how weak the analysis, the next move is either to slam together bits of the university that could have something to do with business (a bad idea), or confuse a business school with having a big building and then slam bits together (an even worse idea). The discipline of marketing, along with a few others, crops up when it is realised that the slammed together bits leave core subjects missing. The consequence is a vortex of decline (see Figure 2) that leaves the disciplines severely damaged. A traditional academic in a research-oriented school would follow track “I” from a taught research-oriented degree, through a doctorate and post-doctoral research to an academic position, and finally to a full professorship. That is the track in strong academic disciplines that are close to marketing, such as economics and psychology. Comparing the ideal track I with that occurring in many leading business schools, track ‘II’, shows many gaps: . Taught degrees. The taught degrees are often aggressively practical and career oriented. A successful business degree has graduates who are highly employable by blue chip companies, but have rarely studied a subject in depth and have little inkling of what social science research involves. The USA has the same problem and is the reason why leading schools recruit their research students from engineering, science, economics or psychology rather than their own business degrees. . Doctoral degrees. The orientation of business graduates or research students coming in from other disciplines means that doctoral programmes have to cover the basics of research, not the advanced research topics studied in US programmes. The frequent lack of methodological skills means many graduating doctors of philosophy cannot even understand articles in the leading journals in our discipline, never mind publish in those journals. . Post-doc. Such is the demand for marketing faculty that the traditional “post doc” research post has disappeared. Instead, students are recruited off doctoral programmes before their research is finished or complete their degrees while

Guest editorial

253

EJM 39,3/4

254

Figure 2. Vortex of decline

.

occupying teaching posts. Another reason for this gap is the lack of large funded research projects on which post-docs traditionally work. Lecturers. There is so much demand for lecturers that Europe has no tradition of the great recruiting conference typified by the summer AMA. At this, US schools vie for the best PhD students supervised by the top researchers. European schools often recruit practitioners, graduates, or people part of the way through their PhD. In some disciplines the academic life and that of practitioners are so

.

.

similar that people can switch between one and the other. This is not so in marketing where practice is completely unlike the life of a research academic. Young academics from aspiring schools (track III) are a rich source of young recruits for the leading schools. Just as soon as they as they show signs of research productivity, the leading schools pounce and steal the talent. Career progression. Budding researchers should have a chance to learn their craft as lecturers. In an ideal school, they would be part of a community of fellow academics at various stages of their career development. Such breeding grounds of excellence are becoming increasingly hard to maintain. Now the aspiring schools retaliate for their earlier losses by offering senior posts to still-developing young researchers. Leaps from lecturer to professor are common. These transfers plug gaps in developing schools but damage the culture in the donating school. Worryingly, these young academics lose the research infrastructure that could allow them to grow and are prematurely plunged into research leadership, or even more onerous administrative responsibilities. Aspiring schools have two strategies for managing their new appointments. The first alternative is to obtain immediate payback from their expensive acquisition by allocating them a major administrative role. That is likely to kill the goose that is supposed to continue laying golden eggs. The second alternative is to privilege the acquisition with low teaching and research support. This keeps the bird laying although, without the research infrastructure, less prolifically than before. The full cost to the aspiring school comes in paying for the cuckoo, which is fed by taking resources from their young colleagues. Professors. The vortex of decline leaves professors in all organisations isolated. Those in the leading schools have lost their lieutenants while those in the aspiring schools never had them. From on high, the answer to this problem is clear. Hire some more professors. Enthusiastic deans no longer talk of an additional professor of marketing, but two or three. Sadly the vortex of decline has ensured that such clutches of research leaders do not exist. The result – another round of head hunting and culture-destroying promotions.

The very success of marketing is destroying it. It takes years to train and allow an excellent marketing academic to develop, but the environment where they can grow and develop is very hard to maintain. Only when universities realise that business and marketing are not a quick fix will our discipline have time to grow organically and gain strength. By increasingly focusing research income, Britain’s HEFCs are trying to transmit the message that mediocrity does not pay. But while universities continue to confuse big and broad with best, they will cultivate the bland. Oh for the focus of London’s LBS, LSE and Imperial; Boston’s Harvard and MIT; or Europe’s HEC, IMD and INSEAD! Another strategic whim that has previously had Aston, Bradford, Birmingham, Cardiff, Huddersfield, Imperial, Leicester, Loughborough, MBS, UCE and UMIST in the headlines is the dream of mergers – another manifestation of big, broad and bland. Part of the rationalisation for this is an endearing faith in scale economies. It was even proposed that Imperial (11,495 students according to HESA (2004)) and UCL (17,805 students) should merge to achieve the scale of Harvard (17,500 students)! It is interesting to note that one of HEFCE’s own studies shows that scale does not reduce

Guest editorial

255

EJM 39,3/4

256

costs significantly since multi-site operations, that are the usual outcome from mergers, drive up central costs (Brignall and Pierce-Brown, 2004). Mergers cover whole universities, not just business schools, but since many highly marketable people work in business schools, it is they that are likely to suffer the mass efflux of staff that mergers precipitate. Some lessons from mergers are true for businesses as well as universities. Mergers are exciting for senior managers but distract attention from the core business, in our case conducting excellent teaching and research. They also have a devastating effect on morale and invariably degenerate into cost cutting to recover the original investment. Another reason to be doleful is the unfortunate scale of Britain’s marketing community. In strategic management “b*****d size” refers to a business unit that is too small to reap economies of scale yet too large to niche. Maybe academic marketing communities have the same problem. The mighty USA has the greatest academic power, but beyond that scale is not the thing. The next group of developed countries by size includes the European quartet France, Germany, Italy and Spain. Although they are economic middleweights, on the world academic scene they are featherweights. This is not because they lack marketing academics but is because they are large enough to have their own self-contained academic marketing communities. German academics become invisible to the rest of the world because they are expected to publish in German journals. Most of the French output goes into French journals and some in the community refer to “A” journals (meaning top international) and “AN” (national journals that are rated “A” in their home country only). In these cases the local publishing opportunities are great enough to absorb most output. This has two effects. First, never mind how good the local research and journals are, the output is invisible elsewhere in the world. Second, local traditions and research conventions emerge that do not necessarily make the local scene a stepping stone to the wider academic community. For a long time in the UK, that meant researching only a few of Figure 1’s dendrites. Until recently much research was managerial (dendrite 1.1.2), case oriented (1.2.2) or surveys of managers (2.1.2). Recently these have been supplemented by an enthusiasm for scale development (2.1.3), critical theory (1.1.1) and postmodern work (1.1.3). So intense is the focus on researching managers that some companies have managers filling out questionnaires for most of their working week, and questionnaire fatigue is growing (Maitland, 2004). Meanwhile the focus elsewhere in the world has been on consumer research (2.1.1 and 2.1.4), observation (2.2), model building (3.1), and analysis (3.2). The national tradition increases the parochialism of national research as the traditions and methodology of the local community grow out of synch with the rest of the world. The countries with large academic communities are the academic “b******d size”. Below them in size are a string of countries that are too small to be independent, so play the international game. Chief among these are the Benelux, Scandinavian and Australasian countries. Still, there is reason to be cheerful. While most people are trapped in their comfortable national community, a few brave souls are venturing forth to cross swords outside their national boundaries. This itself can cause a split. A venturesome academic either frequents both national and international conferences or becomes invisible at home. That is why British marketing

academics see little of people from LBS, and some of the leading Dutch marketing academics ignore EMAC. Reasons to be cheerful: part 5 So why are we cheerful? Publishers who travel to conferences across all disciplines and time zones recognise that there is something odd about marketing academics. They are generally happy. Why? There are a number of possible reasons. First, enough of us are close enough to business to recognise the joys of academic life. We can look in practitioner journals and see that even marketing directors are not paid all that much. We know the demands, uncertainty and risk of practice. We also know the joy of being able to take a long-term view; to take time over trying to do things properly rather than chasing short-term deadlines; the joys of research; indulging one’s inquisitiveness at other people’s expense. The joy of going into a bookshop and reading and realising that it is work! Joys in a marketing academic’s life also come from not having to fight over a fixed domain. Marketing practice is forever changing, always at the frontier of new developments, be they e-business or E coli. Being an eclectic discipline, it borrows, adapts or develops ideas from other fields. Marketing, like engineering, is about results so any means will do. Marketing also glorifies in its relevance. Sometimes for good, sometimes not; marketing is everywhere. With manufactured designs being transient and making “things” getting cheaper, value is in the mind and it is marketing that adds that value. Relevance also means that we are graced with students whose degree gives them a career, a chance in life. Certainly academic life has its difficulties but we are free spirits in an expanding and developing discipline. The vortex of decline makes life difficult but we can intervene. We can develop programmes as academic as our research. Where students learn how to create as well as use knowledge. There is certainly room for e´lite programmes that train marketing scientists with tools and skills that will make them invaluable to market research, consulting and the City in addition to academic life. We are also free to collaborate, so there is no reason why such programmes could not be shared across institutions. The Bologna Accord’s expectation of a 3 þ 2 (three-year undergraduate and two-year postgraduate degree) means that we will have to create a second year of postgraduate study to be able to award “European Masters” degrees. How great it would be if the second year concentrates on advanced research methods and an extended project. In Britain we also have reason to be cheered about the HEFCE’s decisions not to back an e´lite club of universities. The Russell Group tried numerous tricks to concentrate research in the middle-aged universities. The campaign was damaged by their own super-e´lite trying to grasp all the goodies for themselves (Oxbridge, plus the London trio of Imperial, UCL and LSE). Instead, funds will flow to excellent 3* or 4* research, wherever it is located (HEFCE, 2003a). However, the HEFCE research funding model is applying an inexorable pressure for centres of excellence. As research funding is focussed, a few institutions will have a concentration of internationally recognised researchers. One day, even vice-chancellors will recognise what strategists have understood for generations: it pays to concentrate, not dissipate. Figure 3 shows why. Where resources increase with effort, the return is

Guest editorial

257

EJM 39,3/4

258

Figure 3. Returns to scale

higher (3r) if resources are concentrated on one “Unit of Assessment” 2 (UoA) rather than the same effort being dissipated across UoA 1 and UoA 2. The increasing return on effort of research selectivity (HEFCE, 2003b) emphasises a return on scale that is general to academic activities (Figure 4). That explains the focus of the USA’s top universities, the clear focus of two of the self-selected e´lite UK institutions (Imperial and LSE) and the size distinction between 4, 5, and 5* rated business schools. The final reason to be cheerful is the pliability of marketing that allows it to fulfil anyone’s desires:

Figure 4. HEFCE’s increased funding concentration

.

Believe in capitalism? Well, that is what marketing is all about.

.

Anti-global? Then marketing is a way of presenting ideas and marshalling support.

.

Want to win an election? Political marketing will tell you how and will probably be discussed as much as policies.

.

.

Want to help the underprivileged? Marketing will help you raise money, increase awareness of people’s plight, and educate people about how to help themselves. Want to succeed as a marketing academic? Try marketing: find out what your target market is and respond accordingly.

Our job as academic marketers is to find out those things that help ourselves, and help others, achieve what they want. Our mission is to find ever better ways of “finding out what people want” and “responding accordingly”. The quest will never end because people and the technology will always change. That is why we have reason to be cheerful. Notes 1. Calculation based on MBA programme sizes reported in The Economist (2003). 2. The 2003 output of just three authors (with their colleagues) in what the leading US business schools class as tenure track journals includes Dekempe: Deleersnyder et al. (2002), Gielens and Dekimpe (2001), Nijs et al. (2001); Steamcamp: Baumgartner and Steenkamp (2001), Nijs et al. (2001); Widel: Brangule-Vlagsma et al. (2002), Elpers et al. (2003), Brangule-Vlagsma et al. (2002), Haaijer and Wedel (2001), Hofstede et al. (2002a, b), Kamakura et al. (2003), Pieters et al. (2002), Sandor and Wedel (2001). Unfortunately for Europe, Widel has just left for a chair at Wharton. 3. The order of the papers was determined strictly on alphabetical lines (as originally submitted), excepting the commentaries herein and by Littler and Tynan. References Baumgartner, H. and Steenkamp, J.-B.E.M. (2001), “Response styles in marketing research: a cross-national investigation”, Journal of Marketing Research, Vol. 38 No. 2, p. 143. Brangule-Vlagsma, K., Pieters, R.G.M. and Wedel, M. (2002), “The dynamics of value segments: modeling framework and empirical illustration”, International Journal of Research in Marketing, Vol. 19 No. 3, p. 267. Brignall, S. and Pierce-Brown, R. (2004), “Administrative efficiency in UK universities”, paper presented at the EAA Annual Congress, Prague, 1-3 April. Brown, S., Kozinets, R.V. and Sherry, J.F. Jr (2003), “Teaching old brands new tricks: retro branding and the revival of brand meaning”, Journal of Marketing, Vol. 67 No. 3, p. 19. Deleersnyder, B., Geyskens, I., Gielens, K. and Dekimpe, M.G. (2002), “How cannibalistic is the internet channel? A study of the newspaper industry in the United Kingdom and The Netherlands”, International Journal of Research in Marketing, Vol. 19 No. 4, p. 337. Diamantopoulos, A. and Winklhofer, H. (2001), “Index construction with formative indicators: an alternative to scale development”, Journal of Marketing Research, Vol. 38 No. 2, p. 269. (The) Economist (2003), Which MBA?, 15th ed., Economist Intelligence Unit, London. (The) Economist (2004), “Who pays to study?”, 22 January. Elpers, J., Woltman, L.C.M., Wedel, M. and Pieters, R.G.M. (2003), “Why do consumers stop viewing television commercials? Two experiments on the influence of moment-to-moment entertainment and information value”, Journal of Marketing Research, Vol. 40 No. 4, p. 437. Gielens, K. and Dekimpe, M.G. (2001), “Do international entry decisions of retail chains matter in the long run?”, International Journal of Research in Marketing, Vol. 18 No. 3, p. 235. Green, M. (2004), “Tuition fee proposal hailed as ‘role model for Europe’”, Financial Times, 21 January.

Guest editorial

259

EJM 39,3/4

260

Grice, A. (2004), “Chancellor appeals to MPs for unity in vote on top-up fees”, The Independent, 25 March. Haaijer, R. and Wedel, M. (2001), “Habit persistence in time series models of discrete choice”, Marketing Letters, Vol. 12 No. 1, p. 25. HEFCE (2003a), Developing the Funding Method for Teaching for 2003/4, HEFCE 2003/42, HEFCE, London, August. HEFCE (2003b), Recurrent Grant for 2003/4, HEFCE 2003/10, HEFCE, London, March. HESA (2004), available at: www.hesa.ac.uk/holisdocs/pubinfo/student/institution0102.htm Hofstede, F.T., Kim, Y. and Wedel, M. (2002a), “Bayesian prediction in hybrid conjoint analysis”, Journal of Marketing Research, Vol. 39 No. 2, p. 253. Hofstede, F.T., Wedel, M. and Steenkamp, J.-B.E.M. (2002b), “Identifying spatial segments in international markets”, Marketing Science, Vol. 21 No. 2, p. 160. Kamakura, W.A., Wedel, M., de Rosa, F. and Mazzon, J.A. (2003), “Cross-selling through database marketing: a mixed data factor analyzer for data augmentation and prediction”, International Journal of Research in Marketing, Vol. 20 No. 1, p. 45. Kuhn, T.S. (1970), The Structure of Scientific Revolutions, University of Chicago Press, Chicago, IL. Maitland, A. (2004), “Companies face an avalanche of questionnaires”, Financial Times, 26 March, p. 10. Nijs, V.R., Dekimpe, M.G., Steenkamp, J.-B.E.M. and Hanssens, D.H. (2001), “The category-demand effects of price promotions”, Marketing Science, Vol. 20 No. 1, p. 1. Parlfreyman, D. (2002), “Higher education in the United Kingdom: a viable e´lite-mass ‘third way’?”, Occasional Paper, No. 4, OxCHEPS, Oxford. Pieters, R., Warlop, L. and Wedel, M. (2002), “Breaking through the clutter: benefits of advertisement originality and familiarity for brand attention and memory”, Management Science, Vol. 48 No. 6, p. 765. Sandor, Z. and Wedel, M. (2001), “Designing conjoint choice experiments using managers’ prior beliefs”, Journal of Marketing Research, Vol. 38 No. 4, p. 430. Saunders, J., Stern, P., Wensley, R. and Forrester, R. (2000), “In search of lemmus lemmus: an investigation into convergent competition”, British Journal of Management, Vol. 11 No. 1, pp. S81-S95. Further reading Geyskens, I., Gielens, K. and Dekimpe, M.G. (2002), “The market valuation of internet channel additions”, Journal of Marketing, Vol. 66 No. 2, p. 102. Sandor, Z. and Wedel, M. (2002), “Profile construction in experimental choice designs for mixed logit models”, Marketing Science, Vol. 21 No. 4, p. 455. Scheer, L.K., Kumar, N. and Steenkamp, J.-B.E.M. (2003), “Reactions to perceived inequity in US and Dutch interorganizational relationships”, Academy of Management Journal, Vol. 46 No. 3, p. 303. Srinivasan, S., Pauwels, K., Hanssens, D. and Dekimpe, M. (2002), “Who benefits from price promotions?”, Harvard Business Review, Vol. 80 No. 9, p. 22.

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0309-0566.htm

COMMENTARY

Where are we and where are we going? The status and future of research in marketing

The status and future of research 261

Dale Littler Manchester School of Management, UMIST, Manchester, UK, and

Caroline Tynan Nottingham University Business School, Nottingham University, Nottingham, UK Abstract Purpose – The paper strives to assess the current status of research in marketing. Design/methodology/approach – The paper examines statistics showing the current position of market area studies. The paper also looks at the range of supply and demand factors affecting the quality and volume of research in marketing. Findings – The paper argues that there is a range of extrinsic factors that affect in some way how research in marketing is undertaken, how it is perceived more widely and the form of its output. Research in marketing also has its own internal drivers that reflect in particular the current position of its knowledge base, and the manner in which those engaged in research perceive research potential. Finally, on the basis of the analysis the paper suggests some practical means of enhancing research in marketing. Originality/value – This is important in identifying areas of concern that can be addressed in order to enhance the status of marketing. Keywords Market research, Education, Educational development Paper type Viewpoint

Introduction Many marketing academics when asked will generally register their dissatisfaction about the lack of recognition given to research in marketing. They may employ as evidence the low levels of research council support for research in marketing; they may even note the low regard often given to marketing by social scientists, including some of those in other areas of business and management. Even practitioners are voicing their concerns about business’s failure to give marketing the regard they consider it merits. Recent campaigns run by, inter alia, the Chartered Institute of Marketing suggest that marketing does not often have top executive representation, while the contribution they allege marketing makes to overall corporate performance is not acknowledged. Practitioners may not universally accept that academic marketing research is either relevant or makes a significant contribution to practice; and this in turn reinforces the low esteem that academic marketers feel that others in general hold of marketing per se and more specifically of the research that there is pressure on them to undertake.

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 261-271 q Emerald Group Publishing Limited 0309-0566 DOI 10.1108/03090560510581764

EJM 39,3/4

262

The resources to engage in in-depth scholarship and extensive programmes of research are scarce while the considerable external pressures stemming from uncertainties in Government policies, the rapid expansion in business and management and the lack of availability of appropriately trained academic staff all add to the difficulties of marketing. In addition, there is still a dilemma concerning the legitimate domain of marketing (Wensley, 2000) and as recently as 1995, Wensley (1995) noted the lack of cumulative research on issues in marketing including such central topics as market structure and segmentation, market networks and inter-firm relationships, marketing orientation and the implementation of the marketing concept. The paper strives to assess the current status of research in marketing. We argue that there are a range of extrinsic factors that affect in some way how research is marketing is undertaken, how it is perceived more widely and the form of its output. Research in marketing also has its own internal drivers that reflect in particular the current position of its knowledge base, and the manner in which those engaged in research perceive research potential. Finally, on the basis of our analysis we suggest some practical means of enhancing research in marketing. The current position It is widely accepted, certainly by those in the business and management academic community, that Economic and Social Research Council (ESRC) support for research in their area is in general unsatisfactory. Of the 16 disciplines for which the ESRC is responsible, both in 2000/2001 and 2001/2002 Management and Business Studies had the lowest success rate (21 per cent and 16 per cent respectively) of ESRC research grant applications. For comparison, “area studies” had success rates of 67 per cent and 50 per cent, while the figures for economics were 48 per cent and 40 per cent (Dobson, 2003). The success rates for applications to the defined research programmes in 2002-2003 was 33 per cent (ESRC, 2003, p. 75) with a total of 17 out of the 109 proposals supported being in “management and business” of which ten were in the area of “evolution of business knowledge’. For all the research programmes for which full statistics are available, for 2002/2003 only around 1.87 per cent of the total number of original applications was awarded in the area of management and business (ESRC, 2003, p. 77). The ESRC-supported Commission on Management Research established in 1993 identified some possible ways forward. Unfortunately this does not appear to have resulted in any long lasting significant positive results. More recently the AIM (ESRC, 2004) initiative sought to support business and management research although in very specific areas and it is too early to assess the possible effects, although the limited number of fellowships awarded, with only one in marketing, suggests that the effects will be limited. Business and management research often seeks support from elsewhere such as charitable trusts, but these have limited resources for which there are many competing demands. The majority of other funding sources often require solutions to relatively short-term policy and other questions. This scarcity of and major sources of available funding may have a negative effect on the development of basic knowledge in business and management. There is also much current dissatisfaction with the research output, with questions about the “relevance” and contribution of business and management research (see, for example, the comment by Varadarajan, 2003, p. 368) For example, the

Commission on the Social Sciences singled out research in business and management as the “greatest concern” because it was of a standard well below that of other research (Utley, 2003). If the position facing business and management researchers in general is difficult, it is even more so for those in marketing. Given its share of business and management it could justifiably be argued that marketing is significantly under-funded, a point reinforced by Bessant et al. (2003, p. 56), who noted that for some time academic marketing has had to survive without significant funding from the ESRC. The accusation that the ESRC review process is biased against marketing per se can be countered by the fact that the ESRC is only responsible for putting in place the evaluation of proposals that it receives and these are then subject to peer review. Possible explanations for the low success rate may reside in the inadequate “quality” of many of the proposals received or intrinsic inefficiencies in the peer review process. Many proposals may fail to secure support because they are deficient in themselves with poor definition of objectives, an inadequate literature review, insufficient attention to the development of a detailed, robust methodology, or a lack of attention to how the data will be analysed. There may also be difficulties in securing the support of a sufficient body of referees, both in terms of the number of academic marketers who are prepared to act in this role and, of those who do review, the time and effort they can afford or are willing to devote to reviewing. Moreover the lack of agreed paradigms unlike in other social sciences may lead to a higher level of disagreement by reviewers that may not enhance the ability of proposals to secure approval when considered by ESRC committees. The final decisions on acceptance or rejection are made by groups that are generally not exclusively “business and management” oriented; in fact in 2002/2003 only 8 per cent of the academics on the various ESRC boards were from management and business studies (ESRC, 2003). Few of these are likely to be familiar with the challenges of research in marketing. Here another influence may come into play: a possible perception that somehow or other marketing is not of the same calibre as other business and management subjects let alone other social sciences. Whether or not there are intrinsic biases either in perception or in process certainly demands greater analysis. However, the belief that there is such a bias can discourage marketing academics from investing in the development of proposals for this important academic funding body, leading to a reinforcement of the perception, supported by the statistics, that few research in marketing proposals are funded! There is also the question about the international status of marketing: in fact, the analysis of the last research assessment exercise (RAE) returns points to a concentration in UK-based journals, with a significant proportion of these not being seen as particularly prestigious. Some involve limited refereeing. There were few publications in what are regarded as “international journals” which is usually synonymous with top league US publications. It is the US journals that dominate the top rankings of academic marketing journals. Here British authors “cannot compete on an equal basis because they face implicit and explicit barriers when attempting to publish in the US” (Easton and Easton, 2003, p. 16). Apart from a possible prejudice against UK academics publishing there, it is evident that given the intense competition with US academics there are limited opportunities for the several hundred UK academics to publish in these prestigious “international” journals.

The status and future of research 263

EJM 39,3/4

264

There is at least one other dimension to the current position of marketing that merits mention at this stage and that is the increasing concern about its possible future given the difficulties in recruiting marketing academics that are disciplined researchers and effective educators across a range of levels. There are fears that there will insufficient numbers entering the profession to replace those who retire, with possible consequential deleterious impact on research in marketing. There are a range of supply and demand factors that in some way affect the quality and volume of research in marketing. These include the: . availability and source of resources to support research in marketing; . perceptions of research in marketing; . institutional process for assessing research in marketing; . “teaching” demands of marketing academics; . supply of marketing academics; and . internal dynamics of marketing. Resources We have already noted that there is severe competition for resources for what might be termed scholarly, long term, more basic research in the social sciences in general, and more especially in marketing. The ESRC allocates only 9 per cent of funds to management and business that in terms of student numbers represents 50 per cent of the social science community (Learning and Teaching Support Network, 2001). The policy of increasing selectivity with the concentration of funding in those departments that achieved a 5 or 5* suggests that many marketing academics cannot rely for any meaningful support from their own institutions which themselves have seen their research grant curtailed. Even where this policy is not being strictly followed, as in Scotland, the lack of growth in the availability of funds for research per se means the opportunities for universities to allow significant research time will undoubtedly be limited. Marketers then like their cousins from other business and management areas, seek research funding from organisations, such as regional development agencies, with a clear policy agenda; knowledge partnerships schemes (formerly teaching companies); businesses; and from consultancies. The majority of the research pursued in this way is likely to be oriented towards developing analyses and even solutions desired by the funder. The paymaster dictates, and the outcome is more likely to be concerned with assisting relative short-term practice rather than providing theoretical and conceptual insights. It can be specific rather than general knowledge, although this is not inevitable if those participants in practice-based research have the ability to extend the understanding beyond the specific context in which it was gained in order to develop new perspectives and question extant theories and concepts. However, it tends not to be the case that such funders favour supporting such extensions of the research while the researchers often do not have the luxury to undertake the requisite long-term reflection and development because of the pressures of managing relationships with the funders, seeking out new funders and the heavy burdens of teaching and administration. Unfortunately, the articulated policy of generally concentrating more research resources within a top research tier of universities, the increasing allocation of

funds by the Higher Education Funding Council to targeted socio-economic ends and the inadequate supply of both funds and students for doctoral research suggests that the lot of many researchers in marketing may not improve, and indeed may even worsen.

The status and future of research

Research evaluation The UK has been a pioneer in the national evaluation of research quality by discipline and institution through its periodic RAE. This is in essence a peer review evaluation of, in the case of business and management, the research outputs of individual business and management departments, although in arriving at an overall evaluation a significant number of factors are taken into account. Whether or not the quality of the four publications required for each submitted academic in the last RAE is the deciding factor is debatable; however, the view, widely held, is that they do have a significant influence on the grading given the emphasis by the Panel on reading and reviewing a substantial proportion of them. The RAE undoubtedly places pressure on academics to publish, and in particular to ensure that they have at least four “worthy” publications. It could reasonably be argued that this results in “quick fix” research: research that is likely to lead to publishable results in the short term. It may then shape the type of research is marketing that is undertaken. There may, for example, be a tendency for academics not to engage in long-term “blue skies” research, longitudinal studies that demand considerable preparation and extensive fieldwork, and in-depth conceptualisation or theory construction. Instead, one can see the advantages of collecting cross-sectional data generally by questionnaire, and subjecting this to often sophisticated statistical analysis; undertaking a series of mini case studies; or capitalising on student dissertation projects. Undoubtedly academics are now aware that to satisfy institutions’ ambitions to have a significantly financeable research rating (one that is currently 5 or above) it is important to have the requisite number of publications in what might be seen as at least reasonable journals with some form of independent refereeing process. The pressure to publish per se then can have implications for the type of research in marketing that is favoured. There is, however, the question of the frequency with which it is possible to report weighty research. Research, empirically based or not, that is substantial and which is likely to have a significant impact will tend to require considerable “space”: time for scholarship, for reflection, for analysis for dialogue with other scholars: space that is not often available. The time and effort involved in preparing papers for many of the top flight journals is probably beyond the means of the average UK marketing academic. This is reflected in the conclusions of the RAE panel for Business and Management: “The mass of research in marketing . . . does not exhibit the methodological rigour or theory-building that would come close to . . . international academic standards” (Bessant et al., 2003, p. 56). It is also evident that there are increasing pressures on the Higher Education Funding Councils to concentrate funding in those departments where there is a track record of what has been judged to be “quality” research, as reflected in the RAE grading. This policy is founded on the assumptions that institutions will continue to maintain their record, despite changes in structures and personnel, that the criteria which were applied at the time the assessment was made were valid and pertinent and

265

EJM 39,3/4

266

will continue to be so, and that only “good” research can be undertaken. This may result in the concentration of knowledge production in just a few centres, thereby discouraging diversity and stifling innovative and creative thinking. The focusing of Government funding means that many academics will have little if any opportunity to engage in certainly fundamental knowledge creation. As the Higher Education Minister noted in September 2003: . . . we are acutely aware of the vulnerability of our nation’s world class research base. Ensuring that our very best departments receive the funding they need to compete internationally means that we must be selective in spending these new funds. The hard truth is we can no longer fund departments rated 4, 3a and 3b as before, but we are not saying that public funding will stop altogether (Johnson, cited in Research Fortnight, 2003, p. 1).

There has also been renewed questioning of the dual support system for university research. This system relies on two mechanisms for distributing research funding: the RAE to allocate resources for research infrastructure; and the research councils open review system to support specific research projects. In an independent study commissioned by the Royal Society (compiled by the independent research and policy analysts, Evidence) it was concluded that the cost of maintaining the current dual support funding streams amounted to £230 million (Fazackerley and Goddard, 2003). An alternative approach would be to simplify the procedure by using a system of competitive tendering based on full economic costs for the allocation of research funding, an approach, however, that is likely to deprive even more academics of the opportunities to participate in research which in turn could adversely affect teaching. Even if it is accepted that there is only a tenuous link between research and teaching, a link that many marketing academics see as both concrete and symbiotic, such a policy change could have profound deleterious implications for the standard of the education available to a significant proportion of university students. Teaching The massive, under-funded growth in university education since the 1970s has resulted in conditions and teaching loads which are far less than ideal. The whole university system produces three times as many students as it did 30 years ago (Elias and Purcell, 2003), but the resources to do so have not kept pace with the growth. The experience of undergraduates in 2004 with huge class sizes, limited tutorials and inaccessible faculty is unrecognisable to those who graduated 30 or even 20 years ago. The policy of continued expansion of higher education to embrace those from less privileged backgrounds can in some cases place an additional onus on academics who may be faced with groups of more than usual mixed ability with the need to undertake necessary remedial work. The pain associated with this growth in student numbers is not evenly spread throughout the university and disproportionately affects business and management departments because of the high levels of student demand, from both the UK and overseas. Indeed, the demand for business and management degree programmes remains buoyant with in 2001/2002 almost 12 per cent of all degree level students opting to read for degrees in this area, an increase of 11 per cent on the previous year (HESA, 2003 cited in ABS, 2003, pp. 28-9). Moreover, in 2001-2002 business and management degree programmes were the most popular for non-UK students studying in the UK at both undergraduate (16.5 per cent of the total) and postgraduate (22.6 per cent) levels

(HESA, 2003 cited in ABS, 2003, p. 3). Within business and management marketing remains highly popular, particularly at second year and final year levels, with student cohorts for some courses totalling hundreds. The direct contact time may be limited, but the servicing of such courses can be both time consuming and exhausting. This leaves marketing academics with comparatively high contact hours and poor conditions even when compared to academics in other subject areas. The notorious two-tier university system makes this systematic under-resourcing, even more of a problem for those in the post-1992 universities. Students sometimes encouraged by their parents may often regard themselves as “consumers” because they are now both directly and indirectly paying for a service in terms of fees, maintenance, and opportunity costs. Understandably students demand a high level of educational experience and many are increasingly likely to take appropriate action, either through a formal complaints procedure or even litigation, if they perceive that this is not being provided. At both institutional and individual levels processes have been introduced to ensure there is available clear information about programmes and courses, there are transparent procedures for evaluation and complaints, and there are records of inter alia meetings and other contacts with students. The Quality Assurance Agency itself has developed a range of policies the implementation of which are periodically reviewed in time consuming assessments. All of this has made the academic’s job even more onerous. Supply of marketing academics It is widely recognised that there is shortage of research-trained staff to teach business and management at all levels in the universities (Learning and Teaching Support Network, 2001) and most academic managers of marketing departments are familiar with the problems of recruiting high quality marketing academics. Moreover, a recent survey by Baker and Erdogan (2000), reported that 60 per cent of marketing academics were over 40 years old. This confirms the ageing population, which is going to exacerbate recruiting difficulties. In addition there is a lack of good young academics and research students in marketing to replace the retirees. Even now, we are recruiting too few good, young marketing academics. This might be because the current low pay, excessive workloads (more than one-quarter of Baker and Erdogan’s sample reported working over 51 hours per week on academic activities) heavy teaching loads and relatively poor conditions are not attractive. It might also be that for our best graduating students consulting and management offer a lucrative alternative to academe. The increase in indebtedness caused by the introduction of higher fees for a university degree will inevitably exacerbate this trend. For those research students we do manage to recruit, the RAE panel have questioned whether most doctoral training programmes equip them to do research to the necessary standard, be that in a critical or positivist paradigm (Bessant et al., 2003, p. 56). Whether this is related to the lack of a relevant academic background the lack of a doctorate among the faculty is a matter of conjecture but Diamatopoulos et al. (1992, p. 16) raised both issues in their 1992 survey of UK marketing academics. They concluded “Marketing academia at UK universities is male-dominated, rapidly ageing, with little industrial/commercial full-time work experience, under-qualified at the doctoral level, largely inappropriately qualified with regard to subjects read prior to entering the profession and badly equipped to keep up with new developments through

The status and future of research 267

EJM 39,3/4

268

membership of professional organisations”. Burton (2003) recommends that it is an appropriate time for the UK’s marketing academics to rethink research training provision in the light of the ESRC initiatives on reconfiguring the PhD programme to include formalised research training in year one, followed by three years focusing on a thesis. Marketing knowledge At least part of the explanation of the low apparent funding of research in marketing and its perceived low status may be that marketing per se is an embryonic social sciences’ discipline that has yet to develop its own coherent paradigm. Its core concepts are, for the most part, derived from more established disciplines, in particular economics, sociology, and psychology. Many of its methodologies are those that have been initially developed and applied elsewhere, while their application without adequate training can itself result in poor quality research. For instance, the contemporary discovery of ethnography by marketing academics may lead to its adoption without the detailed in depth training that the valid use of this complex methodology demands. In fact it may also be as Pringle (2001, quoting Claycamp, 1971) argues that we spend too much time creating new and sophisticated techniques and too little to operationalising them. Marketing per se may be viewed as an applied discipline driven by the demands of practitioners. Indeed there is a widespread acceptance that marketing should aim to solve business problems by developing “best practice”, a tendency that is reinforced by the availability of resources to undertake practitioner-oriented research, either through some form of consultancy, knowledge transfer partnerships and postgraduate student funding. The obstacles, perceived or otherwise, to securing funding for more conceptual research also support this practice-based orientation. Knowledge about practice should not be undervalued: it has its own validity and is the essence of many technology subjects, as well as medicine where the emphasis is increasingly on evidence-based practice. Whatever the form of knowledge, the methodology by which it is developed must be transparent, together with clear evidence, so that the status of “the knowledge” can be clearly evaluated. Moreover, it should be recognised that the creation and development of knowledge is not restricted to universities or indeed individual institutions, but is often the product of a complex of influences within a diverse range of structures. Indeed the formation of marketing knowledge may necessarily involve intimate dialogue with practitioners. Research therefore can be viewed as not “being institutionalised primarily within university structures . . . [it] involves the close interaction of many actors throughout the process of knowledge production . . . [it] makes use of a wider range of criteria in judging quality control” (Gibbons et al., 1994). There are, however, worries about what is regarded as the lack of “relevance” both of business and management research in general and of marketing in particular. For example, the editor of Marketing Science suggested: “There is concern that our work is becoming less relevant to practitioners and that we are less successful than we could be an impacting practice” (Ratchford, 2001, pp. iii-iv). The focus on “relevance”, however, begs the question of what this means and who defines it? The significance of research output can be time dependent, reliant on developments extraneous to the research and specific to particular users. The

worthiness of much basic science for example was often not realised at the time of its discovery, but only emerged later as opportunities for its exploitation were realised through social, economic, political and technological developments. To take one example: the implications of identifying the structure of DNA were not fully recognised at the time of the “discovery”. The scientists were driven by the challenge of being the first to unravel this scientific mystery; not by its relevance to resolving paternity suits; engineering plants and crops; or predicting and curing inherited diseases. These “relevancies” emerged later. Often the “best” science is driven internally by the curiosity of scientists who have an appreciation of the potential returns from the development of particular knowledge trajectories and who invest their intellectual energy in striving to realise these. This is not to argue that they undertake research free from any extraneous influences; but rather to state the case for a more balanced view of the production of knowledge that recognises that it embraces a range of actors from a variety of institutional contexts, and that relevance itself is a shifting and context specific notion that may in itself not be the initial, primary driver. Nevertheless, it is likely that in any future research evaluation process there will increase emphasis on “relevance” to users. Indeed, even in the 2001 RAE the definition of research adopted for all panels and units of assessment in RAE 2001 embraced the specific inclusion of “work of direct relevance to the needs of commerce and industry” (Baker and Gabbott, 2002, p. 6). The assessment of the value-added to professional practice by research is an issue that was identified as not being properly address in RAE 2001 and is therefore being taken into account in the next RAE (Higher Education and Research Opportunities in the United Kingdom, 2003, p. 23). This suggests that in the future we will be faced with the spectre of the “double hurdle” of academic rigor and practical relevance for research in marketing, as identified by Pettigrew (cited in Wensley, 2002, p. 394).

Conclusions So where does all that leave us? It is evident that there may be some practical steps that marketers can take in order to enhance the status of marketing. First, more effort should be directed towards securing funding to support both basic and practice-oriented knowledge production. Inevitably this demands that we make efforts to improve research support from bodies such as the ESRC and Leverhulme, in addition to obtaining financial support from more applied funding. This demands increasing the volume of applications and ensuring that they meet the criteria applied to assessing the “quality” of any research proposal: clear objectives, familiarity with the appropriate literature, the specification of the methodology and the means of analysis, some of which often lacks the attention to detail that encourages confidence in the proposers. In all cases there is a need for more experienced referees who will invest in providing detailed and productive feedback. Second, more attention should be given to the development of capabilities in research in marketing. There is a need to develop robust methodology training programmes both as an integral component of doctoral training and also for the marketing academy in general to ensure that all of use have the opportunity to familiarise ourselves with a range of methodologies.

The status and future of research 269

EJM 39,3/4

270

Third, effort needs to be invested in ensuring that the scholarship and research robustness of academic papers in marketing is developed through for instance, more careful attention to both then training of academic referees and the investment of time by referees in providing the positive feedback that is a hallmark of effective refereeing. The preparation of papers of substance often demands significant time and there is a need to campaign for more emphasis on quality as against quantity: more can indeed mean less! Fourth, we need to ensure that the issue of “users” is addressed by demonstrating how the results of the results will be disseminated. The “user community” embraces more than practitioners in business, commerce and the public sector; it also includes users of knowledge per se. Research may not be solely concerned or even involve the generation of solutions that have immediate application. It is unquestionably the case that “relevance” in some guise will be a pertinent criterion in the next RAE, but as we have argued there is no definitive interpretation of what this means. Overall, what is important is the “ quality” as reflected in the conceptual development and analysis; where appropriate, the robustness of the methodology employed; and the impact of the research on understanding. Good research is undertaken by enquiring minds: it is the curiosity of the researcher that produces the surprising result that has a considerable impact on our understanding. That rather than pointing to the “blindingly” obvious should be our goal. Here, however, we do offer some “blindingly obvious” practical steps that could have a positive impact on our discipline: . Improve our PhD programmes, seeking external recognition and verification of professional standards. . Publish more and better papers, (increased commitment to research). . Publish accessible versions of good papers with emphasis on managerial implications in practitioner journals. . Develop epistemology and research methodology workshops to facilitate the improvement of research standards in marketing. . Improve the methodological rigor of our work (Bessant et al., 2003). . Apply for more research funding. . Review each others’ work critically, helpfully and strategically. . Make nominations of marketing academics to the research council panels. . Network and mentor to the benefit of the discipline. References (The) Association of Business Schools (ABS) (2003), Pillars of the Economy: How UK Business Schools Are Meeting the Global Competitive Challenge, ABS, London. Baker, M.J. and Erdogan, B.Z. (2000), “Who we are and what we do – 2000”, Journal of Marketing Management, Vol. 16 No. 7, pp. 679-96. Baker, M.J. and Gabbott, M. (2002), “The assessment of research”, The International Journal of Management Education, Vol. 2 No. 3, pp. 3-15. Bessant, J., Birley, S., Cooper, C., Dawson, S., Gennard, J., Gardiner, M., Gray, A., Jones, P., Mayer, C., McGee, J., Pidd, M., Rowley, G., Saunders, J. and Stark, A. (2003), “The state of

the field in UK management research: reflections of the Research Assessment Exercises (RAE) Panel”, British Journal of Management, Vol. 14 No. 1, pp. 51-68. Burton, D. (2003), “Rethinking the UK system of doctoral training in marketing”, Journal of Marketing Management, Vol. 19 No. 7-8, pp. 883-904. Diamantopoulos, A., Schlegelmilch, B.B. and Neate-Stidson, S. (1992), “Who we are and what we do: a profile of marketing academics in UK universities”, Journal of Marketing Management, Vol. 8 No. 1, pp. 5-20. Dobson, M. (2003), “Investing in the future – ESRC funding opportunities”, paper presented at the Academy of Marketing Conference, Business School, Aston, Birmingham, July. Easton, G. and Easton, D.M. (2003), “Marketing journals and the research assessment exercise”, Journal of Marketing Management, Vol. 19 No. 1/2, pp. 5-24. Elias, P. and Purcell, K. (2003), “Too many graduates? The market disagrees”, THES, 20 June. ESRC (2003), The Economic and Social Research Council Annual Report and Accounts 2002-03, The Stationery Office, London. ESRC (2004), available at: www.esrc.ac.uk?esrccontent/researchfundinmg/aim_fellowship.asp (accessed 26 January). Fazackerley, A. and Goddard, A. (2003), “Price of policy put at £230 m”, THES, 28 November, p. 7. Gibbons, M., Limoges, C., Nowotony, H., Schwartzman, S., Scott, P. and Trow, M. (1994), The New Production of Knowledge: The Dynamics of Science and Research in Contemporary Societies, Sage Publications, London. Higher Education and Research Opportunities in the United Kingdom (2003), available at: www.hero.ac.uk/rae/ Learning and Teaching Support Network (2001), Learning and Teaching in Business, Management and Accounting: The UK Landscape, December, Business Education Support Team, York. Pringle, L.G. (2001), “The academy and the practice: in principle they are different but in practice, they never are”, Marketing Science, Vol. 20 No. 4, pp. 373-81. Ratchford, B.T. (2001), “Editorial”, Marketing Science, Vol. 20 No. 4, Fall, pp. iii-iv. Research Fortnight (2003), “Johnson moves top calm VC fears over research”, Vol. 9 No. 198, 17 September. Utley, A. (2003), “Social science failing its users”, THES, 28 March. Varadarajan, P.R. (2003), “Musings on relevance and rigor of scholarly research in marketing”, Journal of the Academy of Marketing Science, Vol. 31 No. 4, pp. 368-76. Wensley, R. (1995), “A critical review of research in marketing”, British Journal of Management, Vol. 6, December, pp. S63-S82. Wensley, R. (2000), “The MSI priorities: a critical view on researching firm performance, customer experience and marketing”, Journal of Marketing Management, Vol. 16 No. 1, pp. 11-27. Wensley, R. (2002), “Bridge over troubled water?”, European Journal of Marketing, Vol. 36 No. 3, pp. 391-400.

The status and future of research 271

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0309-0566.htm

EJM 39,3/4

Beyond incommensurability? Empirical expansion on diversity in research

272

Andrea Davies

Received January 2003 Revised September 2003 Accepted January 2004

Management Centre, University of Leicester, Leicester, UK, and

James A. Fitchett Nottingham University Business School, Nottingham, UK Abstract Purpose – This paper is a practical attempt to contribute to the ongoing reappraisal of the dichotomies and categories that have become prevalent throughout marketing research. Design/methodology/approach – The paper reviews current literature on incommensurability and undertakes a comparative re-examination of two studies. Findings – How the authors view their research is constituted in retrospective terms through a marketing and consumption logic based on the principles of division, distinction and difference. Re-examination of some empirical case material suggests that in practice the perceived duality separating research traditions is unsound. A misplaced reading of paradigm incommensurability has resulted in research practices appearing oppositional and static when they are essentially undifferentiated and dynamic. An over-socialised research epistemology has raised the tangible outcomes of research activities to be dominant in directing research practice. Research limitations/implications – The comparative analysis is illustrative rather than representative. Originality/value – The paper offers an applied exposition of theoretical debates in marketing research concerning paradigm incommensurability. Keywords Market research methods, History Paper type Conceptual paper

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 272-293 q Emerald Group Publishing Limited 0309-0566 DOI 10.1108/03090560510581773

Introduction Marketing and consumer research is once again experiencing a period of reappraisal, looking internally to question the direction and speed of its theoretical development and the jurisdiction of its theoretical boundaries (Levy, 2002; Brown, 2002; Zaltman, 2000; Lowenstein, 2001; Rossiter, 2001; Wells, 2001; Bettany, 2002). Marketing has been defined by its multi-disciplinary status for some time now and the processes of generating knowledge have been characterised by considerable methodological diversity (Wilk, 2001; Murray et al., 2002; Piercy, 2002). As with any field of academic enquiry the activities associated with theory making underpin disciplinary content and progress in marketing science (Gummesson, 2001; O’Shaughnessy and O’Shaughnessy, 2002; Thompson, 2002). However, the extent to which research communities in business research have a complete understanding of what constitutes and drives research practice is less clear and it may be either too much or too little disciplinary self-confidence that accounts for an unwillingness to examine the practices of knowledge production (Remenyi, 2002).

Paradigmatic debates were afforded a high profile in the marketing and consumer research literatures throughout the 1980s and 1990s. During this period questions were frequently asked regarding the authority of different research traditions with several ways to move forward and guide the research community being proposed (e.g. Anderson, 1983, 1986; Deshpande, 1983; Hirschman, 1985; Hirschman and Holbrook, 1992; Hudson and Ozanne, 1988; Ozanne and Hudson, 1989; Hunt, 1990, 1991, 1992, 1994; Sherry, 1991; Kavanagh, 1994). Critical pluralists (Hunt, 1991, 1994; Kavanagh, 1994) claiming openness and tolerance called for the status of qualitative methods to be repositioned in an effort to challenge what they perceived to be increasingly pervasive nihilistic and relativistic tendencies. These conciliatory efforts can perhaps be thought of as an attempt by influential groups within the academy at resolving, ending or offering a “last word” in debates over epistemology or so-called “epistobabble” warfare. Whether one agrees or disagrees with resolution couched solely in an objectivist and realist philosophy (Wilk, 2001), contemporary evidence would suggest that the concerns raised during this era of paradigm debate have not been “resolved” in the everyday practice and experience of researchers. Current debates concerning the legitimacy of different research approaches may be less likely to adopt the oppositional tone that characterised much of this earlier discussion but acknowledgement of divisions and boundaries between research programmes nevertheless continues to feature regularly in respected journals (Thompson, 2002; Wells, 2001; Shankar and Patterson, 2001; Szmigim and Foxhall, 2000; Hackley, 1999). The difficulties of “investigating”, “solving” or “resolving” issues of research diversity has the potential to deter researchers from undertaking empirical research, as well as deterring much needed discussion about the status of theory and knowledge generation and the relationship to research practice. As Wilk (2001) reminds us, pluralism demands that we accept the possibility that two studies or methods cannot be brought to bear on one another and that a single explanation is insufficient. The absence of agreed criteria that could be applied to adjudicate on the merits of different approaches to research investigation calls into question whether discussions about research practice or debates concerning the value of multi-method research constitute valid topics of enquiry. Evaluative criteria are bound up in the same social world in which the operative knowledge they are sought to bear on is generated (Schmidt, 2001) and this produces a situation of underdetermination (Anderson, 1986). In marketing and consumer research (Wilk, 2001) other social sciences (Sayers, 1987; Guba and Lincoln, 1998; Weick, 1989; Lewis and Grimes, 1999) and beyond (Nagel and Newman, 1971) the inability to step outside our own theoretical and epistemological languages is recognised as an impasse, or at the very least a stumbling block, that makes the study of diversity in research traditions fraught with anxiety. Since it may be impossible to “adequately ‘solve’ the problems of being, truth and subjectivity” (Flax, 1990, p. 193) we must perhaps be content to live in “the middle of things” (Wilk, 2001) and to accept the “messiness of this condition” (St Pierre, 1997). Our intention here is to offer another view on the diversity observed in our research traditions. From the outset it is necessary to draw a distinction between the appreciative and instrumental systems in research (Vickers, 1995), what we might summarise as the differences between the way the research process is viewed and theorised on the one hand and the way it is conducted or enacted on the other. The paper focuses on the manner in which marketing and consumer research knowledge is

Expansion on diversity in research 273

EJM 39,3/4

274

produced and represented through research practice to critique and challenge some misunderstandings that have now become entrenched. Marketing research remains constituted primarily in static and retrospective terms through a logic based on the principles of division, distinction and difference. Oppositional terms such as inside/outside, mind/matter, part/whole, self/other, academic/practitioner, qualitative/quantitative and positivist/interpretivist, are the direct products of such logic and have become the key conceptual categories that we rely on to express research experiences and research practices. Kuhn’s (1970a) incommensurability thesis is often drawn on to maintain and justify the integrity of such oppositional dualisms. In the following section we suggest that this thesis has been misunderstood in marketing. Building on this discussion we go on to consider two studies both of which examine a similar subject but are conducted in what would be widely defined as alternative paradigms. Our analysis offers a practical attempt to confront incommensurability by drawing on empirical case material. By critically contrasting the two studies we are able to examine the practice of interpretivism and the practice of positivism in the context of ongoing philosophy of science dualisms. The main outcome of this analysis is to contribute to the growing body of work in the management sciences that seeks to reappraise prevailing dichotomies and categories (e.g. Gioia and Pitre´, 1990; Schultz and Hatch, 1996; Lewis and Grimes, 1999). Incommensurability and beyond There is a widespread rejection of a unified meta-theoretical paradigm in marketing and consumer research (Anderson, 1983; Arndt, 1985; Mick, 1999; Zaltman, 2000). While the Kuhnian cycle of scientific revolution together with the thesis on incommensurability was described some time ago as a ”colourful exaggeration” (Venkatesh, 1985) and marketing was “not considered to have experienced robust alternative schools of marketing thought” (Savitt, 1990, p. 299), it remains a core concept in debates over paradigm credibility. At the workbench as it were, it seems that researchers are now being required to make an “either/or” choice between two seemingly polar paradigmatic opposites of research practice (Howe, 1988, 1992; Rose, 2001). In consumer research, for example, researchers choose between “positivist” and “interpretivist” (or “humanist”) approaches (Hudson and Ozanne, 1988; Wilk, 2001; Wells, 2001; Shankar and Patterson, 2001; Sherry, 2000; Brown, 2002; Thompson, 2002). Although strict conceptual justification for paradigmatic choice might be appealing in theory, inconsistencies begin to emerge when these principles are translated and applied to specific research practices and to the general social context of the academy in which research is undertaken. For example, differences in researchers’ personalities (Brown, 1999; Hirschman, 1985; Leong et al., 1994) and socialisation and cultural environments (Anderson, 1986; Hirschman and Holbrook, 1992; Trocchia and Berkowitz, 1999) have been shown to guide paradigmatic choices. Yet the spectre of incommensurability continues to restrict and restrain advances despite calls for multi-paradigm enquiry in the marketing discipline (Mick, 1999; Zaltman, 2000), social sciences (Schultz and Hatch, 1996; Lewis and Grimes, 1999) and by leading figures beyond the empiricist paradigm (Howe, 1992). This is all the more confounding when one considers the multi-disciplinary agenda of the subject, its focus on diversity (Ferber, 1977; Frank, 1974; Kavanagh, 1994; Kernan, 1995; Wilk, 2001) and history of

mixing, matching and importing ideas and techniques from many different areas (Tellis et al., 1999; Bettencourt and Houston, 2001). Authors in both “positivist” and “interpretivist” camps have focused on the value added to research designed with some form of triangulation (e.g. Arnould and Price, 1993; Deshpande, 1983; McQuarrie and Mick, 1992; Gummesson, 2001; McAlexander et al., 2002; Thompson and Troester, 2002). A problem common to such integrationist stances is that different research traditions are rarely given equal authority in the agenda, findings or conclusions drawn from empirical work (Hirschman, 1985). Indeed, examples of studies that have used a mixture of qualitative and quantitative methods, or cases where attention has been given to the triangulation of different kinds of data, have been criticised for relying on only one or a very narrowly defined set of philosophical and ontological beliefs (Wilk, 2001). Assigning equal value and weight to different traditions and methods of research is often considered problematic due to arguments regarding the apparent lack of common criteria to judge their quality (e.g. Szmigim and Foxhall, 2000; Thompson, 1990, 1993). The potential of realising a workable and acceptable context for multi-paradigm and inter-paradigm research remains uncertain. Theoretical frameworks for multiple paradigm research have been explored in some detail in management studies (Gioia and Pitre´, 1990; Schultz and Hatch, 1996; Lewis and Grimes, 1999). In contrast to the integrationist view common in marketing and consumer research, these frameworks contend that paradigm interplay produces new forms of understanding by stressing the interdependencies between constitutive oppositions. Their approach seeks to embrace disparity to build an enlarged and enlightened understanding of phenomena thus offering another perspective on paradigm reconciliation (Hirschman, 1985; Hirschman and Holbrook, 1992; Hunt, 1994; Hudson and Ozanne, 1988; Ozanne and Hudson, 1989; Szmigim and Foxhall, 2000). Although Schultz and Hatch (1996) offer a meta-paradigmatic position that denies incommensurability, their theoretical framework of meta-triangulation maintains rather than smoothes over notions of differences and boundaries that uphold divisions between different research traditions. This is exemplified by their use of “bridging” and “interplay” and does not seem adequately to attain a meta-paradigmatic level where different research traditions can co-exist (de Cock et al., 1995). Instead these frameworks operate in a grey area between paradigms where the potential to privilege one side of a dualism remains. Eisenhart (2000) views this middle ground as a way to avoid the blandness of synthesis and celebrates the co-existence of dualities. She calls for the need to uphold and accept paradox in practice and questions whether incommensurability is a state of nature. The importance of representing the interdependencies and commonalities among and between diverse research traditions does however remain unattended in theories that adhere to incommensurability as well as those based on meta-paradigmatic explanations. One of the problems researchers face when trying to resolve paradigm debates is the contradiction between the philosophy of research on the one hand, and the practice of research on the other (Kaplan, 1964; Lauden, 1984; Marsden, 2001; Wilk, 2001; Bouchikhi and Kimberley, 2001). The majority of researchers involved in the marketing and consumer research community would probably be able to locate their own and colleagues’ work within either a positivist or interpretivist tradition. The type and style of data collected and analysis employed, as well the type and range of

Expansion on diversity in research 275

EJM 39,3/4

276

research issues attended to, are most likely to be assumed to provide surrogate indicators that locate researchers in one tradition rather than another. However, while current debates about philosophy of science would seem to imply that positivists and interpretivists should be odds, in everyday practice this is not necessarily the case (Heath, 1992; Venkatesh, 1985). Researchers from different research traditions work alongside one another, discuss and argue common ideas and topics, attend similar conferences and publish in the same pool of journals. Furthermore, research conducted in one tradition often draws on work conducted in other traditions (Tellis et al., 1999; Bettencourt and Houston, 2001). There are of course disagreements and conflicts between researchers that should not be overlooked. However, these disagreements are arguably equally prevalent within the interpretivist and positivist traditions as they are between them. Research philosophy and research practice Accounts of the “muddling of method” (Baker et al., 1992; Goulding, 1999), the “stand between” metaphor as a description of research practice (Charmaz, 1995), and Heath’s (1992) boundary blurring description of consumer researchers as “liberal naturalists” or “conservative humanists” highlight the difficulty of linking paradigm theories to research practices. Examples of studies that have applied qualitative methods in a positivist tradition (Iacobucci, 2001) and quantitative methods in an interpretivist tradition (Potter and Wetherell, 1994) emphasise the widely accepted observation that there is a lack of correspondence between qualitative and quantitative methods of research (instruments) and the theoretical paradigms of positivism and humanism in which they are used (Lauden, 1984; Howe, 1988; Silverman, 1997; Riley and Love, 2000). Despite this it has become common to define paradigm debates using surrogate measures of methodology choice, research community conventions, power locales, data collection and analysis procedures, and accepted presentation protocols (e.g. Hunt, 1994; Shankar and Patterson, 2001; Brown et al., 2001). There is a real difficulty, even impossibility, of isolating one theoretical research tradition from another without drawing on more procedural aspects of research conduct. Research papers with recognised and accepted methods, procedures, and reporting conventions, are the tangible representations of theory generating research activity. Reward systems are based on journal publications and citation frequency, and the maturity of the discipline is analysed by the degree of diversity and exchange of reference citations (Zinkan et al., 1992; Bettencourt and Houston, 2001). We are the consumers of new ideas presented in our journal papers and their review and editorial practices guide what is right and good in the discipline. Studies of reference diversity in top journals show a negative relationship between subsequent article citation and the conceptual distance or disciplinary variety of citations contained within them (Bettencourt and Houston, 2001). As consumers of new research we favour theories that are readily understandable and easy to put into operation (Peter and Olson, 1983; Lynch, 1998). There is comfort in the certain and the familiar. As researchers we look for marketable findings, replicating the methodological and rhetorical strategies necessary to conform to the accepted practice of various sub-disciplines as they are constructed (Bouchikhi and Kimberley, 2001; Wells, 2001; Thompson, 2002). We persuade our audiences primarily through the utilisation of specific narrative forms (Van Mannen, 1995; de Cock and Chia, 2001), staging credible truth effects in a

progression of popularity contests (Sutton and Shaw, 1995; O’Shaughnessy, 1997; Shankar and Patterson, 2001). Schmidt (2001) recognises the cultural location of norms and standards held by our research traditions. He accounts for this “oversocialised epistemology” in terms of the “various parochialisms to which most academics (including himself) lean due to processes of deformation professionelle” (Schmidt, 2001, p. 148). It would seem research communities use Kuhn’s incommensurability thesis as an ontological act of differentiation to authenticate the credibility and status of sub-groups and sub-disciplinary areas, and research traditions become “traps invented by researchers to snare and dismember them” (Zaltman, 2000). Researchers focus on dissimilarities rather than similarities and reject the applicability of conceptually distant theories (Bettencourt and Houston, 2001). There is a drive to reduce epistemological uncertainty so as to establish credible research traditions. This has raised methodological and representational concerns to the status of first principle where the means through which marketing knowledge or science has become normalised have become ends in themselves (Haslam and McGarty, 2001). Research papers have become prized and valued over and above the research activities they describe. When valued only as research papers our research activities reified, becoming “thing-like”. One problem is that when viewed retrospectively, what is accepted in our journals becomes viewed as a universal way or theoretical map of what is right and good. This is at best a “counterfeit” account of research practice and disciplinary progression. As Rescher (1996, p. 35) explains “‘things’ [research papers] are no more than stability waves in a sea of process”. Journal papers cannot represent the full extent of research activity. They fail to capture the dialogue and exchange that is part of research practice. Paradigms revisited Kuhn’s (1968) original incommensurability thesis positions all research and knowledge within specific contexts of institutions, expectations, languages, instruments, and modes of understanding. These boundaries define and locate research paradigms. Hassard (1988) emphasises that paradigm change has been misinterpreted as a gradual transition where existing paradigms are seen as nurturing, growing, sitting alongside, and finally being subsumed to new paradigms. For Kuhn (1970a, p. 85), however, paradigm shifts are points of fracture and transformation from one form of “normal” science to another. Such a reading does not adequately reflect the current status of different research communities nor the relationship between interpretivists and positivists for disciplines that see themselves as pluri-paradigmatic and hold constant two (Hudson and Ozanne, 1988; Szmigim and Foxhall, 2000), or more “paradigms” (Hirschman and Holbrook, 1992; Sherry, 1991). Other recent re-examinations of Kuhn’s understanding of paradigms and paradigm revolutions highlight the logical and discursive inconsistency between the idea of scientific progress on the one hand, and paradigmatic conflict or rivalry leading to incommensurability on the other. Conflict and rivalry require at least some level of translation, common location and reference or meaning, whereas theories expressed in alternative incommensurable paradigmatic languages can have no common meaning and comparable content. Davidson (1984) and Putman (1981) show that it is incoherent to suggest theoretical languages are untranslatable, unless the content of theories in

Expansion on diversity in research 277

EJM 39,3/4

278

one paradigm are received as “farm yard noises” with no meaningful expression in other paradigms. Indeed, both Kuhn and Feyerabend did not adhere to an indeterminable notion of semantic incommensurability, but rather to one of “local incommensurability” (Kuhn, 1983). Sankey (1994) also rejects a reading of “paradigm” based on discontinuous transition between theories which have no common reference or which refer to distinct worlds of their own making. Instead he considers there to be a referential continuity where supposedly divergent paradigmatic theories can be compared on content because they only have “divergent referential relations to a common theory-dependent reality” (Sankey, 1994, p. 219). Drawing on Feyerabend (1965) he suggests that translation, at least at the level of non-instantial theory or common descriptive concepts can occur due to the presence of background theories that are in some way generically common and therefore applicable in more than one paradigm. Evidence of semantic understanding and exchange between seemingly alternative paradigmatic traditions would suggest that current paradigmatic boundaries are at least permeable or, perhaps more dramatically, that boundaries and divisions have been drawn pre-maturely and do not reflect substantially different content. O’Shaughnessy (1997) suggests that there may be little substance in the differences identified in marketing and consumer research traditions and that we are subject to the “granfalloon effect”. We believe and act as though we have something in common because we can be, and need to be, identified with a legitimate research tradition. In a similar vein, Wilk (2001) argues that knowledge of any sort is fundamentally built on comparison. He suggests all research papers are in fact re-studies of what has gone before. It would seem that we have created difference as a disciplinary legitimising tactic and this limits our ability to foster discontinuous innovation (Zaltman, 2000; Feyerabend, 1965). The absence of a unified or authoritative paradigmatic position and the subsequent lack of a commonly agreed consensus to define mechanisms of evaluation, has tended to over-emphasise and inflate differences in terms of conflict, and in some cases as “paradigm warfare” (Anderson, 1983, 1986; Hunt, 1990, 1991, 1992; Kavanagh, 1994). As a consequence, evidence of common language (Putman, 1981) or shared cognitive goals (Howe, 1992) is under represented. Methodological approach The incommensurability thesis is a difficult one to examine if paradigmatic boundaries are considered to be fixed or partial incommensurable. Debates surrounding incommensurability question the very assumptions of scientific examination. To seek and “verify” the thesis would from a certain paradigmatic position seem acceptable and necessary, whereas from others such efforts are inappropriate, even irrelevant. In terms of a method we adopt a critically-reflexive re-inquiry (Thompson, 2002). The primary methodological motivation is to address the issue of incommensurability in two research traditions by employing a data-driven agenda. The lack of broadly comparable case studies undertaken in similar research contexts, but from alternate paradigmatic traditions, has hindered the exploration of paradigm commensurability as practised, as opposed to how the debate is conceptualised. The aim of the analysis is to illustrate the contrasts evident between studies with different so called paradigmatic bases, to compare respective assumptions and justifications, and to identify agreement and disagreement in the main findings. Here we shift our attention from the “reality”

being presented in studies from different research to the representation of those realities in each tradition. This enquiry draws on two empirical studies that examine consumption experience and consumer behaviour in the leisure industry, and in particular the museum “sector”. The two studies were conducted independently of each other, with neither research programme or research team involved with the other. Only following the completion and peer review of both studies did the researchers become aware of the other piece of research. Study A was undertaken in accordance with what would be identified as a positivist approach, whereas Study B was undertaken in the interpretivist tradition. Study A takes the traditional psychological/ behaviourist paradigm in consumer research as its basis, applying the Theory of Planned Behaviour to compare several predictive models of museum visiting behaviour, based on a stratified spatial random sample of 400 individuals, and using a two-stage structural equation modelling analysis technique. The intangible experience-outcomes shared by museum visitors, and disliked or undervalued by museum non-visitors, are offered as the basis for enlarging the explanation of museum visiting and non-visiting intentions by improved descriptive depth and predictive validity. Study B draws heavily on Marxist and Post-Marxist critical theory to examine whether the museum experience is organised in terms of a commodity code, and visiting a museum as an act of consumption. Study B offers a cultural/macro explanation of museum visiting, identifying the cultural processes that have enabled aspects of history and culture (museum exhibitions) to become represented and consumed as commodities. Based on a naturalistic/ethnographic methodology where 40 qualitative interviews are supplemented with photography to ensure sensitivity to the natural setting, the design of the study aims to maximise thickness and depth and stresses the authority of the author in developing/presenting the description and interpretation the museum experience. Both studies justify their chosen research philosophies, and conduct, analyse and report the research process and their findings in accordance with the principles ascribed by their respective methodologies. Study A presents and verifies data by drawing on conventions of quantitative modelling. Study B relies on textual data and description. Study A maintains a focus on individual cognitive processes to explain museum visiting, whereas Study B focuses exclusively on the explication of socio-structural codes to describe and understand the experience. A summary of the main conceptual, methodological and analytical components of both studies is given in Table I.

Qualitative and descriptive contrasts To examine the two pieces of research we considered the textual structure and constituency of both studies using a literary critical, or discourse analytical approach. This level of analysis calls for the two studies to be examined as texts written in accordance with certain protocols, languages and traditions. Paradigms have been seen as differentiated by conventions of discourse and modes of representation making communication between paradigms appear unfeasible. While we focus part of this analysis on the findings and data generated by the two pieces of research, we also take the research texts themselves as units for analysis. This analysis considers the

Expansion on diversity in research 279

EJM 39,3/4 Principal literature

280

Conceptual framework

Methodological argument Research design

Structure of main findings

Table I. Composition of Study A and Study B

Study A: marketing goes to the museum

Study B: consumption and cultural commodification

Experiential consumption and museum visitor research Review of motivation theory in social psychology, leisure and tourism studies Review theoretical and methodological concerns for Theory of Planned Behaviour Explain visitor and non-visitor intentions by anticipated experiences moderated by value and museum interpretive style (service environment), skills and resources and social norms Methodological and conceptual modification to the Theory of Planned Behaviour Justification of a modelling approach Qualitative variable definition Quantitative data collection Two-stage structural modelling In-depth interviewing Structured survey questionnaire Random-spatial sample Scenario method with two contrasting pictorial stimuli Patterns of museum behaviour Identification and description of structural model Testing predictive and descriptive utility of alternative models of Theory of Planned Behavior Identification of principal determinants of visiting behaviour

Anthropology and material culture theory of gift/commodity exchange Review of Marxist critical theory and sociological research on commodity culture Review of post-Marxist critique, postmodernism and theories of consumer culture The museum/collection in pre-industrial, industrial and post-industrial society The consumer “experience” Typology of consumer value

Justification of an interpretivist approach Qualitative data collection and analysis methods In-depth interviewing Ethnographic approaches to consumer research Description of the research locations The value of museum visiting Examination of sign value in the museum Commodification of museum exhibits Experiences of authenticity and proximity Narrative construction

discursive framing, modes of justification and argument, styles and language of representation and the underlying assumptions and rationales for each text. In accordance with discourse analysis a close reading of the two texts was used to identify common “codes” or tropes (Elliott, 1996). The approach involved each author preparing a close reading of the other text, that is, the author of Study A began analysing Study B, and vice versa. A series of panel sessions were organised whereby these initial analyses were discussed, with both differences and similarities being confronted and further examined. The first theme discussed below refers to the presence of a “marketing zeitgeist”. It can be read as a “common world view” or shared set of underlying assumptions. The second theme explores the basis of the research assumptions and the manner in which they are constructed, justified and represented in the texts. The third theme goes on to examine the some of the principle research findings developed in both studies.

A marketing zeitgeist When analysed as a socially and historically located phenomenon, consumer research like all modes of enquiry has common traits or dominant themes (Kernan, 1995). Knowledge is emergent and organic rather than explicitly planned or chosen by any one group or institution (Kuhn, 1970b). That two researchers decided to consider the identical subject of museum visiting consumption reveals a great deal about the zeitgeist or underlying state of ideas and theories in marketing at the time. The following excerpts illustrate the extent to which both studies represent and respond to these prevailing: Study A: In an over-supplied heritage attractions market . . . and against a background of increasingly demanding and discerning consumers bent on avid “experience” seeking of the past, present and future . . . it has now become a matter of urgency for museum professionals to understand the determinants of museum visiting intentions, customer satisfaction and the potentials for market development . . . Studies from diverse literatures beyond heritage research have been relatively more substantial in their concern with consumption experiences . . . An examination of studies from [consumer research] offer[s] potential classifications of experience useful to understand the possible “nature” of the museum experience, and to forward the conceptual development of a research agenda focused on consumer and non-consumer experiences of heritage consumption . . . The overall aim of this paper is to contribute to the understanding and knowledge of museum consumption behaviour framed within an experiential management approach. Study B: The museum profession, like many other “culture industries”, has had to respond to a changing environment. The structure of public funding together with the limitations placed on national and regional government expenditure has meant that many museums have been forced to reconsider their priorities and strategies for continued survival . . . As a consequence, the role of the visitor has undergone somewhat of a metamorphosis. As museums become organised like businesses, visitors become increasingly redefined as consumers. The problem however, which is central to this investigation, is what exactly constitutes the museum commodity? If museum visitors are consumers, then what do they consume? The museum, as a site of consumption, presents several problems in terms of existing theories of consumer behaviour, namely the lack of any economic exchange, material acquisition or functional utility relating to “needs”. The primary objective of this empirical investigation is to examine whether a semiotic, cultural theory of consumption can be applied to understand museum consumption.

These excerpts illustrate that there was a motivation to question traditional boundaries around the discipline that have tended to locate consumer research firmly in the sphere of commercial product-based contexts (e.g. Kotler, 1979; Hirschman and Holbrook, 1982; Holbrook and Schindler, 1989). A number of opportunities were emerging in which empirically to examine further increasingly popular conceptual ideas concerning the consumption experience, and tourism and leisure activities as marketing-related processes (Arnould and Price, 1993; Celsi et al., 1993; Hirschman and Holbrook, 1992; McCracken, 1988). During this period, popular ideas relating to postmodernism, consumer culture, and consumption symbolism (Belk, 1988; Brown, 1995; Firat and Venkatesh, 1995) were also being given empirical attention. These trends in the subject demonstrate the importance of disciplinary socialisation. Despite seemingly incommensurable methodologies, both studies share and build on a

Expansion on diversity in research 281

EJM 39,3/4

282

common legacy within the field of consumer research. Issues and topics popular in the discipline do not only determine these underlying areas of similarity, but also the location where the research was conducted, the resources available to support the research, and the specific social and cultural community where the research was undertaken. The broader social and cultural location of the researchers should not be ignored. Both studies were completed in marketing faculties of business schools in Scotland (UK) in the late 1990s. The prevailing cultural conditions in Scotland at the time focussed heavily on questions of Scottish identity, Scottish (as opposed to English) culture, and Scottish institutions, with museums playing a highly visible role in debates surrounding the celebration and representation of Scottish culture. When looked at in relational terms, the salience of Scottish identity issues and the role of museums in representing these issues have influenced project and issue formulation. The role of marketing in promoting and encouraging public support for these institutions of culture, as well as the generic trend towards recreating museums as visitor (consumer) attractions for the increasingly significant tourism and leisure industry in Scotland provides some explanation for why both studies selected a common context in which to conduct the empirical aspects of their research programmes. While both studies can be seen to address a shared basic question or “problem”, the form that this question takes in the studies differs. Study A presents no theoretical justification for the observed shift in consumer research. That other non-commercial settings have been peer reviewed as credible sites for consumer enquiry is a sufficient justification for the museum as a location of research. Unlike Study A, Study B is not concerned with issues that demand explanative, or definitive styles of investigation but rather with the reasons why this set of conditions emerged is a given cultural-historical context. This reveals that debates about commensurability become complex and somewhat open to negotiation when applied empirically. Both studies can be seen to share underlying themes and motivations despite clear paradigmatic differences in terms of their respective theoretical expressions and cognitive aims (Howe, 1992; Sankey, 1994). The prevailing zeitgeist in both the discipline and the social environment provides some kind of meta-paradigmatic context that simultaneously enables alternative paradigmatically bound explanations to be pursued. Research assumptions The two studies have distinct research assumptions that conform to their respective methodological positions and this provides an opportunity to apply a conventional incommensurability thesis. Study A explains consumption at the individual level of behaviour that is influenced by various environmental stimuli. Study B is not concerned with the individual level but rather with the macro, or cultural conditions that have enabled contemporary consumption (represented by the proliferation of commodities) to emerge. The experience, behaviour and beliefs of consumers are taken only as illustrative of this general condition. While Study A is concerned with consumers and their behaviour, Study B is concerned with consumption and consumer society. Claims that the two studies should be considered commensurable because they both investigate the general topic of “consumption” in a museum setting could be questioned since both take a different view on what constitutes consumption and

consumption-related behaviour. However, it is not that Study B rejects or disagrees with the assumption that consumption can be understood as a behavioural condition, or that Study A attempts to refute a cultural critique of consumer society. Rather that the discourses used to develop Study B do not require behavioural readings, and the discourses used to develop Study A provide no space to account for cultural critiques other than as marginal behavioural determinants. However, collective insights can be gained from reading both studies together despite the lack of commonly informed research assumptions, as our comparative assessment of findings in the next section illustrates. Both studies apply a schema or theory to put into operation their respective approaches to the study of museum consumption, although this is achieved in subtly different ways. In common with the hypothetic-deductive approach, Study A is largely predetermined by the Theory of Planned Behaviour (Ajzen, 1991). Moderation of the theoretical model is justified in the literature review and then tested in the study. Study A rejects the conventional method of measuring antecedent attitudinal, subjective norm and control beliefs to suggest a revision based on disaggregation (Bagozzi, 1985) and a modelling procedure previously unused with this schema. The schematic classification used in Study B is not represented as a model, but rather as a simplified, and logical reduction of theories of commodity value. Whereas Study A uses behavioural antecedents as principal conditions for its model (classification), Study B draws on linguistic, or structuralist antecedents for its categorical description. While these schematic structures are not directly commensurable both seek to confirm a schematic premise that can be comprehended from each paradigmatic position. Both studies develop their arguments in a way that is appropriate to their respective root paradigms. For Study B, which locates itself in an interpretivist perspective, theory evaluation is implicit and not stressed. Study A, which associates itself with a positivist perspective, explicitly recognises model testing as an important part of the research design. Comparative assessment of findings Study A reports its findings with little concern for the cultural, historical or social conditions of consumer behaviour, and Study B offers little behavioural explanation for respondents’ actions. The coherence and rigour demanded of all research, whether interpretivist or positivist, by necessity leads to specialised modes of enquiry. The findings evident in both studies, whilst leading to somewhat disparate ends can, however, be seen to share some common traits. The main areas of commonality relate to visitors’ perceptions of authenticity, the issue of active participation and construction of the museum experience, and visitors’ desire to maintain a degree of personal control in the museum experience. Both studies identify consumer research as having traditionally viewed consumption as a passive activity, with the main focus being on the object of consumption and the manner in which it is provided or supplied. The two studies use empirical evidence to argue that consumption in a museum context is a mutually constructive process involving active interpretation. They both further ideas of the consumer as producer (or co-producer) of the consumption experience (Firat and Dholakia, 1998). The excerpts given below illustrate differences in discourse (or syntax) but commonality in terms of meaning (semantics):

Expansion on diversity in research 283

EJM 39,3/4

284

Study A: “Imagineering” or the professional dreaming-up of three dimensional fantasies which are planned and reconstructed for the total experience has become common place in our developing urban landscape . . . [and] characteristic of postmodern society as hyperreality. In terms of perceived positive reactions to the museum exhibitions, the idea and object based museums were found to significantly diverge in terms of hedonic cognitions and wakeful daydreaming, . . . personal and non-personal imagery (Hirschman and Holbrook, 1982). In particular, strong differences were noted for “feeling the past is brought to life” (p , 0:000, Cramer’s V ¼ 0:38) and “creating images in your mind of how and who used the objects” (p , 0:000, Cramer’s V ¼ 0:36). For instance, 80.0 per cent of idea-based museum respondents said it was extremely or quite likely that the “past would be brought to life” (p , 0:000, Cramer’s V ¼ 0:38), compared to only 46.5 per cent of object-based museum respondents. Similarly more respondents in the idea-based museum thought it was extremely or quite likely (77.0 per cent) that this would “create images in their minds of how and who used the objects’, compared to 46.0 per cent respondents in the object-based museum sample (p , 0:000, Cramer’s V ¼ 0:36) . . . These findings confirm that the idea-based museum is far more effective at facilitating respondents’ imaginations. “Imagineering”, has been shown as a positive and enabling change agent rather than an inability to come to terms with the past with its associated experiences of nostalgia and escapism . . . The fourth dimension extracted from principal component analysis (to describe the structure of museum experience) represents the experiential cognitions described by Hirschman (1985). Variables which loaded significantly on this dimension included “feeling a connection with the past”, “imagining what your life would have been like living in the past”, “imagining who used and how the objects were used” and “feeling admiration for the skill and craftsmanship of people in the past”. Study B: When visitors come into contact with museum artefacts, they use this experience to make stories about the past. In some cases this narrative construction was highly personal, relating to aspects of the visitor’s own past. This was particularly evident in the Museum of Transport with visitors who could remember the exhibits when they functioned as vehicles of public transport. In producing these stories about the past, visitors used the artefacts to structure their narratives, as if by being close to objects, the memories and stories about the past were more visual and more real and therefore more useful for visitors when constructing their own narratives . . . Museum artefacts have a unique use value for visitors in the sense that no other form of presentation can provide this insight into the past or provide the components for this type of narrative construction. Visitors did not only construct stories around objects they had personal experience of but also with objects from periods beyond their own life time or from other cultural spaces.

These findings present both studies with the opportunity to develop and progress their respective theoretical starting points although for very different reasons. Study A uses the finding to suggest that behavioural assumptions need to accommodate a greater level of individual action which it has tended to lack, while Study B uses a similar finding to propose that commodity theory (which represents the consumer as being alienated from the productive sphere) needs to be modified to incorporate an active or interpretative component. Common or broadly comparable (commensurable) findings are represented through different syntactic conventions, and are used to substantiate different types of arguments and interpretations. The use of numerical and statistical information in Study A is the clearest example of differing conventions (Potter and Wetherell, 1994).

A joint reading of the two pieces of research explicates a basic deterministic assumption running throughout seemingly incommensurable theories of consumption. Study A considers consumer behaviour to be determined by sets of cognitive antecedents, whereas Study B considers consumption to be structurally determined. Both studies argue for a greater recognition of individual facticity in value creation than is generally permitted by either behaviourist or cultural theories. They view the agency of the consumer as constrained although present. Consumption experiences in which the consumer is permitted to take a significant degree of control are perceived to be of greater value and significance than those where control is restricted or maintained by some other group, for example by museum curators: Study A: The largest differences in perceptions of the two museums were related to the interpretative environment, and in particular, the ability of visitors to engage with the objects displayed and to interact with interpretative media. “Being able to touch objects” (p , 0:000, Cramer’s V ¼ 0:49), “using computers” (p , 0:000, Cramer’s V ¼ 0:43) and “using models” (p , :000, Cramer’s V ¼ 0:41) were all more strongly observed in the idea-based museum. For example, 43.0 per cent of respondents rated “being able to touch objects” as extremely likely in the idea-based museum compared to only 8.5 per cent of respondents who found this to be the case for the object-based museum. Study B: The reality principle is a compromise. On one level a video presentation offers greater potential in terms of an accurate depiction of history because it can show far more than the objects alone. It can depict the artefact in use, it can show how the object was integrated into everyday social life and thus represent history more authentically. However, the level of abstraction that such a presentation imposes is considered to be limiting and verging on simulation. It adds a further barrier between the visitor and the already vague and disappearing past. Since close proximity to objects is of principle concern to museum visitors, the video, film or picture media loses its authenticity because of the distance it imposes between artefact and the object-desiring visitor. The object display, as seen in the museum, allows this disadvantage to be overcome – but at a price . . . The visitors’ choice is unanimous. The object discourse is given principle value and the visitor is prepared to ignore and even fail to recognise, the obvious disadvantages.

The influence of post-modern thought, questions over authentic representation, hyper-reality, and the primacy of the image over referents are evidenced in both studies. The then contemporary debate in the museum profession over the importance of the “authentic” and criticisms about the ”disneyfication” of museum interpretation (Cannon-Brookes, 1991; Terrell, 1991) emerge as common traits. Both studies focus on issues of authenticity as a vehicle to contribute to this debate. Rather than being an externally verifiable construct, authenticity is argued to be a subjectively constructed condition. Even when museum exhibits were clearly not “real” in the sense that they were not original or actual examples of the objects being represented, consumers were able to find authenticity in the “fake”. Reality emerges from the experience of being in the position to experience rather than being seen as located in the object presented. This shared interpretation is the strongest evidence against the incommensurability thesis when applied to these studies. From an empirical lived experience of research, the incommensurability thesis is very difficult to sustain in a clear and simple fashion. It is as if incommensurability is itself inappropriate simply because it negates and

Expansion on diversity in research 285

EJM 39,3/4

286

restrains the commonalities between different paradigmatic schools in terms of research practice.

Concluding remarks Adherence to strict or partial incommensurability is shown to be inappropriate in the context of evaluating the contributions provided by these two studies. Our analysis has clearly demonstrated the need to consider alternative epistemological foundations in consumer research but rejects viewing different paradigms as fixed choice alternatives. This analysis has shown that the post-Kuhnian critique of incommensurability is a viable proposition in terms of research practice. It raises a number of critical questions regarding the theoretical basis and pragmatic incorporation of research findings into the marketing and consumer research canon. The fact that marketing and consumer research, now has a breadth that spans a whole spectrum of research identities, including everything from technical and scientific explications, to studies in the arts and literature of consumption, suggests that a more fluid understanding of paradigms is desirable. Far from representing impermeable worldviews or a lack of relational language or common signification, the analysis of these two studies clearly shows “evidence” of commonality, points of shared interest, and mutual experience. The lasting legacy of the incommensurability debate is that it draws attention to, and demarks difference as a core feature of all disciplinary identity. Differences can be the source of conflict and power but also enhanced understanding. An uncritical adherence to the concept of incommensurability can be seen as placing somewhat artificial barriers around the exchange and discussion of research from studies in other paradigmatic camps. A generation of researchers have been trained to conceptualise research issues in terms of paradigmatic boundaries and to underplay their permeability and interrelation. Marketing and consumer research has been subject to what Giddens (1984) terms “opposing error” that has produced state of disunifying dualism. This “error” is manifest in theoretical and methodological oppositions and in terms practice (experience). This paper has shown that adhering to a misplaced reading of paradigm incommensurability in a search to establish credible research identities has made research practices appear oppositional and static when they are essentially undifferentiated and dynamic. The need for paradigmatic consensus as a process of identity validation for our research traditions has seen our methods, procedures, modes of representation, and journal editorial decisions become the dominant source of reason, which then serves to direct research practice (Haslam and McGarty, 2001). If the procedural aspects and representation of research activities account for most or all of the discrepancies between different types of research then there may be little or no need to reify the higher order construct of paradigm incommensurability. If this is the case, locating disagreement and conflict at the inaccessible and non-observable level of paradigms has the potential to stifle initiatives that seek to bridge the gap between specific aspects of research practice. Legislating methodological concerns to the status of first principle where the means through which marketing knowledge or science has become normalised have become ends in themselves and reveals a discipline moving towards homogeneity and standardisation which is paradoxical to the scientific values our discipline has sought to follow.

Our discipline suffers from false confidence (O’Shaughnessy, 1997). We may recognise that research tools and our representation of the procedural aspects of our research endeavours rarely reflect research experiences, but we comply with these conventions as a matter of expediency to obtain journal article acceptance (O’Shaughnessy, 1997; Zaltman, 2000). Compliance to these now established academic conventions, effectively a privileged methodological and representational code, may relegate and ultimately divorce the experience of research practice from its subsequent publication. The production of research comes to rely more heavily on the conventions of style and adherence to a kind of hyper reality of research expectations. The relevance and presence of good research will inevitably suffer, being indistinguishable from a body research that is dislocated and lacking in any sense of reference to the experiences of either the researching or researched community. This trend is of course evident is other social contexts and well reported. As signs lose their referents to reality they eventually become self-referential resulting in a hyper real image that has no value or purpose other than within the boundaries of the simulation itself (Baudrillard, 1981). The debate between interpretivism and positivism for instance often seems to take on a quality of hyper reality, being relevant only to and between itself. To others outside of the narrow community of scholars who engage in the debate the outcomes of such arguments often seem bewildering and intelligible, not to say irrelevant and esoteric. Kavanagh’s (1994) review of the then ongoing sparring match between Shelby Hunt and Paul Anderson and together with contributions from their respective advocates illustrates that such hyper real wrangling only serves to trivialise research processes and alienate a broader readership from engaging with the often genuine and positive insights that emerge from academic research practices, irrespective of the paradigmatic approach taken. We are in serious danger of marginalising potential contribution and diversity both among academic sub-communities and between academics and industry practitioners (Elliott and Jankel-Elliott, 2003; Wotton et al., 2002; Marsden, 2001; Hackley, 1999; Sankey, 1994). Other research traditions that are not burdened by these debates, such as commercial consumer research for example, will take an increasingly visible and powerful position in which principles of reliability and reflexivity are marginalised. Adopting a pragmatic position, although one that is progressive and non-confrontational, Weick (1999) recommends researchers to “drop their heavy tools of paradigms”, which are only relevant to their own modes of representation and constitute a largely internal and academic debate. Common resources, institutions, experiences, distance and vocabulary are shared resources within the marketing and consumer academic research community that should not be underestimated.

References Ajzen, I. (1991), “The theory of planned behaviour”, Organizational Behaviour and Human Decision Processes, Vol. 50, pp. 179-211. Anderson, P.F. (1983), “Marketing, scientific progress and scientific method”, Journal of Marketing, Vol. 47, Fall, pp. 18-31. Anderson, P.F. (1986), “On method in consumer research: a critical relativist perspective”, Journal of Consumer Research, Vol. 13, September, pp. 155-73.

Expansion on diversity in research 287

EJM 39,3/4

288

Arndt, J. (1985), “The tyranny of paradigms: the case for paradigmatic pluralism in marketing”, in Dholakia, N. and Ardnt, J. (Eds), Research in Marketing: Changing the Course of Marketing, JAI Press, London, pp. 3-26. Arnould, E.J. and Price, L.L. (1993), “River magic: extraordinary experience and the extended service encounter”, Journal of Consumer Research, Vol. 20, June, pp. 636-41. Bagozzi, R.P. (1985), “Expectancy-value attitude models: an analysis of critical theoretical issues”, International Journal of Research in Marketing, Vol. 2 No. 1, pp. 43-60. Baker, C., West, J. and Stern, P.N. (1992), “Method slurring: grounded theory/phenomenology example”, Journal of Advanced Nursing Studies, Vol. 17 No. 11, pp. 1355-60. Baudrillard, J. (1981), For a Critique of the Political Economy of the Sign, Telos Press, New York, NY. Belk, R.W. (1988), “Possessions and the extended self”, Journal of Consumer Research, Vol. 15, September, pp. 139-67. Bettany, S. (2002), “Conceptualising marketing theory as an empirical domain: a speculative taxonomy of research approaches to the practice of marketing theory and suggestions for feminist ethnographic interventions”, Proceedings of the Academy of Marketing Conference 2002, Nottingham. Bettencourt, L.A. and Houston, M.B. (2001), “Reference diversity in JCR, JM and JMR: a reexamination and extension of Tellis, Chandy, and Ackerman (1999)”, Journal of Consumer Research, Vol. 28, September, pp. 313-23. Bouchikhi, H. and Kimberley, J.R. (2001), “‘It’s difficult to innovate’: the death of the tenured professor and the birth of the knowledge entrepreneur”, Human Relations, Vol. 55 No. 1, pp. 77-84. Brown, S. (1995), Postmodern Marketing, Routledge, London. Brown, S. (1999), ‘“Marketing and literature’: the anxiety of academic influence”, Journal of Marketing, Vol. 63, January, pp. 1-15. Brown, S. (2002), “Vote, vote, vote for Philip Kotler”, European Journal of Marketing, Vol. 36 No. 3, pp. 313-24. Brown, S., Hirschman, E.C. and McLaren, P. (2001), “Always historicise! Researching marketing theory in a post-historical epoch”, Marketing Theory, Vol. 1 No. 1, pp. 49-89. Cannon-Brookes, P. (1991), “Museums, theme parks and heritage experiences”, Museum Management and Curatorship, Vol. 10, pp. 351-8. Celsi, R.L., Rose, R.L. and Leigh, T.W. (1993), “An exploration of high-risk leisure consumption through skydiving”, Journal of Consumer Research, Vol. 20 No. 1, pp. 1-23. Charmaz, K. (1995), “Between positivism and postmodernism: implications for methods”, Studies in Symbolic Interaction, Vol. 17, pp. 43-72. Davidson, D. (1984), Inquiries into Truth and Interpretation, Clarendon Press, Oxford, pp. 183-98. de Cock, C. and Chia, R. (2001), “Diversity and organization: an essay with special reference to the work of Mikhail Bakhtin”, Discussion Papers in Management, Exeter University, Exeter. de Cock, C., Richards, T., Weaver, G.R. and Gioia, D.A. (1995), “A rejoinder and reply from Weaver and Gioia”, Organization Studies, Vol. 16, pp. 669-76. Deshpande, R. (1983), “Paradigms lost: on theory and method in research in marketing”, Journal of Marketing, Vol. 47 No. Fall, pp. 101-10. Eisenhart, K.M. (2000), “Paradox, spirals and ambivalence: the new language or change and pluralism?”, Academy of Management Review, Vol. 25 No. 4, p. 703.

Elliott, R. (1996), “Discourse analysis: exploring action, function and conflict in social texts”, Marketing Intelligence & Planning, Vol. 14 No. 6, pp. 65-8. Elliott, R. and Jankel-Elliott, N. (2003), “Using ethnography in strategic consumer research”, Qualitative Market Research, Vol. 6 No. 4, pp. 215-23. Ferber, R. (1977), “Can consumer research be interdisciplinary?”, Journal of Consumer Research, Vol. 4, June, pp. 19-28. Feyerabend, P.K. (1965), in Colodny, R.G. (Ed.), Beyond the Edge of Certainty, Prentice-Hall, Englewood Cliffs, NJ, pp. 145-260. Firat, A.F. and Dholakia, N. (1998), Consuming People: From Political Economy to Theatres of Consumption, Routledge, London. Firat, A.F. and Venkatesh, A. (1995), “Liberatory postmodernism and the reenchantment of consumption”, Journal of Consumer Research, Vol. 22, December, pp. 239-67. Flax, J. (1990), Thinking Fragments: Psychoanalysis, Feminism, and Postmodernism in the Contemporary West, University of California Press, Berkeley, CA. Frank, R.E. (1974), “The Journal of Consumer Research: an introduction”, Journal of Consumer Research, Vol. 1, June, pp. iv-vi. Giddens, A. (1984), Constitution of Society, Polity Press, Oxford. Gioia, D. and Pitre´, E. (1990), “Multiparadigm perspectives on theory building”, Academy of Management Review, Vol. 15 No. 4, pp. 584-602. Goulding, C. (1999), “Consumer research, interpretive paradigms and methodological ambiguities”, European Journal of Marketing, Vol. 33 No. 9/10, pp. 859-73. Guba, E. and Lincoln, Y.S. (1998), Paradigm Dialogue, Sage, Newbury Park, CA. Gummesson, E. (2001), “Are current approaches to marketing leading us astray?”, Marketing Theory, Vol. 1 No. 1, pp. 17-48. Hackley, C.E. (1999), “Tacit knowledge and the epistemology of expertise in strategic marketing management”, European Journal of Marketing, Vol. 33 No. 7/8, pp. 720-35. Haslam, S.A. and McGarty, C. (2001), “100 years of certitude? Social psychology, the experimental method and the management of scientific uncertainty”, British Journal of Social Psychology, Vol. 40 No. 1, pp. 1-21. Hassard, J. (1988), “Overcoming hermeticism in organization theory: an alternative to paradigm incommensurability”, Human Relations, Vol. 41 No. 3, pp. 247-59. Heath, T.B. (1992), “The reconciliation of humanism and positivism in the practice of consumer research: a view from the trenches”, Journal of the Academy of Marketing Science, Vol. 20 No. 2, pp. 107-18. Hirschman, E.C. (1985), “Scientific style and the conduct of consumer research”, Journal of Consumer Research, Vol. 12, September, pp. 225-39. Hirschman, E.C. and Holbrook, M.B. (1982), “The experiential aspects of consumption: consumer fantasies, feelings and fun”, Journal of Consumer Research, Vol. 9, September, pp. 132-40. Hirschman, E.C. and Holbrook, M.B. (1992), Postmodern Consumer Research: The Study of Consumption as Text, Sage, Newbury Park, CA. Holbrook, M.B. and Schindler, R.M. (1989), “Some exploratory findings on the development of musical tastes”, Journal of Consumer Research, Vol. 16, June, pp. 119-224. Howe, K.R. (1988), “Against the quantitative-qualitative incompatibility thesis or dogmas die hard”, Educational Researcher, Vol. 17, November, pp. 10-16.

Expansion on diversity in research 289

EJM 39,3/4

290

Howe, K.R. (1992), “Getting over the quantitative-qualitative debate”, American Journal of Education, Vol. 100, February, pp. 236-56. Hudson, L.A. and Ozanne, J.L. (1988), “Alternative ways of seeking knowledge in consumer research”, Journal of Consumer Research, Vol. 14, March, pp. 508-21. Hunt, S.D. (1990), “Truth in marketing theory and research”, Journal of Marketing, Vol. 54, July, pp. 1-15. Hunt, S.D. (1991), “Positivism and paradigm dominance in consumer research – towards critical pluralism and reapproachment”, Journal of Consumer Research, Vol. 18, June, pp. 32-44. Hunt, S.D. (1992), “Marketing is . . .”, Journal of the Academy of Marketing Science, Vol. 20 No. 4, pp. 301-11. Hunt, S.D. (1994), “On rethinking marketing: our discipline, our practice, our methods”, European Journal of Marketing, Vol. 28 No. 3, pp. 13-25. Iacobucci, D. (2001), “Commonalties between research methods for consumer science and biblical scholarship”, Marketing Theory, Vol. 1 No. 1, pp. 109-33. Kaplan, A. (1964), The Conduct of Enquiry, Chandler, San Francisco, CA. Kavanagh, D. (1994), “Hunt versus Anderson: Round 16”, European Journal of Marketing, Vol. 23 No. 3, pp. 26-41. Kernan, J.B. (1995), “Framing the rainbow, focusing the light: JCR’s first twenty five years”, Advances in Consumer Research, Vol. 22, pp. 488-96. Kotler, P. (1979), “Strategies for introducing marketing into nonprofit organizations”, Journal of Marketing, Vol. 43, January, pp. 37-44. Kuhn, T.S. (1968), “History of science”, International Encyclopaedia of the Social Sciences, Vol. XIV, Macmillan, New York, NY, pp. 74-83. Kuhn, T.S. (1970a), The Structure of Scientific Revolutions, 2nd ed., University of Chicago Press, Chicago, IL. Kuhn, T.S. (1970b), “Reflections on my critics”, in Musgrave, A. and Lakatos, I. (Eds), Criticism and the Growth of Knowledge, Cambridge University Press, Cambridge. Kuhn, T.S. (1983), “Commensurability, comparability, communicability”, in Asquith, P.D. and Nickels, T. (Eds), PSA 1982: Proceedings of the 1982 Biennial Meeting of the Philosophy of Science Association, Vol. 2, Philosophy of Science Association, East Lansing, MI, pp. 669-88. Lauden, L. (1984), “Reconstructing methodology”, in Anderson, P.F. and Ryan, M.J. (Eds), Scientific Method in Marketing, AMA Winter Educators’ Conference, American Marketing Association, Chicago, IL, pp. 1-4. Leong, S.M., Sheth, J. and Tan, C.T. (1994), “An empirical study of the scientific styles of marketing academics”, European Journal of Marketing, Vol. 28 No. 8/9, pp. 12-26. Levy, S.J. (2002), “Revisiting the marketing domain”, European Journal of Marketing, Vol. 36 No. 2, pp. 299-304. Lewis, M.W. and Grimes, A.J. (1999), “Metatriangulation: building theory from multiple paradigms”, Academy of Management Review, Vol. 24 No. 4, pp. 672-90. Lowenstein, G. (2001), “The creative destruction of decision research”, Journal of Consumer Research, Vol. 28 No. 3, pp. 499-505. Lynch, J.G. (1998), “Presidential address: reviewing”, Advances in Consumer Research, Vol. 25, pp. 1-6.

McAlexander, J., Schouten, J. and Koenig, H. (2002), “Building brand community”, Journal of Marketing, Vol. 66 No. 1, pp. 38-54. McCracken, G. (1988), Culture and Consumption: New Approaches to the Symbolic Character of Consumer Goods and Activities, Indiana University Press, Indianapolis, IN. McQuarrie, E.F. and Mick, D.G. (1992), “On resonance: a critical pluralistic inquiry into advertising rhetoric”, Journal of Consumer Research, Vol. 19, September, pp. 180-97. Marsden, D. (2001), “Deconstructing consumer behaviour: theory and practice”, Journal of Consumer Behaviour, Vol. 1 No. 1, pp. 9-21. Mick, D.G. (1999), “From the editor-elect”, Journal of Consumer Research, Vol. 25, March, pp. iv-vi. Murray, J.A., O’Driscoll, A. and Torres, A. (2002), “Discovering diversity in marketing practice”, European Journal of Marketing, Vol. 36 No. 3, pp. 373-90. Nagel, E. and Newman, J.R. (1971), Godel’s Proof, Routledge, London. O’Shaughnessy, J. (1997), “Temerarious directions for marketing”, European Journal of Marketing, Vol. 31 No. 9/10, pp. 677-705. O’Shaughnessy, J. and O’Shaughnessy, N. (2002), “Ways of knowing and their applicability”, Marketing Theory, Vol. 2 No. 2, pp. 147-64. Ozanne, J.L. and Hudson, L.A. (1989), “Exploring diversity in consumer research”, in Hirschman, E.C. (Ed.), Interpretive Consumer Research, Association of Consumer Research, Provo, UT, pp. 1-99. Peter, J.P. and Olson, J.C. (1983), “Is science marketing?”, Journal of Marketing, Vol. 47, Fall, pp. 111-25. Piercy, N.F. (2002), “Research in marketing: teasing with trivia or risking relevance?”, European Journal of Marketing, Vol. 36 No. 3, pp. 350-63. Potter, J. and Wetherell, M. (1994), “Analyzing discourse”, in Bryman, A. and Burgess, R.G. (Eds), Analyzing Qualitative Data, Routledge, London, pp. 47-66. Putman, H. (1981), Reason, Truth and History, Cambridge University Press, Cambridge. Remenyi, D. (2002), “Research strategies – beyond the differences”, Electronic Journal of Business Research, Vol. 1 No. 1, pp. 39-41. Rescher, N. (1996), Process Metaphysics: An Introduction to Process Philosophy, State University of New York Press, New York, NY. Riley, R.W. and Love, L.L. (2000), “The state of tourism research”, Annals of Tourism Research, Vol. 27 No. 4, pp. 164-82. Rose, H. (2001), “Life after the science wars”, in Philo, G. and Miller, D. (Eds), Market Killing, Longman, Harlow, pp. 110-24. Rossiter, J.R. (2001), “What is marketing knowledge? Stage I: forms of marketing knowledge”, Marketing Theory, Vol. 1 No. 1, pp. 9-26. St Pierre, E.A. (1997), “Methodology in the fold and the irruption of transgressive data”, International Journal of Qualitative Studies in Education, Vol. 10 No. 2, pp. 175-89. Sankey, H. (1994), Incommensurability Thesis, Ashgate Publishing, Aldershot. Savitt, R. (1990), “Pre-Aldersonian antecedents to macromarketing: insights from the textual literature”, Journal of the Academy of Marketing Science, Vol. 18 No. 4, pp. 293-301. Sayers, B. (1987), “Wittgenstein, relativism, and the Strong thesis in sociology”, Philosophy of Social Sciences, Vol. 17, pp. 133-45. Schmidt, V.H. (2001), “Oversocialised epistemology: a critical appraisal of constructivism”, Sociology, Vol. 35 No. 3, pp. 135-57.

Expansion on diversity in research 291

EJM 39,3/4

292

Schultz, M. and Hatch, M. (1996), “Living with multiple paradigms: the case of paradigm interplay in organizational cultural studies”, Academy of Management Review, Vol. 21 No. 2, pp. 529-57. Shankar, A. and Patterson, M. (2001), “Interpreting the past, writing the future”, Journal of Marketing Management, Vol. 17, pp. 481-501. Sherry, J.F. (1991), “Postmodern alternatives in the interpretive turn in consumer research”, in Kassarjian, H.H. and Robertson, T.S. (Eds), Handbook of Consumer Behavior, Prentice-Hall, Englewood Cliffs, NJ, pp. 548-91. Sherry, J.F. (2000), “Place, technology, and representation”, Journal of Consumer Research, Vol. 27 No. 2, pp. 273-8. Silverman, D. (1997), Qualitative Research, Theory, Methods and Practice, Sage, London. Szmigim, I. and Foxhall, G. (2000), “Interpretive consumer research: how far have we come?”, Qualitative Market Research, Vol. 3 No. 4, pp. 187-9. Sutton, R.I. and Shaw, B.M. (1995), “What theory is not”, Administrative Science Quarterly, Vol. 40, pp. 371-84. Tellis, G.J., Chandy, R.K. and Ackerman, D.S. (1999), “In search of diversity: the record of major marketing journals”, Journal of Marketing Research, Vol. 36, February, pp. 120-31. Terrell, J. (1991), “Disneyland and the future of museum anthropology”, American Anthropologist, Vol. 93 No. 1, pp. 149-53. Thompson, C.J. and Troester, M. (2002), “Consumer value systems in the age of postmodern fragmentation: the case of natural health micro-culture”, Journal of Consumer Research, Vol. 28 No. 4, pp. 550-71. Thompson, C. (1990), “Eureka! And other tests of significance: a new look at evaluating interpretive research”, Advances in Consumer Research, Vol. 17, pp. 25-30. Thompson, C.J. (1993), “Modern truth and postmodern incredulity: a hermeneutic deconstruction of the metanarrative of scientific truth in marketing research”, International Journal of Research in Marketing, Vol. 10, pp. 325-38. Thompson, C.J. (2002), “A re-inquiry on re-inquiries: a postmodern proposal for a critical-reflexive approach”, Journal of Consumer Research, Vol. 29 June, pp. 142-5. Trocchia, P.J. and Berkowitz, D. (1999), “Getting doctored: a proposed model of marketing doctoral student socialisation”, European Journal of Marketing, Vol. 33 No. 7/8, pp. 746-59. Van Mannen, J. (1995), “Style as theory”, Organization Science, Vol. 6 No. 1, pp. 133-43. Venkatesh, A. (1985), “Is marketing ready for Kuhn?”, in Dholakhia, N. and Arndt, J. (Eds), Research in Marketing: Changing the Course of Marketing, JAI Press, London, pp. 45-68. Vickers, G. (1995), The Art of Judgement: A Study of Policy Making, Sage, London. Weick, K.E. (1989), “Theory construction as disciplined imagination”, Academy of Management Review, Vol. 14, pp. 516-31. Weick, K.E. (1999), “Theory construction as disciplined reflexivity: tradeoffs in the ’90s”, Academy of Management Review, Vol. 24 No. 4, pp. 797-806. Wells, W.D. (2001), “The perils of N ¼ 1”, Journal of Consumer Research, Vol. 28 No. 3, pp. 494-8. Wilk, R.R. (2001), “The impossibility and necessity of re-inquiry: finding middle ground in social research”, Journal of Consumer Research, Vol. 28, September, pp. 308-12. Wotton, N., Moore, K. and Elliott, R. (2002), “When theories become tools: toward a framework for pragmatic validity”, Human Relations, Vol. 55 No. 10, pp. 1227-49.

Zaltman, G. (2000), “Consumer researchers: take a hike!”, Journal of Consumer Research, Vol. 26 No. 4, pp. 423-8. Zinkan, G.M., Roth, M.S. and Saxton, M.J. (1992), “Knowledge development and scientific status in consumer behavior research: a social exchange perspective”, Journal of Consumer Research, Vol. 19, September, pp. 282-91.

Expansion on diversity in research

Further reading Kuhn, T.S. (1962), The Structure of Scientific Revolutions, 1st ed., University of Chicago Press, Chicago, IL. Lacher, K.T. (1989), “Hedonic consumption: music as a product”, Advances in Consumer Research, Vol. 16, pp. 488-96.

293

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

EJM 39,3/4

294 Received January 2003 Revised September 2003 Accepted January 2004

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0309-0566.htm

Grounded theory, ethnography and phenomenology A comparative analysis of three qualitative strategies for marketing research Christina Goulding Department of Marketing and Economics, Wolverhampton Business School, Wolverhampton University, Wolverhampton, UK Abstract Purpose – The paper aims to look at some of the problems commonly associated with qualitative methodologies, suggesting that there is a need for a more rigorous application in order to develop theory and aid effective decision making. Design/methodology/approach – The paper examines three qualitative methodologies: grounded theory, ethnography, and phenomenology. It compares and contrasts their approaches to data collection and interpretation and highlights some of the strengths and weaknesses associated with each one. Findings – The paper suggests that, while qualitative methodologies, as opposed to qualitative methods, are now an accepted feature of consumer research, their application in the truest sense is still in its infancy within the broader field of marketing. It proposes a number of possible contexts that may benefit from in-depth qualitative enquiry. Originality/value – The paper should be of interest to marketers considering adopting a qualitative perspective, possibly for the first time, as it offers a snap-shot of three widely-used methodologies, their associated procedures and potential pitfalls. Keywords Marketing theory, Ethnography, Qualitative research Paper type Conceptual paper

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 294-308 q Emerald Group Publishing Limited 0309-0566 DOI 10.1108/03090560510581782

Introduction Within the field of marketing the positivism versus interpretivism debate is showing signs of slowing down, with a growing acceptance of the diverse range of methods of representing marketing phenomenon (Brown, 2003). The last two decades have seen a steady increase in the number of qualitative papers appearing in the premier journals, and while we may be a long way from reaching a full Kuhnian paradigm revolution, it is fair to say that qualitative research is no longer viewed as merely “speculative”, or “soft”, as was generally held to be the case by many in the past. However, that is not to say that some of the criticisms levied against qualitative research in marketing are not entirely without foundation. Just as there are many quantitative papers that fail to give full attention to design and statistical checks, there are many instances of papers claiming to be, for example grounded theory, that are based on purposive sampling and a handful of interviews which are then described, but lack theoretical sensitivity (Glaser, 1992). This has led to accusations that grounded theory is being used as an “anything goes” approach (Locke, 1996, p. 244). There are also “phenomenological” accounts which are free from any guiding philosophy and described in terms of content analysis and even statistics (see for example Masberg and Silverman’s (1996)

phenomenological study of visitor experiences at heritage sites), and ethnographic descriptions which are based on snap-shot observations and limited participatory interaction (Goulding, 2002). However, there is increasing acknowledgement, not only in academic circles, but also among marketing practitioners, of the need for the application of qualitative methodologies in their truest and most fundamental sense in order to gain valid insights, develop theory and aid effective decision making. This paper examines three qualitative methodologies, grounded theory, ethnography, and phenomenology, which have seen a steady growth in their application within marketing. It points to a number of studies that have applied the techniques and summarises the key procedures associated with each in order to differentiate between them and highlight their potential for future marketing research. A grounded theory approach Although traditionally associated with sociology (Glaser and Strauss, 1967; Glaser, 1978, 1992, 1998; Strauss and Corbin, 1990), nursing and health (Sandelowski, 1995; Morse, 1994), and organisational studies (Parry, 1998; Hunt and Ropo, 1995; Brown, 1994, 1995; Turner, 1981, 1988), grounded theory has, in recent years, started to enter the repertoire of marketing and consumer research (Goulding, 1998, 1999a, 2000a; Pettigrew, 2000). For instance, the “Odyssey” team (Belk et al., 1989) utilised aspects of grounded theory in their ground breaking analysis of the sacred and profane in consumer behaviour. Other examples include Hirschman and Thompson’s (1997) grounded theory analysis of advertising and the mass media, Burchill and Fine’s (1997) examination of product concept development, De la Cuesta’s (1994) study of marketing and health visiting, Houston and Venkatesh’s (1996) interpretation of health care consumption practices by Asian immigrants, and Goulding’s (1999b, c, 2000b, c) research into consumer experiences at heritage sites and museums. However, in comparison to other qualitative methodologies its application is still fairly confined to experiential consumer behaviour, despite its potential for theoretical developments across a range of marketing phenomenon that are predicated on a behavioural component. Contextualising grounded theory Essentially, grounded theory has its origins in symbolic interactionsim, a paradigm which holds that individuals engage in a world that requires reflexive interaction as averse to environmental response. Accordingly, behaviour is goal driven, evolving from social interaction that is highly symbolic in itself. This behaviour involves various forms of communication, both verbal and non-verbal and the notion of symbols is intrinsic to the perspective (Schwandt, 1994). Nonetheless, while symbolic interactionism was a key school within sociology particularly during the 1950s and 1960s, there were few guidelines for using its concepts to conduct research. Much of the work evolving from this paradigm was also emerging at a time when qualitative research was under attack for lacking “scientific” procedures and rigour. Consequently there was recognition of the need for a methodology that could track and validate the process of theory building. During this period two American sociologists, Barney Glaser, who was trained at Columbia University New York and was heavily influenced by the formal theorising of Paul Lazarsfeld, Paul Merton, and Herbert Hyman, and Anselm Strauss, a scholar

Analysis of three qualitative strategies 295

EJM 39,3/4

296

tutored in the philosophies of critical and qualitative methodologies of the Chicago School (Locke, 1996), set out to develop a more systematic and defined procedure for the collection and analysis of qualitative data. The methodology they devised was labelled grounded theory to reflect, as the name suggests, theory that is grounded in the words and actions of those individuals under study. Indeed the richness of the methodology has largely been attributed to the influences of the broader sociological schools of symbolic interactionism, ethnography and Glaser’s background in formal theorising. Consequently, grounded theory is suitable to the study of any behaviour that has an interactional element to it. Grounded theory process According to the basic principles of grounded theory, once an area of research has been identified, the researcher should enter the field as soon as possible. Consequently the literature is not exhausted prior to the research, as in many studies, rather it is consulted as part of an iterative, inductive and interactional process of data collection, simultaneous analysis, and emergent interpretation. In other words, the developing theory should direct the researcher to appropriate extant theories and literature that have relevance to the emerging, data grounded concepts. However, it must be noted that the way this has been expressed in the earlier works of Glaser and Strauss has resulted in confusion regarding the nature of “induction”. A common misconception is that the researcher is expected to enter the field ignorant of any theory or associated literature relating to the phenomenon and wait for the theory to emerge purely from the data. Such perceptions have resulted in accusations that both originators have “conned” those they have influenced by preaching the notion of theoretical avoidance, while marching into the field themselves armed with a life-time of concepts and experiences (Goulding, 2002). Nevertheless, this was not necessarily the case. Indeed Glaser and Strauss (1967, p. 253) state that: The core categories can emerge in the sociologist’s mind from his reading, life experiences, research and scholarship; [furthermore] no sociologist can possibly erase from his mind all the theory he knows before he begins his research. Indeed the trick is to line up what one takes as theoretically possible or probable with what one is finding in the field.

For marketers the dilemma is just the same. It involves a delicate balancing act between drawing on prior knowledge while keeping a fresh and open mind to new concepts as they emerge from the data. This means using the literature differently as the process evolves, getting closer to direct sources as the conceptual categories take shape and gain explanatory power. In addition to the use of literature grounded theory differs in a number of respects from other qualitative methodologies, particularly with regard to sampling. According to Coyle (1997), most sampling is purposive and defined before data collection commences. In the case of grounded theory, sampling begins as a “commonsense” process of talking to those informants who are most likely to provide early information. This information is then analysed through the application of open coding techniques, or line-by-line analysis (looking for words and sentences in the text that have meaning), which should help to identify provisional explanatory concepts and direct the researcher to further “theoretically” identified samples, locations, and forms of data. According to the original rules of grounded theory, the researcher should not leave the

field and stop sampling until saturation is reached, or when no additional information is found in the data. On this subject, one of the appeals of grounded theory is that it allows for a wide range of data, the most common of which are in-depth interviews, observations, and memos which describe situations, record events, note feelings and keep track of ideas. However, these are not exhaustive, as noted by Glaser (1978, p. 6):

Analysis of three qualitative strategies

Grounded theory method although uniquely suited to fieldwork and qualitative data, can be easily used as a general method of analysis with any form of data collection: survey, experiment, case study. Further, it can combine and integrate them. It transcends specific data collection methods.

297

Therefore, data may comprise of life histories (McKinley-Wright, 1995; Clondinin and Connelly, 1994), secondary data (Szabo and Strang, 1997), introspection (Corbin, 1998), and even numbers (Glaser and Strauss, 1967). Nevertheless, despite the open and flexible nature of the data that may be used in a grounded theory study, there exist a set of specific principles for analysing and abstracting the information. These include the “constant comparison” method, where, for example, interview texts are analysed line by line, provisional themes noted, and subsequently compared with other transcripts in order to ensure consistency and also to identify negative cases. The next stage is to search for links through the identification of concepts that may go some way to offering an explanation of the phenomenon under study. This process is normally associated with axial coding that is achieved by specifying relationships and delineating a core category or construct around which the other concepts revolve. Axial coding is the appreciation of concepts in terms of their dynamic interrelationships and they should form the basis for theory construction (Spiggle, 1994). The final stage of the theory development process is the construction of a core category (Glaser and Strauss, 1967). A core category pulls together all the concepts in order to offer an explanation of the phenomenon. It should have theoretical significance and should be traceable back through the data. This is usually when the theory is written up and integrated with existing theories to show relevance, fit, and/or extension. Grounded theory as a methodology emerged from the discipline of sociology, an area of enquiry that is focused on society and the individual. However, given the broadening of the marketing discipline over the last two decades to incorporate such issues as ethical marketing, social marketing, green issues and experiential consumption, all of which have significant behavioural implications, the application of grounded theory would seem appropriate. The rigours of the approach force the researcher to look beyond the superficial, to apply every possible interpretation before developing final concepts, and to demonstrate these concepts through explication and data supported evidence. The main problems associated with the methodology appear to stem largely from its misuse and abuse. Examples include studies labelled grounded theory that have not followed the principles of theoretical sampling, inductive coding, constant comparison and so on. However, this is possibly truer of studies stemming from the organisational behaviour literature than it is of marketing. For example, Mullins and Roessier’s (1998) research into employment outcomes ignored theoretical sampling and constant comparison, drew on an established model and collected data through structured mail surveys which featured a rating scale. Consequently, while the study may well have been rigorous, it is hard to claim that it produced a “grounded” theory in the accepted sense.

EJM 39,3/4

298

This practice of methodological muddling (Baker et al., 1992) has traditionally been perceived as a violation of the principles of grounded theory (Locke, 1996), although there is now recognition that it is valid to use grounded theory in part, alongside other approaches so long as the objectives are made clear (Skodal-Wilson and Ambler-Hutchinson, 1996). A good example of this can be seen in Belk et al.’s (1989) “Odyssey” paper which combined a number of methodologies in order to address various questions as the research progressed. Pettigrew (2000) on the other hand discusses the “marriage” of grounded theory with ethnography to gain a deeper insight into consumer experiences. Yet despite the evolving nature of grounded theory and its application to an increasing number of contexts, there remains one outstanding issue; namely that most grounded theory studies are concluded at the “substantive” level (context specific) rather than developed to the general level. The answer to this may lie in a number of factors, not the least of which is time. The development of a general theory involves taking the research into a variety of contexts, ensuring full theoretical sampling and the production of a theory that has applications to other settings and populations (Morse, 1994). This may take years and involve research teams as averse to lone academics, which combined with the pressure to publish means that few have either the time or the resources to commit to such an endeavour. A final significant issue that may act as a deterrent to using grounded theory is that theoretical saturation of the data and the interpretation of that data can make it difficult to anticipate an accurate time scale for the research. This, while problematic in the general sense of planning research strategies, creates even greater difficulties when applying for research grants (Goulding, 2002). However, grounded theory is only one of a wide range of qualitative methodologies currently being used to investigate contemporary marketing phenomenon. The next section examines the ethnographic position within marketing by first discussing its application before providing an overview of the ethnographic process. An ethnographic approach The roots of ethnography lie in cultural anthropology, with its focus on small-scale societies and the original central concept remains paramount today; that is a concern with the nature, construction and maintenance of culture. Ethnographies are always informed by this concept as ethnographers aim to look beyond what people say to understand the shared system of meanings we call “culture”. In her review of ethnography in consumer research Pettigrew (2000) argues that consumption represents a phenomenon that can be effectively addressed with the use of ethnographic techniques, based on the understanding that the social meanings found in material possessions can be viewed as cultural communicators. Arnould (1998, p. 86) provides an in-depth discussion of consumer-oriented ethnography suggesting that “ethnography attempts to explicate structured patterns of action that are cultural and/or social rather than merely cognitive, behavioural or affective”. Stebbins (1997) further illustrates the potential of ethnographic research to the study of lifestyles given the cultural or sub-cultural context within which they are enacted, while Arnould and Wallendorf (1994) discuss the relevance of market-oriented ethnography to developing marketing strategy. Indeed within marketing, the ethnographic turn in enquiry is becoming widespread (Brown, 1998) and critically appraised (Brownlie, 1997). Moreover, there is a growing body of literature, based on ethnographic methods, which

act as exemplars for those wishing to apply the methodology. For example, Arnould and Price’s (1993) much quoted “River magic”, Hill’s (1991) study of homeless women and the meaning of possessions, and Schouten and McAlexander’s (1995) longitudinal study of the new “biker” culture in the USA. From a British perspective Ritson and Elliott’s (1999) analysis of adolescents and their use of advertisements was based on extended encounters with groups of teenagers, including teaching part-time in the schools that they were attending in order to develop trust and rapport, and gain a sense of the dynamics and systems within the environment. Ethnography, therefore, has its own distinct procedures for collecting data, but it is also highly concerned with the cultural context. The next part of this paper attempts to contextualise ethnography before examining the primary techniques associated with ethnographic research. Contextualising ethnography It is important for the researcher considering using ethnography to understand the various types of investigation that may potentially form the framework for analysis. Ethnography can be any full or partial description of a group (ethno – folk, graphy – description), as a means of identifying common threads, whether these be religion, social relationships or management style. Ethnographies may be cross sectional such as Goffman’s (1961) study of asylums which looked at a cross section of “total” institutions (Fine and Martin, 1990), or ethnohistorical, which describe the cultural reality of the present as the historical result of events in the past. They may be classified on the basis of spatial or geographic dimensions, by language, by theory, or in any number of diverse ways; there are few limitations to the cultural contexts to which it can be applied (Boyle, 1994). Ethnographic process A key feature of ethnography is that it is labour intensive and always involves prolonged direct contact with group members in an effort to look for rounded, holistic explanations. The hallmark of ethnography is fieldwork; working with people in their natural settings. The voices of participants are an important source of data and should be allowed to be heard in the written end-product, which should be a coherent, fluent and readable narrative (Boyle, 1994; Muecke, 1994). Arnould (1998) offers a useful summary of ethnography and its role in consumer research, which includes: . Ethnography should aim to explain the ways that culture constructs and is constructed by the behaviours and experiences of its members. . Ethnography involves prolonged participation within a specific culture or sub-culture. . Ethnography in consumer research tends to be particularistic rather than generalisable, espousing pluralistic accounts of consumption. . The potential for ethnography lies in applying multiple data collection methods at a single phenomenon. These may range from surveys to observational data, video tapes, photographs and recordings of speech in action. . Ethnography requires tactics for representing research findings. These representations should aim to unravel the layered meanings that marketing activities hold for the customer.

Analysis of three qualitative strategies 299

EJM 39,3/4

300

Holbrook (1998) provides an interesting departure from the traditional participatory observational approach to ethnography in his description of his “ethnoscopic” journey to Kroywen. In his study he borrows from cultural anthropology and visual sociology, making use of photographs and other pictorial records to reveal revelationary collages based on intense introspection and reflection. The reflexive nature of ethnography is a characteristic which implies that the researcher is part of the world that is under study and is consequently affected by it (Boyle, 1994). In a catch 22 situation, this view has resulted in criticisms of “value laden” interpretations (Borman and Preissle-Goez, 1986). However, the interpreter does not take the data on face value, but considers it as a field of inferences in which hypothetical patterns can be identified and their validity tested. This usually entails a long and discursive process over data interpretation, reasoning, and the fit to related studies (Agar, 1983; Borman and Preissle-Goez, 1986). Additionally, with ethnography, insider and outsider views combine to provide deeper insights than would be possible by the “native” alone. This two-sided view produces a third dimension that rounds off the ethnographic picture, which is a theoretical explanation of the phenomena under study. The emic perspective is at the heart of ethnography, while the etic perspective is the researchers abstractions or scientific explanations of reality (Boyle, 1994). Boyle further suggests, that although ethnographies vary, most combine some element of etic and emic analysis, although the emphasis may differ according to the philosophy of the researcher. With regard to the analysis of ethnographic data, this involves the search for patterns, and ideas that will help explain the existence of these patterns, taking into consideration emic and etic interpretations. This is frequently done through the application of content analysis, a technique for making inferences from text data. Each word or phrase in a text is categorised by applying labels that reflect concepts such as aggression, denial and so forth inherent within it. Some counting may be done (but not always), and some researchers use factor analysis, although this runs the risk of substituting numbers for rich description and it is rare to see this in an ethnographic study. More often than not, the ethnographer identifies categories and instances within the data by desegregating the text (notes) into a series of fragments which are then regrouped under a set of thematic headings. Comprehension is thought to be complete when the researcher can describe the events, incidents and exceptions from an emic perspective (Morse, 1994). Synthesis, involves coding and content analysis where the data is pooled and the constructed categories are linked. Often, however, ethnographic analysis is not developed beyond the level of “thick description”, presented as informants’ stories and case studies. Furthermore, the analytical phase of theorising is seldom separated from the descriptive discourse and treated as a separate level with a distinct purpose (Morse, 1994). With regard to theory, ethnography is embedded in cultural theory and the establishment of macro/micro, or etic/emic, distinctions are ultimately linked to established theory. Finally, recontextualisation is achieved by forcing the theory to a level of abstraction, the degree of which determines the generalisability of the theory. However, according to Atkinson and Hammersley (1995) the degree to which ethnography is conducted in its “truest” form is sometimes controversial. To some it constitutes a philosophical paradigm to which one makes a total commitment, for others it designates a method that one uses where and when appropriate. Furthermore

they propose that the nature of participatory observation is something that has attracted criticism on ethical grounds and it is argued that the following should be made explicit: . Whether the researcher is known to be a researcher by those studied. . How much, and what is known by the participants about the research. . What sort of activities are and are not engaged in by the researcher. . What is the orientation of the researcher and how involved is he/she in the situation. While these factors are not necessarily unique to ethnography, the matter of involvement and participation brings them very much to the fore. A further bone of contention stems from those who adopt a critical stance toward ethnography. Critical ethnography, comprising largely of feminist and post-modernist perspectives, is thought to present an impressionistic collage, an image that represents only a “slice” of time and context and therefore challenges the claims of holistic interpretations (Muecke, 1994). The final part of this paper looks at phenomenology, a methodology that has seen increasing use, especially in the field of consumer behaviour. A phenomenological approach Phenomenology, as both a philosophy and a methodology has been used in organisational and consumer research in order to develop an understanding of complex issues that may not be immediately implicit in surface responses. However, within the field of marketing, it is probably the work of Craig Thompson that has done more to highlight both the underlying principles of phenomenology (Thompson, 1997, 1998; Thompson et al., 1989) and its application to various research situations. Examples of these include Thompson’s (1996) exploration of gendered consumption and lifestyle, Thompson and Hirschman’s (1995) analysis of self care practices and self conceptions, Thomson et al.’s (1990) study of the everyday consumption practices of married women, and Thompson and Haykto’s (1997) deconstruction of the meaning of fashion discourses and the link to identity and self conceptions. Other consumer studies which have adopted a phenomenological position include Mick and Demoss’s (1990) exploration of self gift giving, O’Guinn and Faber’s (1989) work on compulsive shopping, Woodruffe-Burton et al.’s (2002) research into gender and addictive consumption, Goulding et al.’s (2002) analysis of dance culture, the link to postmodern identity fragmentation and the emergence of neo-communities, and Seebaransingh et al.’s (2002) insightful account of the relationship between cosmetic surgery and the construction of identity. These studies provide examples of phenomenology in practice, however it is necessary to understand the origins and particular techniques associated with the methodology in order to distinguish it from other forms of qualitative enquiry. Contextualising phenomenology With regard to the development and application of phenomenology it is probably fair to say that it has had a long, controversial, and often confusing history within the social sciences (Rehorick and Taylor, 1995). Moreover, depending on one’s epistemological and ontological position, it is either conceptualised as a philosophy,

Analysis of three qualitative strategies 301

EJM 39,3/4

302

for those who adhere to the thinking of Husserl (1962) and Heidegger (1962), or a methodology, for those who adopt the position put forward by Schutz (1967). While Husserl’s intention was to develop a schema for describing and classifying subjective experiences of what he termed the life world (Langenbach, 1995), Schutz (1967) developed the approach as a method which incorporated details of experience often at the level of mundane everyday life. The lifeworld is defined as the world in which we, as human beings among fellow human beings, experience culture and society, take a stand with regard to their objects, are influenced by them, and act on them (Schutz, 1966). Gregova (1996), distinguishes between the lifeworld and the social world, proposing that the lifeworld consists of formal structures about which we are less explicitly aware, while the social world relates to everyday familiar actions and experiences. Essentially, the goal of phenomenology is to enlarge and deepen understanding of the range of immediate experiences (Spiegelberg, 1982). Merleau-Ponty (1962, p. vii) suggests that the results of phenomenological enquiry should be “a direct description of our experience without taking account of its psychological origin”. Phenomenology therefore is a critical reflection on conscious experience, rather than subconscious motivation, and is designed to uncover the essential invariant features of that experience (Jopling, 1996). Phenomenological process As one of the major influences on phenomenological enquiry Schutz proposed that individuals approach the life world with a stock of knowledge made up of common sense constructs and categories that are essentially social in action. These stocks of knowledge produce familiarity, but they are always incomplete and open ended. Naming requires the interpretative application of a category to the concrete particulars of a situation (Holstein and Gubrium, 1994). Language is the central medium for transmitting meaning and as such provides a methodological orientation for a phenomenology of social life that is concerned with the relation between language use and the objects of experience. The meaning of a word is taken to be what it references, corresponds with, or stands for in the real world. This is based on the premise that the essential task of language is to convey information and describe “reality”. It is also assumed that there is a degree of commonalty in that others experience the world in fundamentally the same way, intersubjectively sharing the same meaning. The basic assumption is that a person’s life is a socially constructed totality in which experiences interrelate coherently and meaningfully. With regard to the process of enquiry, the phenomonologist has only one legitimate source of data, and that is the views and experiences of the participants themselves. This in itself assumes that the participant’s view is taken as “fact”. Furthermore, participants are selected only if they have lived the experience under study. Sampling is therefore purposive and prescribed from the start and the main instrument of data collection is the interview. Thompson (1998) suggests that the relevant implication of the life-world is that the consumer as text metaphor needs to be reflectively adapted and enriched through a dynamic combination of consumers who interpret their experiences and consumer researchers who interpret consumer’s narratives. In other words “it is important to recognise that human existence has some text like qualities but it is not text per se” (Thompson, 1998, p. 129). As a means of interpretation, Thompson (1997) in his analysis of consumer experiences, advocates part to whole analysis of participant

accounts by proceeding through an interactive process. This involves reading texts (interview transcripts) in full, in order to first gain a sense of the whole picture. After several readings of the text, the next stage is hermeneutic endeavour (Thompson et al., 1990), or intertextuality (Thompson, 1997), whereby patterns and differences are sought across transcripts. This strategy of interpretation must broaden the analysis to include a wider range of considerations that helps the researcher arrive at a holistic interpretation. There must also be recognition that the final explanation represents a fusion of horizons between the interpreter’s frame of reference and the texts being interpreted (Thompson, 1997). This is largely concurrent with the process outlined by Colaizzi (1978) who suggests a series of seven steps: (1) The first task of the researcher is to read the participants narratives, to acquire a feeling for their ideas in order to understand them fully. (2) The next step “extracting significant statements”, requires the researcher to identify key words and sentences relating to the phenomenon under study. (3) The researcher then attempts to formulate meanings for each of these significant statements. (4) This process is repeated across participants’ stories and recurrent meaningful themes are clustered. These may be validated by returning to the informants to check interpretation. (5) After this the researcher should be able to integrate the resulting themes into a rich description of the phenomenon under study. (6) The next step is to reduce these themes to an essential structure that offers an explanation of the behaviour. (7) Finally, the researcher may return to the participants to conduct further interviews or elicit their opinions on the analysis in order to cross check interpretation. These structures provide the researcher with an understanding of the world that contributes towards the development of theory. On the question of theory Morse (1994) suggests that phenomenologists do not label themselves as theorists in the strictest sense, rather linkages from the data to theory are based on reflections of theoretical literature. Recontextualisation comes from writing and rewriting, which as a practice sensitises the researcher and provides new insights. This increases the level of abstraction by moving from the “particular” sphere to a “universal” sphere where themes are readily comprehensible to the humans which they seek to describe. However, there are a number of misconceptions and ad hoc interpretations of phenomenology to be found in the literature. First, the philosophy that underpins phenomenology is often misused to refer to the qualitative paradigm as a whole. This is not the case. Phenomenologists seek guidance from existential philosophers in the interpretation of their data. Through careful study of individuals they hope to discover the deeper meaning of the “lived” experience in terms of the individual’s relationship with time, space and personal history (Stern, 1994). The collection and analysis of data using this methodology is also specifically prescribed. Phenomenology demands that intense reflection is an integral part of the process, but above all, the primacy of the subjective experience is felt to be crucial. Analysis is conducted by scrutinising the text

Analysis of three qualitative strategies 303

EJM 39,3/4

304

for meaning “units” which describe the central aspects of the experience. These are then synthesised to provide a general description of the “whole”. Conclusion This paper has looked at three different qualitative methodologies, each quite distinct in their procedures for data collection, interpretation and theoretical development. For example, grounded theory is flexible in terms of data, but insists on theoretical sampling and saturation of both data and theory before theory development can be claimed. It is also a methodology particularly suited to situations that have a symbolic and, or, an interactional element to them. Consequently, it has potential for a number of research directions and contexts that go beyond consumer behaviour, for example, relationship marketing or even the sales situation. Ethnography is generally concerned with culture and power with the main forms of data generally participatory observation and interviews. Again, research agendas that focus on, for example, inter-departmental dynamics, gender issues, ethical marketing or green consumption, may well benefit from the application of an ethnographic approach. Finally, phenomenology has its own unique characteristics and philosophy, which may be beneficial in terms of theory building based around lived experiences, whether these be consumption or strategic decision making. In sum, each has its own strengths and weaknesses, but each are suited to the study of marketing phenomenon. However, despite the flexibility of the methodologies described, their application remains largely in the field of experiential consumer behaviour with their potential for theory development within the broader field of marketing yet to be fully realised. Possibly, this may be the next step in the development of qualitative research with marketing. References Agar, M. (1983), “Ethnographic evidence”, Urban Life, Vol. 12 No. 10, pp. 32-48. Arnould, E.J. (1998), “Daring consumer-oriented ethnography”, in Stern, B. (Ed.), Representing Consumers: Voices, Views and Visions, Routledge, London. Arnould, E.J. and Price, L. (1993), “River magic: extraordinary experience and the extended service encounter”, Journal of Consumer Research, Vol. 20, June, pp. 24-45. Arnould, E.J. and Wallendorf, M. (1994), “Market oriented ethnography: interpretation building and marketing strategy formulation”, Journal of Marketing Research, Vol. 31, November, pp. 484-504. Atkinson, P. and Hammersley, M. (1995), Ethnography, 2nd ed., Routledge, London. Baker, C., West, J. and Stern, P. (1992), “Method slurring; the grounded theory/phenomenology example”, Journal of Advanced Nursing, Vol. 17 No. 11, pp. 1355-60. Belk, R.W., Wallendorf, M. and Sherry, J.F. (1989), “The sacred and the profane in consumer behaviour: theodicy on The Odyssey”, Journal of Consumer Research, Vol. 16, June, pp. 1-37. Borman, K. and Preissle-Goez, J. (1986), “Ethnographic and qualitative research design and why it doesn’t work”, American Behavioural Scientist, Vol. 30 No. 1, pp. 42-57. Boyle, J.S. (1994), “Styles of ethnography”, in Morse, J.M. (Ed.), Critical Issues in Qualitative Research Methods, Sage, Thousand Oaks, CA. Brown, A. (1994), “Politics, symbolic action and myth making in pursuit of legitimacy”, Organization Studies, Vol. 15 No. 6, pp. 861-78.

Brown, A. (1995), “Managing understandings: politics, symbolism, niche marketing and the quest for legitimacy in IT implementation”, Organization Studies, Vol. 16 No. 6, pp. 951-69. Brown, S. (1998), Postmodern Marketing 2: Telling Tales, Thompson Business Press, London. Brown, S. (2003), “Crisis, what crisis? Marketing, Midas, and the Croesus of representation”, Qualitative Marketing Research: An International Journal, Vol. 6 No. 3, pp. 194-205. Brownlie, D. (1997), “Beyond ethnography: towards writerly accounts of organising in marketing”, European Journal of Marketing, Vol. 31 No. 3/4, pp. 263-82. Burchill, G. and Fine, C. (1997), “Time versus market orientation in product concept development: empirically-based theory generation”, Management Science, Vol. 43 No. 4, pp. 465-78. Clondinin, D.J. and Connelly, F.M. (1994), “Personal experience methods”, in Denzin, N. and Lincoln, Y. (Eds), Handbook of Qualitative Research, Sage, London. Colaizzi, P. (1978), “Psychological research as a phenomenologist views it”, in Valle, R. and King, M. (Eds), Existential Phenomenological Alternatives for Psychology, Oxford University Press, New York, NY. Corbin, J. (1998), “Alternative interpretations: valid or not?”, Theory and Psychology, Vol. 8 No. 1, pp. 121-8. Coyle, I.T. (1997), “Sampling in qualitative research. purposeful and theoretical sampling; merging or clear boundaries?”, Journal of Advanced Nursing, Vol. 26 No. 3, pp. 623-30. De la Cuesta, C. (1994), “Marketing: a process in health visiting”, Journal of Advanced Nursing, Vol. 19 No. 2, pp. 347-53. Fine, G. and Martin, D. (1990), “A partisan view: sarcasm, satire, and irony as voices in Erving Goffman’s asylums”, Journal of Contemporary Ethnography, Vol. 19 No. 1, April, pp. 89-115. Glaser, B. (1978), Theoretical Sensitivity, The Sociology Press, Mill Valley, CA. Glaser, B. (1992), Basics of Grounded Theory Analysis: Emergence v. Forcing, The Sociology Press, Mill Valley, CA. Glaser, B. (1998), Doing Grounded Theory: Issues and Discussions, Sociology Press, Mill Valley, CA. Glaser, B. and Strauss, A. (1967), The Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine, Chicago, IL. Goffman, E. (1961), “Asylums”, in Cressey, D. (Ed.), Holt, Rhinenhart, Winston, New York, NY. Goulding, C. (1998), “Grounded theory: the missing methodology on the interpretivist agenda”, Qualitative Marketing Research: An International Journal, Vol. 1 No. 1, pp. 50-7. Goulding, C. (1999a), “Consumer research, qualitative paradigms, and methodological ambiguities”, European Journal of Marketing, Vol. 33 No. 9/10, pp. 859-73. Goulding, C. (1999b), “Heritage, nostalgia, and the ‘grey’ consumer”, The Journal of Marketing Practice: Applied Marketing Science, Vol. 5 No. 6/7/8, pp. 177-99. Goulding, C. (1999c), “Museum culture and consumer behaviour”, Journal of Marketing Management, Vol. 15, November, pp. 647-72. Goulding, C. (2000a), “Grounded theory and consumer behaviour: principles, practice, and pitfalls”, Advances in Consumer Research, Vol. 27, September. Goulding, C. (2000b), “The museum environment and the visitor experience”, European Journal of Marketing, Vol. 34 No. 3/4, pp. 433-52.

Analysis of three qualitative strategies 305

EJM 39,3/4

306

Goulding, C. (2000c), “The commodification of the past, postmodern pastiche, and the search for authentic experiences at contemporary heritage attractions”, European Journal of Marketing, Vol. 34 No. 7, pp. 835-53. Goulding, C. (2002), Grounded Theory; A Practical Guide for Management, Business and Market Researchers, Sage, London. Goulding, C., Shankar, A. and Elliott, R. (2002), “Working weeks, rave weekends: identity fragmentation and the emergence of new communities”, Consumption, Markets and Culture, Vol. 5 No. 4, pp. 261-84. Gregova, M. (1996), “Restructuring of the life world of socialism”, International Sociology, Vol. 11 No. 1, pp. 63-78. Heidegger, M. (1962), Being and Time, Harper & Row, New York, NY. Hill, R. (1991), “Homeless women, special possessions and the meaning of home: an ethnographic case study”, Journal of Consumer Research, Vol. 18, December, pp. 298-310. Hirschman, E.C. and Thompson, C. (1997), “Why media matter: toward a richer understanding of consumers’ relationships with advertising and mass media”, The Journal of Advertising, Vol. 26 No. 1, pp. 43-60. Holbrook, M. (1998), “Journey to Kroywen: an ethnoscopic auto-auto-auto-driven stereographic photo-essay”, in Stern, B. (Ed.), Representing Consumers: Voices, Views and Visions, Routledge, London. Holstein, J. and Gubrium, J. (1994), “Phenomenology, ethnography and interpretative practice”, in Denzin, N. and Lincoln, Y. (Eds), Handbook of Qualitative Research, Sage, Beverly Hills, CA. Houston, H.R. and Venkatesh, A. (1996), “The health-care consumption patterns of Asian immigrants: grounded theory implications for consumer acculturation theory”, Advances in Consumer Research, Vol. 23, pp. 418-23. Hunt, J.G. and Ropo, A. (1995), “Multi-level leadership: grounded theory and mainstream theory applied to the case of General Motors”, Leadership Quarterly, Vol. 6 No. 3, pp. 379-412. Husserl, E. (1962), The Crisis of European Sciences and Transcendental Phenomenology, trans. by Carr, D., Northwestern University Press, Evanston, IL. Jopling, D. (1996), “Sub-phenomenology”, Human Studies, Vol. 19 No. 2, pp. 153-73. Langenbach, M. (1995), “Phenomenology, intentionality, and mental experiences”, History of Psychiatry, Vol. 5, pp. 209-24. Locke, K. (1996), “Rewriting the discovery of grounded theory after 25 years?”, Journal of Management Inquiry, Vol. 5 No. 3, pp. 239-45. McKinley-Wright, M. (1995), “I never did any fieldwork, but I milked an awful lot of cows! Using rural women’s experience to reconceptualise models of work”, Gender and Society, Vol. 9 No. 2, pp. 216-35. Masberg, B. and Silverman, L. (1996), “Visitor experiences at heritage sites: a phenomenological approach”, Journal of Travel Research, Vol. 34 No. 4, pp. 20-5. Merleau-Ponty, M. (1962), Phenomenology of Perception, trans. by Smith, C., Routledge & Kegan Paul, London. Mick, D.G. and Demoss, M. (1990), “Self gifts: phenomenological insights from four contexts”, Journal of Consumer Research, Vol. 17, December, pp. 322-32. Morse, J.M. (1994), “Emerging from the data: the cognitive process of analysis in qualitative enquiry”, in Morse, J.M. (Ed.), Critical Issues in Qualitative Research Methods, Sage, Thousand Oaks, CA.

Muecke, M.A. (1994), “On the evaluation of ethnographies”, in Morse, J.M. (Ed.), Critical Issues in Qualitative Research Methods, Sage, Thousand Oaks, CA. Mullins, J. and Roessier, R. (1998), “Improving employment outcomes; perspectives of experienced counsellors regarding the importance of counselling tasks”, Journal of Rehabilitation, Vol. 64 No. 2, pp. 12-18. O’Guinn, T.C. and Faber, R.J. (1989), “Compulsive buying: a phenomenological exploration”, Journal of Consumer Research, Vol. 16 September, pp. 147-57. Parry, K.W. (1998), “Grounded theory and social process: a new direction for leadership research”, Leadership Quarterly, Vol. 9 No. 1, pp. 85-105. Pettigrew, S. (2000), “Ethnography and grounded theory: a happy marriage?”, Advances in Consumer Research, Vol. 27, pp. 256-60. Rehorick, D. and Taylor, G. (1995), “Thoughtful incoherence: first encounters with the phenomenological-hermeneutic domain”, Human Studies, Vol. 18, pp. 389-414. Ritson, M. and Elliott, R. (1999), “The social uses of advertising: an ethnographic study of adolescent advertising audiences”, Journal of Consumer Research, Vol. 26, June, pp. 260-77. Sandelowski, M. (1995), “Focus on qualitative methods: sample size in qualitative research”, Research in Nursing and Health, Vol. 18, pp. 179-83. Schouten, J.W. and McAlexander, J.H. (1995), “Subcultures of consumption: an ethnography of the new bikers”, Journal of Consumer Research, Vol. 22, June, pp. 43-62. Schutz, A. (1966), “Some structures of the lifeworld”, Collected Papers, Vol. 3, Martinus Nijhoff, The Hague. Schutz, A. (1967), The Phenomenology of the Social World, Northwestern University Press, Evanston, IL. Schwandt, T.A. (1994), “Constructivist, interpretivist approaches to human enquiry”, in Denzin, N.K. and Lincoln, Y.S. (Eds), Handbook of Qualitative Research, Sage, Thousand Oaks, CA. Seebaransingh, N., Patterson, M. and O’Malley, L. (2002), “Finding ourselves: women, breast augmentation and identity”, Gender, Marketing and Consumer Behaviour, Vol. 6, pp. 15-16. Skodal-Wilson, H. and Ambler-Hutchinson, S. (1996), “Methodological mistakes in grounded theory”, Nursing Research, Vol. 45 No. 2, pp. 122-4. Spiegelberg, H. (1982), The Phenomenological Movement: A Historical Introduction, Martinus Nijhoff, The Hague. Spiggle, S. (1994), “Analysis and interpretation of qualitative data in consumer research”, Journal of Consumer Research, Vol. 21 No. 3, pp. 491-503. Stebbins, R. (1997), “Lifestyle as a generic concept in ethnographic research”, Quality and Quantity, Vol. 31 No. 4, pp. 347-60. Stern, B. (1994), “Classical and vignette television advertising dramas: structural models, formal analysis and consumer effects”, Journal of Consumer Research, Vol. 19, June, pp. 601-15. Strauss, A. and Corbin, J. (1990), Basics of Qualitative Research: Grounded Theory Procedures and Techniques, Sage, London. Szabo, V. and Strang, V.R. (1997), “Secondary analysis of qualitative data”, Advances in Nursing Science, Vol. 20 No. 2, pp. 66-74. Thompson, C.J. (1996), “Caring consumers: gendered consumption meanings and the juggling of lifestyle”, Journal of Consumer Research, Vol. 22, March, pp. 388-407.

Analysis of three qualitative strategies 307

EJM 39,3/4

308

Thompson, C.J. (1997), “Interpreting consumers: a hermeneutic framework for deriving marketing insights from the texts of consumers’ consumption stories”, Journal of Marketing Research, Vol. XXXIV, November, pp. 438-55. Thompson, C.J. (1998), “Living the texts of everyday life: a hermeneutic perspective on the relationships between consumer stories and life-world structures”, in Stern, B. (Ed.), Representing Consumers: Voices, Views and Vision, Routledge, London. Thompson, C.J. and Haykto, D.L. (1997), “Consumers uses of fashion discourses and the appropriation of countervailing cultural meanings”, Journal of Consumer Research, Vol. 25, March, pp. 139-53. Thompson, C.J. and Hirschman, E.C. (1995), “Understanding the socialised body: a poststructuralist analysis of consumers’ self conceptions, body images, and self care practices”, Journal of Consumer Research, Vol. 25 No. 2, pp. 139-53. Thompson, C.J., Locander, W.B. and Pollio, H.R. (1989), “Putting consumer research back into consumer behaviour: the philosophy and method of existential phenomenology”, Journal of Consumer Research, Vol. 16 September, pp. 133-46. 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. Turner, B. (1981), “Some practical aspects of qualitative data analysis: one way of organising the cognitive process associated with the generation of grounded theory”, Quality and Quantity, Vol. 15, pp. 225-47. Turner, B. (1988), “Connoisseurship in the study of organizational cultures”, in Bryman, A. (Ed.), Doing Research in Organizations, Routledge, London. Woodruffe-Burton, H., Eccles, S. and Elliott, R. (2002), “The effect of gender on addictive consumption: reflections on men, shopping and consumption meaning”, Gender, Marketing and Consumer Behaviour, Vol. 6, pp. 239-56. Further reading Anderson, P.F. (1986), “On method in consumer research: a critical relativist perspective”, Journal of Consumer Research, Vol. 13, September, pp. 155-73. Annells, M. (1996), “Grounded theory method: philosophical perspectives, paradigm of enquiry, and postmodernism”, Qualitative Health Research, Vol. 6 No. 3, pp. 379-93. Atkinson, P. and Hammersley, M. (1994), “Ethnography and participant observation”, in Guba, G. and Lincoln, Y. (Eds), Handbook of Qualitative Research, Sage, Beverly Hills, CA. Bloch, C. (1996), “Emotions and discourse”, Human Studies, Vol. 16 No. 3, pp. 323-41. Calder, B. and Tybout, A. (1987), “What consumer research is”, Journal of Consumer Research, Vol. 14 June, pp. 136-40. Chock, P. (1986), “Irony and ethnography”, Anthropology Quarterly, Vol. 52 No. 2, pp. 87-96. Hirschman, E. (1985), “Scientific style and the conduct of consumer research”, Journal of Consumer Research, Vol. 12 No. 2, pp. 225-39. Srubar, I. (1998), “Phenomenological analysis and its contemporary significance”, Human Studies, Vol. 21, pp. 121-39. Strauss, A. and Corbin, J. (1994), “Grounded theory methodology: an overview”, in Denzin, N. and Lincoln, Y. (Eds), Handbook of Qualitative Research, Sage, Beverly Hills, CA.

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0309-0566.htm

Qualitative research in marketing Road-map for a wilderness of complexity and unpredictability Evert Gummesson

Qualitative research in marketing 309

School of Business, Stockholm University, Stockholm, Sweden Abstract

Received March 2003 Revised November 2003 Accepted January 2004

Purpose – To discuss and analyse three themes in qualitative research in marketing which are objects of both frustration and confusion: analysis and interpretation; theory generation; and a quest for scientific pluralism and individual researcher lifestyles. Design/methodology/approach – Underpinning the discussion is that complexity, ambiguity, fuzziness, chaos, change, uncertainty and unpredictability are characteristics of a market economy; that qualitative and subjective interpretation is necessary to add the spark of life to marketing data; and that general marketing theory needs more attention from researchers. Practical implications – The proper use of methodology and the generation of better marketing theory will make it easier for practitioners to reach the right decisions. Findings – Quantitative and qualitative research processes are not by nature antagonistic, although their advocates may be; quantitative methodology carries qualitative “bugs”, necessary for its sustenance. Originality/value – The article ends with a recommendation that every researcher in marketing should design his or her individual research approach, one that suits the personality of the researcher. As an example, the author presents his own current methodology-in-use, interactive research. Keywords Qualitative research, Marketing theory Paper type Conceptual paper

Marketing: established – but without clout? Let me begin with my perception of the current state of research in marketing. It can of course be argued against, but it can also be accepted as a vantage point for a scientific discourse. The purpose then is not to prove whether I am right or wrong, but to assess what value a journey from my personally chosen departure point can add to research in marketing. Marketing is an established discipline in business schools and an established function in enterprises, where, at best it is part of the corporate culture, at worst it is encapsulated in its own silo. But has marketing indeed got any clout in the boardroom and among top management? Very often marketing seems to be reduced to customer satisfaction studies and planning. Despite confessions to customer-orientation, my feeling is that the prime time of the CEO’s agenda is occupied by finance, accounting, and technology, most commonly information technology. On the academic side Brown (1993, p. 28) laments “marketing’s perennial search for academic respectability” and “the discipline’s lowly standing in a scholarly caste system”; and Piercy (2002, p. 352) argues that in research in marketing “. . . there is a pathology of mediocrity and a process of trivialization . . .which threatens its scientific status, its practical importance, and its future as a major part of business-school curricula”. I too have doubts about the contribution of marketing as an academic

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 309-327 q Emerald Group Publishing Limited 0309-0566 DOI 10.1108/03090560510581791

EJM 39,3/4

discipline and the way research is performed. According to Alvesson and Willmott (1996, p. 128): . . . marketing is perhaps the subdiscipline of management to which critical theory (and related intellectual traditions) can contribute most, and yet it is also in this specialism that the influence of critical analysis is weakest.

310

Their approach is critical theory and Habermas’ (1972) “illusion of pure theory”, and the necessity to challenge the foundation of research and the societal consequences of its practice. A dialogue on the standing of marketing could also start with Kuhn’s (1970) conflict between mainstream science and paradigm shifts; DeBono’s (1971) vertical thinking versus lateral thinking; or Feyerabend’s (1975) doubts about the value of scientific methods. Saunders (1999, p. 85) succinctly expresses a solution: Quantitative methods take marketing from an art to a science, from conjecture to rigour.

This article is inclined to argue that: . being quantitative can contribute to raise the scientific status of marketing but is not sufficient; . quantitative methods cannot achieve scientific excellence without a clear awareness of their qualitative dependency; and . a merger of the best of both worlds – rather than a one-sided acquisition – will add substantial synergy to research in marketing. This resonates the words of Van Maanen (2000, p. x): Meaning and interpretation are required to attach significance to counts and classifications and these are fundamentally qualitative matters. The two approaches are then bound together, neither capturing truth alone nor trumping the other.

The theme of this article is qualitative methods, but seeing those as intertwined with quantitative methods; their shared ground – which is less noted than their differences – will be part of the discussion. Several questions could be raised about research in marketing and its contribution to academe, business practice and society. Do we really come up with results of any impact? Do we offer anything novel that also has practical relevance? Do we make things happen? Do we even react when things happen or do we just follow the “research as usual” daily routine? Where were we during the dotcom craze? How many warned and asked the simple but crucial questions about what the e-shopping advocates and trend gurus were peddling when their prophecies were so blatantly inconsistent? Is our voice heard in the current recession? Is research methodology with its techniques mainly producing trivial and shallow results with little import? This article cannot offer answers to all these questions, but it attempts to explore issues that I consider important for improving research in marketing and its relevance. From the many areas of qualitative research that could be discussed, I will limit myself to three that particularly engage my mind: analysis and interpretation; theory generation; and the individual researcher’s approach to science. Each in turn will be treated in the following sections.

Analysis and interpretation Marketing situations are not easy to grasp. It should be unnecessary to say that marketing decisions and actions are not just based on analyses of data, because data are mostly hard to find, hard to define, and they are incomplete. The reason I mention this is that there seems to be a scholarly dream that if we just do more research and collect more data (“get all the facts”) we will eliminate uncertainty and risk and make rational decisions. However, the real world does not allow this. The marketing manager entering 2003 suffered the aftermath of a brief passage of fast-growing dotcoms with skyrocketing stock prices on volatile exchanges, partly based on fraudulent information. The stock prices took a sharp dive and hit the ground. Soon after, the recently deregulated and more than 100-year-old telecommunications sector imploded and was left in severe crisis. The general economic and political climate facing the marketing manager was a long drawn-out recession; deregulation of markets turning out to be more irregular than expected; European Union expansion plans with unknown future effects on marketing; the business effects of global terrorism; and a host of other uncertainties. Although it is tremendously important how you collect data, it is just as important what you do with the data once collected. In reports based on qualitative research, data collection is usually described, sometimes at length, whereas analysis and interpretation stand out as the Achilles’ heel. Analysis and interpretations are part and parcel of the same issue: how to make sense of data. The term “analysis” is primarily associated with techniques and research designs that are explicit and rigorous and can be replicated by others, that is primarily quantitative research. Interpretation is more linked with qualitative approaches. Its process is not as explicit, transparent and orderly as in analysis, and replication is more difficult. Hermeneutics was once the science of interpreting biblical texts. In today’s research practice, hermeneutics is a general approach to interpretation. Marketing hermeneutics is perhaps not so far from the original application, considering the shopping and consumption evangelism of the modern market economy. Interpretation is a necessity in all human effort to understand the world and specific aspects of interpretation appear in all types of research, not only in qualitative approaches as will be shown later. Hermeneutics does not offer a set of rules for the researcher; it is rather a name for a conscious search for meaning and understanding. I am, however, inspired by the work of O¨dman (1985) and the guidelines he has worked out for hermeneutic research practice. Noting that researchers in marketing have difficulties in analysing and interpreting qualitative data, I have chosen to list a series of strategies that might be helpful in research practice. These can be extracted from manuals on qualitative market research, sociology, educational science and ethnography. More books are also appearing that apply qualitative methods specifically to research in management (for example, Remenyi et al., 1998; Easterby-Smith et al., 2002) and in marketing (for example, Carson et al., 2001). Instead of complying with the traditional article format of having a number of references to back up whatever I say, I have taken the liberty to merge literature with my own experience, and furthermore to mix philosophical issues with technical, handbook-like instructions. It is not until you have tried methods in research practice that they become a living reality.

Qualitative research in marketing 311

EJM 39,3/4

312

Strategies for qualitative analysis/interpretation In listing strategies I have not attempted to make a clear separation between analysis and interpretation as the two partly overlap: . Causality and understanding. Quantitative techniques are mostly used to try to pinpoint causality, usually between two or a few variables where the independent and dependent variables are defined. Qualitative methodology is primarily directed to understanding the complex and the elusive in a systemic perspective more than to establish unambiguous cause and effect relationships between single variables. It is built on the notion that life cannot be broken down into well-defined constituent components. Reality is more than the sum of its parts; there are also synergy effects. . Simultaneity and data generation. Qualitative research is characterised by data collection, analysis and interpretation in part taking place simultaneously, and tentative conclusions being drawn during fieldwork and the reading of archival records. I prefer the term data generation to data collection since data in a social environment do not consist of objects, which can be readily collected. Data are generated, meaning that they are the construction of the researcher. Even if data are taken from secondary sources such as statistical tables and annual reports, the researcher will have to assess and check their credibility, make a selection and combine them in formats chosen by the researcher, such as graphs or texts. Data are often created in interaction, for example, with a respondent in an interview. It means that even at this early stage of the empirical research, the researcher is engaged in analysis/interpretation; pure, value-free and complete description is the exception – if it exists at all. . Comparison. A key to qualitative research analysis is comparison. Data are compared with data, with existing theory, and with results from previous research. This continuous comparison is part of a sense-making process where patterns are formed and turned into concepts, categories and eventually theories. . Condensing data. To have access to an overwhelming richness of data is positive per se, but difficult when it comes to analysis. Therefore it is often said that we need to reduce data to make them manageable. The term is unfortunate as it may lure us to look for the average or the typical, thus rejecting variety and anomalies as exceptions. Instead we need to condense data, to make the same information more compact and manageable but not lose weight. . Transparency. A desirable property of scholarly research is that the reader can follow the actions and thoughts of the researcher. This may seem difficult to do with ordinary verbal language where verbal trickery may blur the real insights. A consolation is that formalised mathematical language is no guarantee for objectivity, and lying with statistics is a well-known and prevalent scam. Considering the large amount of data and links between them that are common in qualitative applications, for example, in case study research, as well as the researcher’s implicit paradigm, pre-understanding and tacit knowledge, it is often not possible to make all steps in the analysis transparent to the consumer of the research. This may feel uncomforting, but is a reality that we had better learn to live with constructively rather than sweep it under the carpet. It makes it

.

.

.

difficult, often even irrelevant, to try to fulfil requirements for reliability and replicability which is adamant in quantitative research and experiments. Techniques. The literature on qualitative analysis offers a smorgasbord of techniques to structure data, such as matrices and graphs (Miles and Huberman, 1994; Yin, 1994). Strauss and Corbin (1991) offer a stepwise approach to qualitative analysis inspired by grounded theory, and books on grounded theory give instructions for analysis (Glaser, 1978). Next to being supervised by a researcher who is experienced in implementing qualitative analysis, reading other researcher’s reports where the implementation of projects is accounted for is recommended. In Bryman and Burgess (1994) researchers tell how they implemented qualitative projects, and Glaser has compiled several readers of grounded theory projects (see, for example, Glaser, 1993). Computer software. Software can store data in an orderly way, provide structures and hierarchies of data, perform certain analytical tasks and respond to questions that the researcher puts to the data, thus facilitating the life of the qualitative researcher and increase research productivity (see, for example, Bazeley and Richards, 2000). Software assists, but does not take over the human researcher’s role as analyst/interpreter and the need continuously to fine-tune analytical/interpretive skills. This conclusion is just as valid in quantitative research which will be explained further with the help of “The research edifice” (see Figure 1). Intuitive and experiential – but systematic and rigorous. Rigorous research techniques need to be combined with such human qualities as intuition, common sense, experience, sound judgement, wisdom, insights, tacit knowledge, empathy

Qualitative research in marketing 313

Figure 1. The research edifice

EJM 39,3/4

314

.

.

.

and ethics. It is not only in exceptional cases that research can be irrelevant and erroneous and thus harmful if the human aspect is overlooked. In qualitative research the researcher remains the most important instrument and whether we acknowledge it or not, analysis/interpretation is dependent on the individual’s intuition. Intuition is often scorned at as anti-science, but good intuition means that we quickly process and synthesise data and draw conclusions based on the enormous masses of data and synapses (linkages) which are stored in our brains. Intuition can of course harbour bias, but that is also true with other techniques for analysis/interpretation. In our daily efforts to get on with our lives, we keep interpreting events. What then differentiates research practice from mere personal day-to-day practice? The difference is rather a matter of degree and transparency. It is high time to merge the systematic and overt procedures of western science with the methods of handling experiences. All is data. Hermeneutics in a broad sense can be applied to all types of data. In the flow of spoken and written words and numbers, symbols, observations, feelings and thoughts, interpretation becomes an integral part of a marketer’s daily life. Marketing hermeneutics can help us find meaning and it mirrors what a marketer does in his or her practice. Hermeneutics is also concerned with the interpretation of subtle expressions of human life, where the researcher tries to transform tacit knowledge into words. Nonaka and Takeuchi (1995), with reference to Japanese industry, have stressed the importance of making tacit knowledge explicit. In comparing the US and Japanese knowledge strategies, Cohen (1998) notes that the Japanese focus is on tacit knowledge and the nurturing of knowledge cultures and communities instead of as in the USA on explicit knowledge, projects and measurement. Linestone and Zhu (2000) note that the combined perspectives of the East and the West can offer insights that could not be obtained by just applying one of the perspectives. Both perspectives are necessary to address the variety and complexity of systems such as corporations and marketing networks. By not only merging western science with western common sense, but also with eastern philosophy we can arrive at a higher state of knowledge than by just applying either of these. The hidden agenda. All research builds in part on assumptions and restrictions that established researchers share – intersubjectively, not objectively – through tradition or agreement. In the extension of interpretation, hermeneutic processes also embrace the researcher’s paradigm, pre-understanding, understanding and contributions to explanation. Pre-understanding is what we know about the phenomenon of study when we start out on a research expedition; understanding is the (hopefully) improved knowledge that we come up with as a result of the research. Explanation is usually claimed to require unambiguous cause and effect relationships established through numbers, but as business life is in many ways ambiguous, less definitive and more transient explanations are required in practice. The hermeneutic helix. This is an expansion of the hermeneutic circle; the helix version bringing more dimensions in focus. Going from a circle to a helix we introduce advancement, moving research frontiers upwards, not just staying on the same level. We move from pre-understanding to understanding on a higher

.

.

.

.

level; we move from parts to the whole and to parts again, but now with greater understanding; and we move back and forth between the substantive, specific level to the abstract and general level, each time stretching ourselves along the helix. There is thus interaction between what we knew and what we have just learned, between slices of data and a systemic whole, and the concrete and the abstract. We can only give meaning to parts if we can put them in a context, a theory. We interpret and re-interpret data in a continuous trial-and-error process of both theory generation and theory testing. Through this process, qualitative research is not just left to generate theory, but it also tests theory in the sense of constantly improving its validity. Evaluation of sources. Researchers in marketing should be more critical to data and their sources – but in a constructive spirit. Newspapers and gossip are unreliable sources that need to be double-checked, we are aware of that. It seems to be less understood that official statistics can be a pack of lies, defined for a specific political or business purpose and compiled under a series of assumptions that are not known to the user of the statistics. Interviews with CEOs, marketing directors, consultants and politicians are often misleading, as they are speaking for their sake. Annual reports and accounts as well as statements from financial analysts can be outright deception, not just in such spectacular cases as in 2002 in Enron, Arthur Andersen and many other companies throughout the world. The dazzling elegance of statistics and the equally dazzling eloquence of words can communicate almost anything that has been asked for by a client, a political party or a company. Without sound judgement and ethics, the rigour and systematic modus operandi of researchers, does not become knowledge but ignorance, albeit rigorous and systematic ignorance. Full report. Data should be accounted for in a transparent, thick, rich and complete way and contradictory data must not be left out. This strategy takes readers closer to reality, but a mere detailed description is not sufficient; the account must offer conceptualisation and condensation or the researcher has not contributed interpretation and meaning. Alternative interpretations. To increase credibility, the researcher should offer possible alternative interpretations and argue both for and against them. As marketing reality is often complex, there is a shortage of simple cause and effect explanation between well-defined independent and dependent variables; there are multiple, sometimes innumerable, possibilities. When causal explanations are straightforward, however, they should be dealt with accordingly. Privileged bias. Two privileges influence interpretation. One is the privilege to define the problem to be researched. In an ideal and free academy this should be the privilege of anyone, but in reality a PhD candidate is dependent on the preferences of the supervising professor and the research culture of a university department. The professor in turn is dependent on the preferences of evaluation and grant committees, companies and governments that control the money. The other is the privilege to interpret the research. Who draws the conclusions and who has the credibility or power to endorse a certain interpretation and publish it?

Qualitative research in marketing 315

EJM 39,3/4

316

Enter the research edifice “The research edifice” in Figure 1 structures analytical/interpretive and objective/subjective phases, underscoring the omnipresence of the qualitative in research and the dependency of quantitative research on qualitative features. The qualitative bug is there; in fact, you do not have much choice as there is no pesticide to kill it without serious side-effects. It is a benign bug though, and you will find it to be a necessary part of the ecology of social life – once you accept it on its own terms and do not see it as a threat. Let’s sneak into the building through the basement. The basement. All research starts with a foundation of the researcher’s paradigm and pre-understanding. Here we make a mix of subjective, intersubjective and objective choices and assumptions such as what to research, which research questions to ask, how to find an answer, and in marketing that a market economy is better than a planned economy and that competition spawns innovation, productivity and low prices. These are mainly qualitative assumptions representing our interpretation of the world. They can be very personal but also be embedded in a specific research culture and discipline and be influenced by objective knowledge. The middle floors. If we press the lift button to the middle floors of the research edifice, we find ourselves confronted with data generation and analysis/interpretation. If the researcher chooses to do a quantitative study, a goal is to be explicitly systematic and rigorous, with little subjective influence on the research by the individual researcher or the team. Accompanying this process are, however, both intersubjectively agreed assumptions and choices, and subjective judgement call. On these floors, data should also be conceptualised and compared to theory and other research. If the goal is theory generation, the researcher has to conceptualise the data and compare them with extant models and theory, or they remain mere description leaving the finishing interpretive touch to the reader and listener. If it is a quantitative survey the questionnaire and its scales, response alternatives, sample, the collection of incoming questionnaires, and the registration and analysis of data can be rigorously handled, but still give rise to judgement calls and interpretation. One example of pseudo-objectivity in data generation and analysis is offered by surveys in which questionnaires are sent to businesses to find out, for example, how their marketing planning is done, how their customer relationship management (CRM) systems are working, how their key account managers operate, or what product or service development they will offer in the future. A lower and lower response rate of 10 per cent or even less is currently considered adequate – but only intersubjectively, not objectively. The resulting slices of data are accepted as amenable to statistical treatment. The explanation for accepting low response rates is that: . it is difficult to get businesses to reply; and . if you send out enough questionnaires you get a lot of data to process anyway; and that there is (after some additional checks) nothing to indicate that the data are skewed and not representative of the studied population. If instead qualitative case study research is performed, there will usually be little numbers and statistical conjecture (note, though, that case study research does not exclude quantitative elements). By doing case studies and applying various data generating techniques – all the way from reading texts to in-depth interviews, ethnographic observation and action research – complex and ambiguous issues can be

penetrated. The rigour may not be as obvious as with quantitative methods but it can still be substantial. However, this kind of research would require much more personal and subjective interpretation of data, a necessity in order to attract and accept the wealth of data that complex, chaotic and unpredictable markets offer. The penthouse. The last stop is the top floor, the penthouse. We are now as close to the sky as we can get in research. Here the research data, results and conclusions are presented, traditionally in written, oral and visual form. Even if it is pure academic research and the focus is on balanced and reflective conclusions, it might also include more speculative, innovative results. If the research is focused on practical action, the researcher could make recommendations and then those concerned have to make decisions, execute decisions, monitor the outcome, and make amendments. Whether the research is aimed to theory generation or to consulting to help solve a specific problem, interpretation is required. There were no simple, exclusively objective formulas to be found in the basement, and the dilemma returns in the penthouse. In fact, the results from our current research should ideally contribute to a more solid foundation of the research edifice and become part of the basement of future research. In summary, Figure 1 shows that the totally systematic and objective pursuit of the truth is a myth. The systematic and objective part is only a fraction of the research, albeit sometimes a pivotal fraction. In this sense it is immaterial whether research is qualitative or quantitative. Generation of marketing theory Four observations in particular have prompted me to bring theory generation into focus. First, theory generation is more often the outcome of a conceptual and qualitative process, whereas theory testing is more associated with empirical, quantitative hypotheses testing. Strange as it is to me, hypotheses testing has been placed on the highest level of scientific excellence in social sciences, including academic work in marketing. Second, researchers in marketing seem to settle for theory on a low level of abstraction or generality and have difficulties seeing the broader, systemic context; the core of a phenomenon is obscured by details and fragments. Third, researchers seem to get stuck in the middle, neither being firmly based in real world data, nor reaching a sufficient level of abstraction. Fourth, researchers are also stuck in the past and an unwarranted notion that theory advancement is a cumulative process. Marketing theory must reinvent itself and be refined, redefined, generated, and regenerated – or it will inevitably degenerate. Much of marketing knowledge resists time and change and should be retained; much is obsolete or was never up to par and should be dumped, and much is not yet discovered. This is in line with Kuhn’s (1970, p. 169) statement that new paradigms “usually preserve a great deal of the most concrete parts of past achievement and they always permit additional concrete problem-solutions besides”. This, however, is misunderstood by many. A paradigm shift, even if the new paradigm includes elements from the past, is not an incremental change based on the bulk of previous literature; it is a new foundation. It is the job of the scholar to separate the chaff from the wheat and not keep the established or accept the new just because it appears in certain journals or comes from the writings of currently popular “masters”. Even if the new is contrary to our beliefs, vested careers and interests, and poses a threat and initial frustration, we should receive a paradigm shift with enthusiasm rather than with apprehension. Schumpeter’s

Qualitative research in marketing 317

EJM 39,3/4

318

(1950) famous demand for creative destruction through innovation seems universal and applicable to marketing theory, too. My general stance is that theory in marketing is not really developing beyond conventional concepts, piecemeal models, current events and success stories (Gummesson, 2002a). The outcome is a world of “textbook theory“ that is taught in business schools. Textbooks are too quick to jump on the hype, for example dotcom, e-shopping, and software-based CRM. They are too slow to merge – not just list or add in a paragraph or note – fundamental, long-term developments, for example, services marketing, business-to-business (B2B) marketing, and the philosophy of relationship marketing, with already existing marketing management knowledge. Research in marketing is often limited to collecting standardised data on consumers, competitors and others without getting beyond statistical or verbal description. To earn the label “marketing intelligence” the research must include informed interpretation, conceptualisation and theory generation. To me it stands out as increasingly important that researchers help to interpret what is happening in the market and in the application of marketing strategy, not least in today’s high-velocity economy. Marketing metrics with relevant indicators can assist, but they are primarily historical and rarely self-sufficient and future-oriented. Their predictive capacity rests on certain assumptions, and qualitative interpretation is indispensable. We need theory to condense reality to something comprehensible so that we can adapt and manage. That is what theory does – when it is good. I should like to draw special attention to two key issues here: conceptualisation and contextualisation. They are interwoven and stress different aspects of theory generation. Concepts are needed, and in times of major changes new concepts – reconceptualisation – are urgently needed. There is a common mix-up between the term, which is a label, and a concept, which is an idea and the content behind the fac¸ade provided by the label. This mix-up is referred to as reification, mistaking the representation of an object for the object itself. Re-labelling is not the same as re-conceptualisation although branding and re-branding are used in marketing to create perceptions among customers about the content of a product or service, sometimes linked to improvements, but sometimes without altering the content. A postmodern marketer, though, may claim that the customer’s perception of the brand is the reality and that the real reality is unreal. An old term in marketing theory gives the reader associations to the old, even if the content is revised or new. For example, to understand services it is not sufficient to compare it to goods, which has so far been common (Lovelock and Gummesson, 2004). Services are treated with the use of goods terminology, and are defined in comparison with goods. This does not give room for understanding services on their own premises. It turns out, however, that it can be difficult to introduce a new term to represent a new concept and it seems quite accidental if the term is accepted or not. Contextualisation refers to the need to place single data in a broader context, that is, generate theory. Theory orders data in a context. A theory is a roadmap and a good roadmap makes it possible to navigate in a territory that is unknown to the traveller. When conceptualisation has taken place but not been extended enough to provide a context, bits and pieces of knowledge risk to be misunderstood and misused. Customers do not live isolated lives and consume one service or product at a time as if these were stand-alones. Customers live in a context. They consume a mix of many

goods and services in all sorts of combinations. If I am treated well in a store I am likely to go back. If the taxi driver to the hotel was unpleasant, I will be in a bad mood entering the hotel and that will affect my behaviour and my perception of the hotel. Theory generation, moving from raw data and description to conceptualisation and contextualisation, may be the most valuable contribution a scholar can offer. As researchers in marketing we are rarely if ever innovators; we rather start out as observers and messengers. However, it is not enough to be reporters of events. We have to add value to the phenomena we present; that is what scholarship is all about. One of Glaser’s (2001) latest books is specifically concerned with the move from description to conceptualisation. The further discourse is linked to Figure 2, a structure of the route from mere description to conceptualisation, from slices of raw substantive data to contexts of increasingly abstract, formal and general theory. The ensuing text will take the reader through the figure. The general framework of the figure starts at the bottom with real world data and a process to generate substantive, specific theories within a limit area of application. In the next phase, these areas are synthesised to a higher level of abstraction and theory of more general application is tentatively formed. The generated theory is compared with other theories that the researcher considers have an affinity or can add a contribution. The process goes on in alignment with the hermeneutic helix towards higher abstraction and generality. For the purpose of illustration, the general framework is clad in my own research in marketing, where the heavy parts are marketing management with services marketing and relationship marketing/CRM at its core, and with input from other relevant research. It starts with service research based on data from different service areas. The data are categorised and conceptualised, forming the beginnings of specific theories: a theory for the marketing of professional services, another for e-services, and so on. These theories in turn are merged into a general theory of services marketing, which later becomes one of several inputs to a theory of relationship marketing/CRM. This is en route to a general theory of marketing and a theory of management. However, the journey as such may be the significant result as long as the journey is guided by vision and commitment. To add to the realism of research, many iterations in the spirit of the hermeneutic helix may be required. If you start inductively with the raw real world data and move on the way up to becoming more general, you will have to dive down again to connect with the descriptive level. There is an ever-present risk that pre-defined categories may not be adequate because they are not properly grounded in the real world and thus are not a valid representation of the studied phenomena. If you start on the abstract level, you have no lifeline to the ground. Soon you hover freely in thinner and thinner air. Unfortunately, when theories are not well grounded, going back and test is not enough; because of the preconceived hypotheses, concepts and categories the test may just become a self-fulfilling prophecy. Is research in marketing really climbing a generality ladder, or does it stay put on the substantive level? Maybe not even that. Because of too superficial contact with reality, it may just stay put on a medium conceptual level with little contact upwards and little contact downwards. What is called marketing management today has its origin in consumer goods marketing of packaged mass products in the USA. It moved up from

Qualitative research in marketing 319

EJM 39,3/4

320

Figure 2. The route from specific description to increasingly general theory

substantive data to a certain level of abstraction and claimed to be general marketing theory. From that point it tried to force, for example, services and B2B marketing into its theory, thus missing the substance of these areas. When services marketing and B2B marketing through research based on real world data, extracted relationships, networks and interaction as core variables, these areas were marked as special cases, despite the fact that they constitute the bulk of all marketing. Later relationship marketing and CRM where labelled a special case, whereas in my view they offer a paradigm shift (Gummesson, 2002b). By generating theory continuously and following the format of the hermeneutic helix there is no strict distinction between generating and testing theory. If the researcher is guided by a desire to understand more and improve theory continuously, the never-ending journey becomes a continuous test of ever-tentative theory. Invitation to scientific pluralism and individual methodologies-in-use In my view, no one could ever claim that he or she has found the golden key to knowledge and science. They are elusive phenomena. We have to respect that; that is their karma and lifestyle. We are therefore in an eternal dilemma, constantly having to ask the questions: What do we really know? What is it worth? How do we get to know more? This ongoing journey is a fascinating treasure hunt for many but for others it is the avoidance or assassination – a scary monster. We probably all would like law and order and simplicity and security. To get that, however, we will have to do away with the magic of life and convert science to a set of regulations. In the process we restrict the scientist – the discoverer and entrepreneur – to a controlled and manageable bureaucrat heeding pre-set routines. But new knowledge should not be a threat even if it temporarily might shatter our careers, received theories and beliefs. The genes of researchers differ and so do researcher experiences, incentives and motives. This variety of researcher personalities offers science a disparate army of two-legged, individual paradigms. To me this is preferable to what I perceive as progressively stronger power demonstrations from established mainstream professors, promotion and grant committees, quality and ranking agencies to beget a centipede paradigm where all 100 feet march in the same direction. With a pluralistic strategy we will develop a diverse community of individual “knowers”, offering a medley of “processes of knowing” and consequently a variety in what is “known” (to use the Samitha concepts from the ancient Vedic culture; see Gustavsson (2003) and Zhu (2003)). In my view, pluralism is the only way to keep science alert, innovative and entrepreneurial. In marketing terms, it is the same as saying that a free market of science is likely to function better than a planned command economy. Interactive research: a two-legged paradigm illustration Here is an example of my personal choice of methodology – a two-legged paradigm. It is an example and not the doctor’s orders; use it if you find it appealing or compose your own methodological symphony (for a fuller account, see Gummesson (2001)). My experience of marketing and research is derived from a wide range of roles: being an around-the-clock consumer in a commerce-centric society; doing scholarly research and reading that of others; doing market research for companies as a consultant; designing and implementing marketing strategy; being a buyer of market research data and strategy advice from consultants; using market research data as

Qualitative research in marketing 321

EJM 39,3/4

322

input to decisions and action; and seeing how market research enters management processes, not seldom as part of internal politics rather than for rational decision making, planning and action. From these experiences I have arrived at a synthesis of research that I have found to be meaningful for me. I have chosen the label interactive research for my current methodology-in-use, the reason being that interaction has stood out as its core variable. Progressively my mindset has adopted a view that life, including both marketing methodology and theory, can be seen as networks of relationships within which interaction takes place. Instead of searching for strict and partial causality, I search for the understanding of a systemic whole, a context, with individual and complex patterns of interactive relationships. Important guidelines throughout my research are close access to reality – which requires researcher involvement – and validity of results. Interactive research embraces five components, which will be briefly reviewed in the next paragraphs. Case study research. This is my overarching approach. In case study research one or several cases are used to arrive at specific or general conclusions about certain phenomena, recognising the multitude of variables, complex interrelations and ambiguities of business life. Case study research provides the researcher with an input of real world data from which concepts can be formed and propositions and theory can be tried. A case study could be primarily inductive where the case provides data for conceptualisation and theory generation, or primarily deductive where cases are used to test existing theory. Case study research is systemic and holistic, aimed to give full and rich accounts of the relationships and interactions between a host of events and factors. It becomes largely interpretive. Quality criteria for quantitative studies, such as reliability and representativeness, cannot in general be applied to case study research. For example, a general rule for the number of cases needed to draw conclusions cannot be set up. The sample is theoretical and purposeful, looking for cases that give maximum information, and preferably be guided by saturation, the point where no or little new information is added. A single case study of a successful launch of new technology not only helps us understand the specific case, but also can teach us general lessons about marketing. For an extensive checklist of quality criteria for case study research, see Gummesson (2000, pp. 185-8). Inductive research and grounded theory. Simply put, inductive research lets reality tell its story on its own terms and not on the terms of extant theory. I have found inductive, empirical research a great source of inspiration, especially the strategies of grounded theory as developed by Glaser and Strauss (see Glaser, 1978, 2001; Locke, 2001). Grounded theory strives to be realistic and valid. To start a theory generating research project by first designing clear-cut categories and criteria in a complex and dynamic domain like marketing can severely impede our perception. But isn’t the researcher’s paradigm and pre-understanding – the basement of the research edifice – the capital of the professional’s work? It may seem odd to ignore existing knowledge to be able to receive new knowledge; we are used to hear that knowledge is cumulative and that what we do must have support in previously published journal articles. Viewing all knowledge as tentative, however, researchers have to train themselves to listen to reality without preconceived ideas. At a later phase the results can be

compared with existing concepts and theory and will thus proceed as an interplay between the inductive and the deductive. Grounded theory concepts and guidelines are clearly underused in marketing. All are not necessarily unique but they have been coherently ordered and reached a high degree of completeness in combining theoretical sensitivity, memos, comparative analysis, theoretical sampling, saturation, open and selective coding, the identification of core variables, and the generation of specific and general theory. Marketing anthropology. The prevailing qualitative research strategy in marketing is verbal data from personal interviews and surveys. In anthropology/ethnography data are generated through direct or participant observation, supplemented by interviews and conversations. The research is systematic and in-depth, documented not only in field-notes, but also in photos, films, audiotapes and artefacts. To make anthropology operational in a business setting, an adapted variant could be called corporate anthropology, or for marketing-specific applications marketing anthropology. Characteristic of true anthropology is the long periods over which a culture is studied – several months or years – as compared to the minutes or hours allocated to interviews and self-administered questionnaires. When a company and market is local and accessible, the researcher can be reasonably present and register what happens. Companies that operate in many locations and time zones require careful selection of the points of observation. Their marketing and sales staff is more often on the road or in the air, in hotels, and at their client’s premises, than they are at their home base. In consumer research – for example, to find out how people really eat breakfast or watch television – the true and complete behaviour can only be observed by being there, or let cameras act as “spies”. Marketing action research. In a job we get professional experience, perhaps 40 hours a week. As consumers we live in a marketing laboratory 24 hours a day. Action research entails dialogue and reflection based on data from experience through active involvement in the process being studied. It is akin to introspection found in postmodern consumer research and self-ethnography is an introspective and reflective extension of participant observation, but its scope is broader. We are actors, we make decisions and we have to face the real life consequences. But data formed through experience are sparsely used as empirical input in research, mainly because the method is unknown or off putting to the conventional researcher who favours detached research and is suspicious of subjective data. Action research gives rise to reflection, an inner dialogue. To paraphrase terminology from CRM: through systematic research based on personal experiences we can build data warehouses and models for future datamining. Involvement gives better access to data than detached research. To claim that involvement and the use of the researcher’s personality and subjectivity makes the researcher corrupted by the phenomenon under study and thus biased, and that the detached researcher is objective and uncorrupted, is no more than a subjective assumption. The scientific quality of subjective reflection and introspection is dependent on the researcher’s personal maturity and professional skills. The experience of a single researcher, however reflective and introspective that person may be, is usually not sufficient, although it has the advantage of datamining from close access. Data could be mined parallel to a dialogue with others and their experiences. The term action research is reserved for situations when researchers assume the role of change agents of the processes and events they are simultaneously studying.

Qualitative research in marketing 323

EJM 39,3/4

324

We have that role both as consumers and when working in organizations. Management action research is an application to the study of business phenomena, and a subdiscipline could be named marketing action research. The action researcher does scholarly research and is both an academic researcher and either a marketing practitioner or an external consultant. His or her purpose is twofold: to contribute to science and to help solve a practical problem. By being involved, the object of study creeps under the skin of the researcher in a way that is not possible in the study of documents or in interviews, even in participant observation. The access is as close as can be, and tacit and embedded knowledge can be uncovered. Further discussion on action research applied to management and marketing is found in Gummesson (2000), Coghlan and Brannick (2001), and a special issue of European Journal of Marketing (Perry, 2004); and for a review of introspection research, see Wallendorf and Brucks (1993). Narrative research. Narrative research is concerned with the ways “. . . in which social actors produce, represent and contextualise experiences through narratives” (Coffey and Atkinson, 1996, p. 54). Narratives are accounts – stories – about experiences, and they can take many forms. There is usually an initial state of affairs, then actions and events occur and there is perhaps a plot, and there is an end, at least a temporary end and more rarely the definitive ”and they lived happily ever after”. Narratives can be chronological but can also weave a web of events around various themes or concepts: story-telling. If marketing anthropology could be documentary that intimately reveals decisions, conflicts and actions – we even find this as entertainment in soap operas these days – narrative research is more of an epic film. By presenting research as a story, we avoid the fragmentation that is inevitable when we break down networks of events into abstract concepts and categories. Minett (2002) consults with B2B companies using cases of successful marketing to communicate with the general public and specific trades through the media, but also to help a selling company better understand its marketing. These stories must be told in a readable and condensed way or they will not get published, nor get read by practitioners. They become story-telling and an informed interpretation of reality, but not fiction. Story-telling in the research sense has close affinity to investigative journalism but with proper research and efforts to pin-point the essentials of reality, it becomes both a marketing tool and an input to general understanding of marketing. From a scholarly perspective such stories must of course be critically scrutinised as they are published in a company’s self-interest. But so must interviews and other data, not least official statistics, provided to researchers from companies or governments. Summing up Interactive research represents various interactions, such as between the researcher and the object of study and its actors; between your consciousness and qualities of your inner self; between substantive data and general concepts; between the parts and the whole; between words, numbers, body language and tacit language; between the researcher and audiences; and between data, analysis, interpretation and conclusions. The elements of interactive research all strive to reach a high level of validity. Interactive research perhaps only operationalises the best of common sense and sound judgement, adding a more systematic and transparent quality to everyday experience. It encompasses a string of strategies that are less recognised

internationally in business schools although they are well established in social sciences. Especially case study research and grounded theory are frequently used in Northern Europe within an eclectic Nordic School tradition, where syntheses from marketing and management disciplines are reflected in theory generating efforts in services marketing, B2B marketing and relationship marketing as illustrated in Figure 2 (see also Gro¨nroos, 2000). It is easy to criticise this personal package of methods – especially if the criticism deploys criteria from a quantitative paradigm. Case study research is called anecdotal evidence as it does not build on randomised samples; inductive, grounded theory research has no support in previous, accumulated research; action research and anthropology/ethnography are subjective and biased because of the individual researcher’s involvement; and narrative research is journalism or novel-writing. However, there is no need for the qualitative researchers to worry and feel forced to justify their research against a norm that has been set up by quantitative research; criticism can be raised against all methods and techniques. The crucial directive is awareness of strengths, weaknesses and relevance of what the researcher does. This has been an invitation to each individual researcher, research team and marketing department in business schools to consider multiple research paradigms. To use marketing as a metaphor for research in marketing: allow a market economy in research in marketing – a two-legged paradigm – and do not force on its members a planned economy – the centipede paradigm – with its standardisation and bureaucracy. In conclusion If we want marketing to be a productive and honourable discipline within management, both from a scholarly, practitioner and societal point-of-view, we need to use all our senses including both cognitive and emotional abilities. If we philosophically accept that there is a marketing reality, we will have to adjust our methods to understanding this reality as best we can. What we cannot afford to do is to say that the reality does not fit the mainstream paradigm and its techniques. All reality is amenable to scientific research and all efforts to focus just on a limited number of researcher-friendly issues will impede the development and usefulness of research in marketing. True scholarship in marketing is engendered by a committed entrepreneurial researcher spirit. References Alvesson, M. and Willmott, H. (1996), Making Sense of Management, Sage, London. Bazeley, P. and Richards, L. (2000), The NVivo Qualitative Project Book, Sage, London. Brown, S. (1993), “Postmodern marketing”, European Journal of Marketing, Vol. 27 No. 4, pp. 19-34. Bryman, A. and Burgess, R.E. (Eds) (1994), Analyzing Qualitative Data, Routledge, London. Carson, D., Gilmore, A., Perry, C. and Gronhaug, K. (2001), Qualitative Marketing Research, Sage, London. Coffey, A. and Atkinson, P. (1996), Making Sense of Qualitative Data, Sage, Thousand Oaks, CA. Coghlan, D. and Brannick, T. (2001), Doing Action Research in Your Own Organization, Sage, London.

Qualitative research in marketing 325

EJM 39,3/4

Cohen, D. (1998), “Toward a knowledge context: report on the first annual UC Berkeley forum on knowledge and the firm”, California Management Review, Vol. 41 No. 3, pp. 22-39. DeBono, E. (1971), New Think: The Use of Lateral Thinking in the Generation of New Ideas, Avon Books, New York, NY. Easterby-Smith, M., Thorpe, R. and Lowe, A. (2002), Management Research, Sage, London.

326

Feyerabend, P. (1975), Against Method, Verso, London. Glaser, B.G. (1978), Theoretical Sensitivity, Sociology Press, Mill Valley, CA. Glaser, B.G. (1993), Examples of Grounded Theory: A Reader, Sociology Press, Mill Valley, CA. Glaser, B.G. (2001), The Grounded Theory Perspective: Conceptualization Contrasted with Description, Sociology Press, Mill Valley, CA. Gro¨nroos, C. (2000), “Relationship marketing: the Nordic School perspective”, in Sheth, J.N. and Parvatiyar, A. (Eds), Handbook of Relationship Marketing, Sage, Thousand Oaks, CA. Gummesson, E. (2000), Qualitative Methods in Management Research, 2nd rev. ed., Sage, Thousand Oaks, CA. Gummesson, E. (2001), “Are current research approaches in marketing leading us astray?”, Marketing Theory, Vol. 1 No. 1, pp. 27-48. Gummesson, E. (2002a), “Practical value of adequate marketing management theory”, European Journal of Marketing, Vol. 36 No. 3, pp. 325-49. Gummesson, E. (2002b), Total Relationship Marketing, 2nd rev. ed., Butterworth-Heinemann, Oxford. Gustavsson, B. (2003), “The nature and understanding of organization from a Samhita perspective”, in Das Gupta, A. (Ed.), Human Values of Indian Management – A Journey from Practice to Theory, Macmillan, Delhi. Habermas, J. (1972), Knowledge and Human Interest, Heinemann, London. Kuhn, T.S. (1970), The Structure of Scientific Revolutions, 2nd ed., University of Chicago Press, Chicago, IL. Linestone, H.A. and Zhu, Z. (2000), “Towards synergy in multiperspective management: an American-Chinese case”, Human Systems Management, Vol. 19, pp. 25-37. Locke, K. (2001), Grounded Theory in Management Research, Sage, London. Lovelock, C.H. and Gummesson, E. (2004), “Whither services marketing?”, Journal of Service Research, Vol. 7 No. 1, pp. 20-41. Miles, M.B. and Huberman, A.M. (1994), Qualitative Data Analysis, Sage, Thousand Oaks, CA. Minett, S. (2002), B2B Marketing, Financial Times/Prentice-Hall, London. Nonaka, I. and Takeuchi, H. (1995), The Knowledge Creating Company, Oxford University Press, New York, NY. O¨dman, P.-J. (1985), “Hermeneutics”, in Huse´n, T. and Postlethwaite, N.T. (Eds), The International Encyclopedia of Education, Pergamon, Oxford. Piercy, N.F. (2002), “Research in marketing: teasing with trivia or risking relevance?”, European Journal of Marketing, Vol. 36 No. 3, pp. 350-63. Perry, C. (Ed.) (2004), European Journal of Marketing, special issue: Action research in marketing, Vol. 38 No. 3/4. Remenyi, D., Williams, B., Money, A. and Swartz, E. (1998), Doing Research in Business and Management, Sage, London.

Saunders, J. (1999), “Quantitative methods in marketing”, in Baker, M.J. (Ed.), The IEBM Encyclopedia of Marketing, Thomson, London, pp. 85-99. Schumpeter, J.A. (1950), Capitalism, Socialism and Democracy, 4th ed., Harper & Row, New York, NY. Strauss, A. and Corbin, J. (1991), Basics of Qualitative Research, Sage, Thousand Oaks, CA. Van Maanen, J. (2000), “Foreword”, in Gummesson, E. (Ed.), Qualitative Methods in Management Research, 2nd rev. ed., Sage, Thousand Oaks, CA. Wallendorf, M. and Brucks, M. (1993), “Introspection in consumer research: implementation and implications”, Journal of Consumer Research, Vol. 20, December, pp. 339-59. Yin, R.K. (1994), Case Study Research: Design and Methods, Sage, Thousand Oaks, CA. Zhu, Z. (2003), “What is to be managed: knowledge, knowing, or knower, and does it matter?”, Working Paper, University of Hull Business School, Hull. Further reading Miles, M.B. and Huberman, A.M. (1992), Qualitative Data Analysis, 2nd ed., Sage, Thousand Oaks, CA. Stake, R.E. (1995), The Art of Case Study Research, Sage, Thousand Oaks, CA.

Qualitative research in marketing 327

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0309-0566.htm

EJM 39,3/4

The rise and fall of the Latin Square in marketing: a cautionary tale

328

Robert P. Hamlin

Received January 2002 Revised October 2003 Accepted January 2004

Department of Marketing, University of Otago, Dunedin, New Zealand Abstract Purpose – This paper aims to illustrate the strengths and weaknesses of experimental design and development in academic marketing since 1950. Design/methodology/approach – The paper does so by taking one experimental design, Latin Square, and describing its history and development within academic marketing in detail. Findings – The Latin Square is a powerful experimental technique that first rose to prominence in agriculture in the 1920s and has remained a key tool in this discipline ever since. The technique was introduced into marketing in 1953, and enjoyed a period of great influence and popularity until 1973, when it abruptly disappeared from the publications of the discipline. Careful investigation of the research record of this period revealed that its demise was due to increasingly poor application method that led to compromised results, combined with an erroneous assignation of superior capabilities to full and fractional factorials that occurred at approximately the same time. Practical implications – Two major implications arise from these findings. First, the discipline has incorrectly retired a tool that is still unmatched in some key research situations. Second, the errors that led to the technique’s demise led to the rise of other techniques that do not have the capabilities that many researchers appear to think they have. Originality/value – This is the first longitudinal historical case study of a single research technique that has appeared in print in a major journal, and it reveals aspects of the discipline’s approach to science that could not have been illustrated in any other way. Keywords Marketing models, Market research, Brand management, Country of origin Paper type Conceptual paper

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 328-350 q Emerald Group Publishing Limited 0309-0566 DOI 10.1108/03090560510581809

Introduction Many commentaries exist on the state of research in academic marketing. Brown (1996, 1997), for example, has written extensively in varying formats on his views of the nature and shortcomings of research and thinking in the discipline. He is not alone (Piercy, 2002). These commentaries take a broad view and discuss trends in thinking and broad methodological approaches. No commentator up to this point has provided a historical analysis of the state of research in the discipline that is based on a study of the development, application and demise of a single technique over an extended period of time. Such an approach does have advantages. Although the study of a single technique may appear to offer only a very restricted picture of events within the discipline, a single tight storyline can help to put larger movements into perspective in much the same way as the romance between Leonardo di Caprio and Kate Winslet was used to The author would like to thank Dr Sarah Todd, and Mr Damian Mather for their suggestions during the preparation of this manuscript.

scope and illuminate the events of the Titanic disaster in the film of that name. It can also stiffen up broad observations and conclusions with the starch of specific example. The Latin Square is in many ways the Titanic of the marketing world. As one of a range of experimental techniques introduced from agricultural science in the early 1950s, It enjoyed a period of fame and wide application in research that sought to establish “hard” causal relationships between specific marketing activities and specific consumer responses to those activities. During this period the technique was accompanied on its journey by an interesting crew of characters, many of whom are well known in the research world. Then, in the early 1970s, the Latin Square abruptly, and quietly, disappeared as a widely used academic marketing research technique[1]. Like the Titanic, the Latin Square had a fleet of very similar “sister ships”, including the full factorial, and the fractional factorial designs that now underlie the widely used techniques of conjoint analysis. Despite being very similar in nature, and sharing many of the Latin Square’s weaknesses, these related techniques did share its fate. In many ways the primary significance of the story of the demise of the Latin Square does not lie within the reasons why it perished, or in its possible rehabilitation, but the reasons as to why, and how, its sister techniques survived this period, and then prospered to the point that they now underlie the majority of experimental research performed within this discipline, and above all if they were in any way compromised during this process. This paper describes the progress of the Latin Square by means of an extensive historical case study. Like the Titanic, the Latin Square has been dead in academic marketing for a very long time, and good quality, recent material on its application is hard to find. As a consequence, the first half of this case study has been devoted to a discussion of what the Latin Square is, what it can be used for and how it should be used. The second half of the case study examines its history, its rise, and more specifically its fall from favour in 1972-1973. The response of many readers to a discussion of a technique that has been dead for the best part of 30 years would be “So what?” The answer is that a close historical examination via this single in depth case study reveals that the Latin Square was done to death by errors, omissions and poor disciplinary structure that persist in the foundations of our research practice to this day, and that may taint a large proportion of our findings. The history of the Latin Square offers an insight as to what might happen to the discipline, on a broader basis, if these problems are not addressed. Because of its subject matter, this paper focuses very heavily on classical experimental design, the acquisition of high quality data and the algebraic models and methods that underlie both of these activities. Owing to the subtlety of some of the issues related to the use and misuse of the Latin Square technique, the discussion can become quite involved in parts of this case. For readers more used to in-depth discussion of multivariate statistical techniques, this paper may therefore be a bit of a novelty – that is also an issue in market research today. The arena: where the Latin Square can be profitably applied Do consumers like this brand mark? Is this brand/product combination a good one? Will this charity’s name improve consumers’ opinions of our product if we pay them for the use of it on the package? Do consumers prefer square bars of soap or round ones? How much is this brand mark worth with respect to this product? These

The rise and fall of the Latin Square 329

EJM 39,3/4

330

questions are posed hundreds of times a day in companies around the world. The marketing practitioners that ask them are not after some general marketing law. They usually ask these questions with respect to a specific situation defined by a single product type and a single group of consumers. They are asking for an estimate of the main effect of an independent variable such as a brand or some other distinguishing cue on an aspect of consumer behaviour that is acting as a dependent variable. Such dependent variables include attitudes, willingness to purchase and willingness to pay a premium. The marketer may use a variety of techniques to acquire an answer to these questions. They may turn to surveys, until they discover that surveys cannot establish the validity or otherwise of the causal relationships that are central to these questions. This drawback is shared by the exotic survey based statistical techniques that rely on covariance matrices. Interviews may offer more detail, and may yield insights that are valuable in other areas of marketing research, but they still cannot establish a causal link between the application of a specific feature and any changes in consumer behaviour that may occur as a result. In order to establish the causal links that lie at the heart of these questions, the marketer has to adopt an experimental approach. Usually this involves observing, recording and comparing consumer reactions towards products with and without the features that are of interest. This approach has its problems. If the same products (with and without the feature) are shown to the same consumer they will become aware of the purpose of the research, and they may react in a way that they think is appropriate rather than displaying the reaction that the feature would normally give rise to. In order to avoid this problem, the marketer may show the two versions of the product to different consumers, and then compare their reactions in order to estimate the effect of the feature. However, if the two consumers are not similar any conclusions based on the differences of their reactions may be erroneous. The product itself, the vehicle on which the mark is presented to the consumer, could offer incidental cues that may distort the consumer’s reactions in a way that would also render the conclusions of the research invalid. In order to investigate questions of this type the practitioner requires an experimental pattern that can estimate the main effects of the independent variable (feature) that they are interested in. At the same time it must also control for the effects of two other extraneous or nuisance variables, these being the individual consumer (group) and the product or cue vehicle. Luckily the marketer has access to several experimental techniques that are capable of doing this. The three most important techniques are the full factorial, the fractional factorial (including conjoint analysis) and the Latin Square. In many cases the Latin Square is the most appropriate technique. It is capable of identifying the required main effects. With a minor modification it is also capable of determining if a significant interaction between these main effects is present. This simple determination is all that is required. If a significant interaction of any type exists the size and valence of main effects cannot be reliably estimated by any of these three techniques. Indeed the concept of main effects becomes meaningless if a major interaction is present (Banks, 1965, p. 155; Davies, 1956, p. 255). The other two techniques therefore offer detailed information on interaction that is not required by the researcher in these situations, and this superfluous information comes at a heavy

cost in efficiency (the number of observations required). The relative inefficiency of full and fractional factorial designs not only costs the marketer time and money; it also makes these techniques less reliable than Latin Square when applied to research of this specific type.

The rise and fall of the Latin Square

An example: examining the impact of the country of origin cue The impact of the country of origin cue on consumer evaluations of wine was examined using the Latin Square technique. A great deal had been written about the country of origin cue, and about the merits of the various wine-producing countries (Han, 1990, 1989; Robinson, 1994). However, nobody had tested the fundamental assumption that consumers paid any attention whatsoever to the country of origin cue when evaluating a bottle of wine. The research was performed in New Zealand and replicated in the UK (Hamlin, 1997; Leith, 1993; Hamlin and Leith, 1995a, b). The experiment was set up in the manner shown in Table I. The independent variables were the four country-of-origin cues (A1-A4): Chile, France, New Zealand and Australia. France had an established reputation for producing wine of all types. Australia had a reputation for good reds, and New Zealand had a reputation for good whites. At that time Chile was just emerging as an exporter of wine, and therefore had no established reputation in either New Zealand or the UK. The country-of-origin cue was presented to the consumer as a simple name on the cue vehicle, which was a bottle of wine. The first set of extraneous variables was the four groups of consumers (B1-B4): each group consisted of 30 consumers recruited from the same liquor store on the same day of the week, and at the same time of day for four successive weeks. The objective was to acquire four groups of consumers who were homogenous in their lifestyle and drinking habits. The Latin Square can control for small variations within the consumer groups as an extraneous variable, but this capability is not extensive. Increasing variations in the extraneous variables will eventually create interactions, which will mask the significance of any main effects in the analysis of variance. The very small size of the experiment (4 £ 30 ¼ 120 subjects) was due to the efficiency of the Latin Square, and to the practice of blocking on consumer groups. It has been mentioned that this efficiency is a major advantage of the Latin Square[2].

331

C1 C2 C3 C4 “Belleville” bottle “Maitres” bottle “Vesselle” bottle “Villiere” bottle B1 30 consumers, week 1 A1 Chile

A2 France

B2 30 consumers, week 2 A2 France B3 30 consumers, week 3 A3 New Zealand B4 30 consumers, week 4 A4 Australia

A3 New Zealand A4 Australia

A3 New Zealand A4 Australia

A4 Australia A1 Chile

A1 Chile

A2 France

A1 Chile

A2 France

A3 New Zealand

Table I. Latin Square arrangement for investigation of effect of country of origin on consumer attitudes towards bottled wine

EJM 39,3/4

332

The advantages of such efficiency include access. It is unlikely that the hosts for this research would have tolerated the researchers’ presence had they chosen to undertake the full factorial equivalent of this research with 1,520 subjects. Efficiency gives rise to other benefits such as reduced expense, speed in the acquisition of results, and increased security. The researcher can be there and gone before their competitors have a chance to observe the research. However, the major advantage of the Latin Square’s efficiency is increased reliability via the reduction of uncertainty. Unlike survey samples where bigger is almost always better, the instruments of parallel comparison that include all factorial and Latin Square designs work best on a minimum sample. The reason for this is that they rely on the assumption of a homogenous experimental population and environment – that is the variations in terms of independent and extraneous variables introduced by the researcher are the only variations in the experimental population and environment (Fisher, 1950, p. 510). It is generally much easier to assure homogeneity in a small experimental population than a large one. As the sample gets larger so variations will begin to occur in the experimental population. Random variations in the experimental population will reduce the significance of any main effects. Systematic variations will produce erroneous results. For example, if a full factorial design had been used, the experimental population would have to have been drawn from several sites, or from several time slots each week, or over a much longer time period. All of these, and especially the last one, could have led to unpredictable variations in the experimental population and environment that could not have been identified or quantified by the researcher. A Latin Square with 30 consumers to each cell would require an experimental population of 480. By “blocking on consumer groups” this can be reduced to 120. Blocking on consumer groups is a procedure by which each consumer group is treated as an extraneous variable, and contributes to multiple cells of the experiment[3]. It is a procedure that is unique to marketing, and it does not occur in agriculture where these techniques were developed. It is seen in marketing in work by (Asam and Bucklin, 1973). Blocking on consumer groups increases the efficiency of the Latin Square design, with all of the advantages described above, and allows the researcher to check that differences between the groups/locations do not give rise to significant main effects and interactions. It has one drawback in that the researcher is not free completely to randomise their treatments among the experimental population. They are constrained to show a specific group of cue/vehicle combinations to each consumer group (see Table I). As a result, if it is discovered that variations in the consumer groups do give rise to significant main effects, the estimation of other main effects may be highly suspect if a test for interaction has not been conducted. In practice the very small experimental populations required by Latin Squares blocked on consumer groups, coupled with careful design of the research, should reduce the likelihood of such an occurrence to a very low level. The incorporation of a test for interaction in the design phase of the research means that even if these main effects do occur, they can be discounted if no interaction has occurred. The second extraneous variable, the cue vehicles (C1-C4), consisted of four very high quality mock-up/fake products[4]. Each fake product was a bottle of varietal wine made in a respected wine growing area by one of four non-existent producers, Belleville, Villieres, Maitres and Vesselle. These fake products offered very specific

challenges to the researcher. As with the consumer groups, wide variations in the products were highly undesirable due to the increased risk of interaction. However, a certain degree of variation within the cue set presented by the vehicles was essential, as each group of consumers would see all four levels of the independent variable (A1-A4). If there was too little variation in the vehicles, the variation in the independent variables would be obvious to the consumer, and their responses would be compromised. Variation in the cue set presented by the four vehicles was also essential to test the stability of the impact of the independent variable (country of origin) to variations within the vehicle on which it was presented to the consumer. The fakes had to be plausible as a product of any one of the four countries whose name would be applied to them. They had to look right, not only in terms of the quality of the labels and the bottles, but also in terms of the design and legal conventions the wine industry within each country. It is essential to research of this type which attempts to recreate the rapid “snap” evaluations made by consumers when buying fast-moving consumer goods (FMCG), that the fakes are of a quality that allows the researcher to present them to the consumer as a real product that they wish to evaluate. Trying to use a sub-fake quality mock up in such research is a complete waste of time. It invariably involves lengthy explanations to the consumer as to what they are expected to do. These discussions resemble explaining a bad joke to someone who does not understand it, and who does not find it particularly amusing when they do. The consumers’ bewilderment/contempt leads to incorrect/flippant responses. It is only in the last four years that technology has allowed the market researcher to produce labels of commercial quality within a reasonable price and time scale. As this is an artistic as well as a scientific process it helps if one is, or has access to, a competent graphic designer. The prototypes of the vehicles were designed and tested against existing products and each other using focus groups and in depth interviews of industry practitioners and consumers. At the time France represented real problems as it was only just starting to produce varietals[5]. The wine producing areas within each country had to be very carefully selected for equivalence. The four neutral labels shown in Figure 1 were eventually produced, and applied to (nearly) identical bottles of Chardonnay varietal wine. Four versions of each label carrying the names of the four countries of origin were produced (Figure 2). The dependent variable was the consumers’ evaluation of the product. This was recorded via a battery of 14 (1-7) Likert scales each dealing with a separate distinguishing characteristic of wine that had been generated by a factor analysis (Leith, 1993). A dependent variable of this type was used as the researchers assumed that the consumers’ selection of wine was a rational decision based on evaluations that followed the pattern described by the Theory of Reasoned Action (Fishbein and Ajzen, 1975). Consumers were asked to evaluate the wines by looking at the bottles and recording their reactions to it on the forms (no tasting sample was given). They were asked to evaluate it quickly in less than a minute, as they would in a shop, after which their evaluation form for that product was taken off them. Each consumer evaluated four products. An example of one of the Likert scales and the statistical analysis of it is given in Figure 3.

The rise and fall of the Latin Square 333

EJM 39,3/4

334

Figure 1. The four label variants used, set for “New Zealand” as a country of origin

This work was replicated for two varietals, Chardonnay and Cabernet Sauvignon, and in two countries, New Zealand and the UK. Very strong results were obtained, and it was possible to build up detailed profiles of the impact of the name of each country of origin on consumer evaluations of it using the 14 distinguishing characteristics[6]. Preliminary work on evaluating the “value” of the names by attempting to measure differences in consumer expectations of price also gave very encouraging results.

The rise and fall of the Latin Square 335

Figure 2. “Vesselle” label showing variations of country of origin cue

The issue of generalisability and interaction It is when this work of this type is presented to outside researchers that the serious problems with the Latin Square in academic marketing are revealed. A presentation of this work would almost invariably create negative comment in two main areas: (1) “This result is too narrow. The results are not generalisable, and it’s therefore not a significant contribution”. The issue of academic marketing’s obsession with generalisable results has been extensively covered by other writers

EJM 39,3/4

336

Figure 3. Example of analysis of data from a simple Latin Square experiment for one component of evaluation of a Chardonnay varietal

(Brown, 1996). This obsession with generalisable results has been particularly damaging with respect to the application of experimental instruments of parallel comparison such as Latin Square. Their reliance on small homogenous samples makes the acquisition of generalisable results impossible when using these techniques. The comments above are usually accompanied by a recommendation to introduce wide variations into the extraneous variables. For example a Latin Square set up with the same independent variables, but using automobiles, hair spray, wine and holidays, and four consumer groups of students, young adults, middle aged people and old age pensioners. The intention being to create a “generalisable” result for country of origin in New Zealand. It is very unlikely that such an experiment would produce a significant result. The main effects and interactions caused by the wide variations in the extraneous variables would mask any country of origin main effect that may exist. For example, the research described in the previous section was replicated for two varietals, Cabernet Sauvignon a “red” varietal and Chardonnay a “white” varietal. This minor change in one of the extraneous variables (cue vehicle) produced very different profiles of the dependent variable. Combining the two datasets to produce a consolidated generalisable result for “bottled wine” produced a weak and distorted result. The destruction of main effects by such a minor broadening of the design would indicate that the results produced by the recommended generalisable research would therefore not only be insignificant – they would also be erroneous. The conclusion would be that no country of origin effect existed, whereas in fact very strong and specific country-of-origin effects do exist in specific situations. There is nothing wrong with research designed to investigate very specific situations. Most commercial research follows this pattern. Situation specific results in academic research can also be useful. If a general theory is to hold good, then it should work in all situations. Testing such theories in highly specific situations represents a very aggressive/strong test of such theories. The approach corresponds to the deductive/falsificationist approach of Popper (1935), and the instruments of parallel comparison are well suited to it. For example, the dependent variable in this research relied on the assumption that consumer evaluations could be approximated to by using the Theory of Reasoned Action. This assumption held good in this instance. It did not hold good when this research procedure was applied to equally specific situations involving meat and food products. These highly specific results may well have generalisable deductive theoretical implications for the Theory of Reasoned Action. (2) “You’re using a Latin Square – everybody knows you shouldn’t use Latin Square because you can’t detect interaction. Can you prove that you don’t have interaction in this result?”. There appears to be widespread, although unpublished, view in academic marketing that the Latin Square is “Badthing”, and is to be avoided if at all possible. It is not an opinion that is shared by either industry or agriculture where the successful application of this technique is still widespread. It is perfectly true that the Latin Square cannot detect interaction. However, as it is a fully confounded design, a significant interaction usually has

The rise and fall of the Latin Square 337

EJM 39,3/4

338

the effect of eliminating significant main effects in the analysis of variance. Therefore if a researcher is reporting significant main effects, the chances are that they are not suffering from a significant interaction. However, they have a real problem if they do not have a significant main effect. If this happens, then they really do not know for sure whether there is nothing there, or if there is a main effect that is being masked by an interaction. While the researcher may be confident that their method has reduced the likelihood of interaction to an acceptably low level (p , 5 per cent), the current culture within academic marketing academic marketing requires a statistical test rather than a personal judgement. As a result, the work of any researcher that cannot furnish such a statistical test is unlikely to be published. There are two statistical tests for interaction in the Latin Square. Both of them are old, and they lurk in the biology/agriculture literature. Both attempt to test for the significance of interaction by dividing the error sum of squares in the analysis of variance table into a sum of squares for residual error and a sum of squares for non-additivity (general interaction). The first of them is due to the statistician John Tukey (Snedecor, 1955; Tukey, 1949). It attempts to calculate a term for the non-additivity sum of squares directly from the dataset of an unmodified Latin Square. Unfortunately it does not work. The reasons for this go beyond the scope or space of this paper, but the procedure by which a “k-array” is generated from the original dataset, and is then fitted back to it, is tautological (Hamlin, 1997). It will report that interaction is insignificant even if it is presented with datasets that are nothing but interaction. The procedure was introduced into the marketing literature by Kohli (1988), who gave an example showing significant interaction. It has so far proved impossible to reproduce this result using Kohli’s dataset and Tukey’s original equations. A rework of the example given in Tukey’s original paper reproduced his reported (insignificant) result. The second method by Youden and Hunter (1955) attempts to estimate the sum of squares for residual error, so that the sum of squares for interaction can be estimated by subtraction in the analysis of variance table. This method involves a modification to the experimental design, so it cannot be applied “after the event” to the data from an experiment. This is a drawback, but its major advantage is that it works, and it works well[7]. This technique relies on a minimal balanced replication (n cells) of an n £ n Latin Square: one cell in each row, column and treatment is replicated. The sum of squares for error is generated for the variations between these replicated cells (Figure 4). While this design is a fractional factorial, the replication is solely to find a general term for interaction. Therefore the authors’ (Youden and Hunter, 1955) description of their design as a partially replicated Latin Square seems reasonable. There is a slight loss of efficiency due to the replication, but this is more than outweighed by the fact that the researcher can now include a statistical test for the presence of interaction in papers submitted to journals. The history of the Latin Square in marketing Why has academic marketing come to regard the Latin Square as “Badthing”? The reasons lie in its history in the discipline, which is both an interesting and cautionary

The rise and fall of the Latin Square 339

Figure 4. Example of analysis of data from a partially replicated Latin Square experiment. Consumer evaluations of ready meals using meat group as an independent variable

EJM 39,3/4

340

tale. The Latin Square first appeared in a paper published by the mathematician Euler in 1782 (Euler, 1782). It remained a mathematical curiosity until, from 1925 on, R.A. Fisher described the Latin Square and other factorial designs as experimental techniques in a series of books (Fisher, 1970, 1950, 1937). Fisher’s work showed how agricultural research workers could use these instruments of parallel comparison to obtain relative rather than absolute empirical data with which to test their theories[8]. This represented a major breakthrough (Fisher received a knighthood) as the experimental environment in agriculture was too “dirty” to allow the researcher to obtain the absolute empirical data that their colleagues in the “hard” sciences were used to working with. These experimental techniques became the basis of agricultural research, and the Green Revolution was based on them. However, the techniques as laid down in writing by Fisher and others only formed a part of the story. Method, the ways in which the techniques could be reliably applied to research in the field, was a verbal tradition. Method was derived from the collective experiences of researchers in the field. Method was highly stressed in agricultural teaching and research institutions. However, texts on method were not available. As method was situation specific it was (and still is) impossible to write a sensibly sized text on the subject. In agriculture this was not a problem, as agricultural institutions tended to be large and only used a very limited number of experimental techniques. Large staff, a high throughput of experiments and the presence of dedicated experimental statisticians meant that experimental method was swiftly learned by researchers who were fresh to the discipline. In 1953 a group of agricultural researchers at the Field Research Station, University of Cornell, decided to apply the techniques by which they had been learning how to grow apples, to the equally experimentally dirty business of learning how to sell them (Brunk and Federer, 1953a, b). Their recommendations on display layout, based on a Latin Square experiment, were applied with spectacular results. The Latin Square and its sister techniques of parallel comparison entered the field of marketing. Large low order factorials and Latin Squares appeared regularly in the leading journals (e.g. Cox, 1964; Holland and Cravens, 1972; Hoofnagle, 1965; Kennedy, 1970; Sawyer, 1973), and a major text covering Latin Square and other agricultural style factorial and blocked experiments was published (Banks, 1965). However in 1973 the Latin Square virtually disappeared from the leading market research journals leaving only a few isolated examples (e.g. Currim and Sarin, 1983; Sheluga et al., 1979. Other agricultural style low order large sized factorials and blocked experiments became much rarer, and were replaced by small high order (2n) factorials, fractional factorials and later by conjoint analysis (e.g. Della-Bitta et al., 1981; Darley and Smith, 1993; Kent and Allen, 1994; Yadev and Monroe, 1993, etc.). There is no evidence that Latin Square became extinct through lack of a purpose. Research questions similar to those outlined earlier in this paper continued to be asked. However, different methods were used to address them. The extinction was so abrupt that it was unlikely to have been a matter of fashion. The extinction of the Latin Square appears to have been caused by two major factors: A slow deterioration in the method and technique of its application, and the sudden emergence of an alternative approach that was thought to be superior. The first of these processes is best demonstrated by an example. Dean et al. (1972) published a study in the Journal of Marketing using the Latin Square. The objective

was to evaluate the impact of the package copy claims “new”, “improved”, “new/improved” and “no claim” on housewives’ evaluations of products using a 4 £ 4 Latin Square. The researchers introduced wide variations into their extraneous variables. The first extraneous variable, the cue vehicles were facial tissue, hosiery, scouring cleanser and underarm deodorant. These could just about be defined as non-food FMCG products, but they were drawn from three very distinct product categories. The chances that these wide variations in the cue vehicles would produce significant main effects and interactions were high. The second extraneous variable was consumer groups, used as a blocking variable. These consumers contained a wide spectrum of users from heavy to non-users of the products, but they were not allocated to groups using these characteristics. The basis of allocation was not given. Introducing a wide between-cell variation in an extraneous variable was poor method. Deliberately introducing a major within cell variation, which would only contribute to the error term in the analysis of variance, is impossible to rationally justify. The analysis of this research was highly unorthodox. The correct way to set up the single analysis of variance table to analyse the Latin Square had been published (Banks, 1965). In this research an array of “one-off” analysis of variance (ANOVA) tables was used. In the first a single comparison between the within consumer groups and the between consumer groups sum of squares was made. The conclusion was that, as the sum of squares for within group variation (1,976) is so much larger that the sum of squares for between group variation (2.46), then the groups were equivalent. As the within groups sum of squares not only included random error, but also included all the variations that could be attributed to the main effects and interactions of all levels of the other extraneous (vehicle) and the independent variable itself (package claim) no such conclusion could be made. The fact that the researchers had already stated that they had organised their experiment so that the majority of all consumer variation occurred within the consumer groups makes the conclusion all the more unreliable. The analysis continued in this vein. The significance of the main effects of the independent variable was disposed of in a similar analysis of variance table. The main effect of the independent variable, package claim, was pitted against not only random error, but also the main effects of the consumer groups, the widely varying products being evaluated, and any interactions that might have been going on between all three of them. The main effect of package claim not surprisingly lost this contest for significance. None of the degrees of freedom in each table (which varied wildly) could be exactly related to either the number of participants, or the number of evaluations that they should have given. The entire analysis could not be related back to any summary data, and the conclusions drawn from it were therefore insupportable. The article appeared in the leading journal of the discipline. It was not challenged, and it was cited at least one text (Tull and Hawkins, 1987). The unavoidable conclusion is that by 1972 very few, if any, researchers in academic marketing had an adequate knowledge of how to use the Latin Square, either in terms of the method necessary to acquire good data, or the basic techniques required to make sense of it. This deterioration in application may have been due to a number of causes. One possibility is that the rate of research in terms of experiments performed is much slower in marketing than it is in agriculture. The range of research tools is so also much wider in Marketing, encompassing survey and related approaches. As a result, the “rate of use”

The rise and fall of the Latin Square 341

EJM 39,3/4

342

of the Latin Square may have been below the point where good technique and method could be acquired and passed on within the marketing research community. This situation spelt trouble for the Latin Square. The technique required expertise to acquire significant results. This expertise was clearly not available in academic marketing by 1972. The result would have been a very high rate of insignificant results. Researchers with an insignificant result from a Latin Square would have been very vulnerable to the interaction “killer question”. As there is no evidence that either of the two tests for interaction in the Latin Square were known in academic marketing at this time, the Latin Square may have acquired a reputation as a “weak” technique as a result. Nevertheless, such a process would have taken time. As the Latin Square suffered an immediate extinction, the rapid emergence of a superior alternative must be considered. It is not hard to see what this alternative was. It was mentioned in an earlier section that researchers presenting Latin Square based research are told that the Latin Square is “Badthing”. They are usually urged to use a factorial design as these can “deal with” interaction. This is not true. A factorial can give broad indication as to whether certain groups of interactions are present. However it cannot individually identify the size and above all the valence (positive or negative) of individual interactions. This is not surprising. A 4 £ 4 £ 4 factorial has 116 possible individual interactions, and these can be recombined into billions of unique interaction patterns for the design. Furthermore, even a 2 £ 2 factorial referenced to the experimental population mean is algebraically insoluble for individual main effects and interactions if interaction is present. If the objective of the researcher is to obtain an estimate of the size and valance of main effects, then the factorial is no more capable of “dealing with” interaction than the Latin Square. The ability of the factorial to deal with interaction under these circumstances exists only in the minds of academic marketing researchers. It is none the less potent for that. The important issue is when and how did it acquire this capability? “Smoking guns” are extremely hard to find in these circumstances, but in this case a page-by-page study of the two leading American journals of the day ( Journal of Marketing and Journal of Marketing Research) for 1972 and 1973 did reveal the two to three key sentences that equipped factorials and fractional factorials with this imaginary capability. In 1973 Holland and Cravens published an article in the Journal of Marketing Research titled “Fractional factorial designs in marketing research” (Holland and Cravens, 1972). This paper was one of the most influential ever published in a marketing journal. It appears to have caused the demise of classical agricultural experiment as the major type of experiment in marketing, and to have introduced the high order 2n factorials and fractional factorial/conjoint designs that have dominated academic marketing research to this day. The paper discussed the drawbacks of the existing designs in marketing, and singled out the Latin Square for particular attention. The paper then discussed high order (2n) factorial designs. It is in this section that the key misunderstanding appears to have arisen. The misunderstanding arose not from an error, but an omission. Holland and Cravens showed how the size and the valence of main effect and interactions could be calculated from the results of a 2n factorial. Their discussion owed much to a section in Davies’ (1956, pp. 249-73) text on experimental design. Regrettably neither Davies nor

Holland and Cravens noted that these calculations relied on one major assumption, that the “null condition” could be identified, and the experiment referenced to this null condition rather than a population average. Figure 5 shows the implications of this assumption for a 22 factorial. Unlike a factorial referenced to its population mean, a factorial experiment referenced to the null condition, usually defined as treatment (1), can indeed yield figures for main effects and interactions. However, it ceases to be an instrument of parallel comparison, and instead becomes a controlled experiment with treatment (1) acting as a control. The major assumption that is made when such an experiment is devised is that no main effects and interactions exist within that null condition. Such an assumption can be easily supported if the null condition is defined by the zero point of a set of treatments defined by ratio data (e.g. concentration of chemicals)[9]. It becomes shaky if the treatments are defined by ordinal data (e.g. high and low advertising). It becomes completely insupportable if the treatments are defined by nominal data (e.g. brands). Is the null condition of a brands/products layout, no

The rise and fall of the Latin Square 343

Figure 5. The null condition used to solve for main effects and interaction in a 2 £ 2 full factorial (calculations given for rows only)

EJM 39,3/4

344

Figure 6. The experimental implications of an arbitrary null condition

brand/product, the most popular brand/product or the most “ordinary” brand/product? In such a situation the choice of the null condition becomes completely arbitrary. As a consequence, the assumption that no interaction exists in the null condition becomes exactly the same as assuming that no interaction exists within the whole layout of a Latin Square experiment. Figure 6 shows the implications for this on an experiment based on such an assumption. It can be undermined by further work that reveals the presence of interaction within the previously selected null condition. The ability of the full and fractional factorial to definitively quantify main effects and interactions is therefore restricted to situations where all the variables defining a specific treatment are ratio data. This is a rare condition in marketing. If any of the variables are ordinal the assumption regarding the null condition that supports the

calculations is suspect, and if any of the variables are nominal it is completely insupportable. This does not seem to have stopped the full and fractional factorial designs from acquiring the general ability to identify and quantify main effects in the presence of interaction in the minds of marketing researchers. A factor that more than any other would have led to the demise of the Latin Square.

The rise and fall of the Latin Square

Conclusions The Latin Square is only one experimental technique among many. To say that it should be used in all situations would be as bad as saying that it should not be used at all. Even within areas where experimentation is the optimal approach, it has a narrow range of application, but this range includes some of the most important areas of commercial and academic marketing research. One area that has been illustrated by example is the impact of package and other point of purchase cues on low involvement consumer evaluation and decision making. Given how much money changes hands to acquire the “goodwill” that may or may not be represented by major FMCG brands that are associated with this type of purchase, the Latin Square has not only academic, but also commercial potential. Within this narrow arena, it is a technique of simplicity, elegance, accuracy and matchless efficiency. It requires care in its application, but it has no weaknesses that cannot be addressed by a conscientious researcher. Despite its poor reputation, and its lack of recent widespread use in academic marketing, there is absolutely no reason why a researcher should not reach for this particular tool if it suits their purposes to do so.

345

Implications Having used the Latin Square successfully and profitably in commercial agriculture for years before a change of discipline, the primary motivation of this researcher when this project began in 1994 was to rehabilitate the Latin Square within academic marketing to the point where it could be reapplied with some hope of the results being published. However, the research also revealed shortcomings within our discipline that are alarming to say the least. The general requirement for “generalisable” results has been noted earlier in this paper. There is no doubt that this pressure led to the progressive misapplication of the Latin Square and other instruments of parallel comparison up to 1973. My own experience of submission of research that seeks to test theory deductively by aggressive testing of specific “commercial-type” propositions indicates that this pressure has not reduced. This pressure will continue to lead to serious misapplication of the currently popular instruments of parallel comparison. This observation particularly applies to “stressed” fractional factorial designs which include wide variations in multiple independent variables in the search for a generalisable result/conclusion. Knowing, as I do, the extreme vulnerability of fractional factorials to interaction in the reference treatment, the widespread application within academic marketing of stressed versions this particular controlled experimental design – in various commercially disguised computerised forms – truly makes me tremble. Future directions The role of pressure created within a discipline as a cause of methodological misapplication cannot be overestimated. The philosophical error made by John Tukey

EJM 39,3/4

346

in his procedure for detecting interaction in the Latin Square appears to have been the result of pressure to somehow come up with an after the fact “fix” for badly designed Latin Square experiments. It is a pressure that is acknowledged in Youden and Hunters’ (1955) article. It shows that even the finest research researchers can fall over if they are pushed hard enough. As a discipline we urgently need to review our requirements with respect to generalisable experimental results in order to relieve this destructive pressure. We could begin by examining just what we mean by a generalisable result. Do we mean a result that can be directly extended to all situations? I, for one, do not believe that there is any such thing in marketing. The variations in consumers themselves, their environments, aspirations and problems, coupled with the variety of ways in which marketing practitioners address them combine to defeat any such generalisations. However, by continuing to stress our experimental designs beyond their capabilities in the search for such generalisable results and theories, a marketing E ¼ mc 2 if you will, we end up with results which are not only generally, but also specifically useless. Are their any pointers as to how we could relieve this pressure? Yes there are. Over the last 50 years, most of the designs that we use in marketing, including the Latin Square, have continued to be the basis for sustained theoretical and practical progress in agriculture, the discipline that developed them. How has this been achieved? The answer is that agricultural research, and particularly commercially-driven agricultural research proceeds largely by the testing of specific propositions tested with specific results derived from specific experimentation using unstressed designs. The theoretical limitations of these results are explored by equally specific replications if that is deemed to be necessary. There is a widespread acceptance that individual research results can only very rarely be extended to general theory and practice without extensive replication. The well-established progression within agriculture and the plant and animal sciences from laboratory trials to field trials to limited commercial testing and release is the most obvious expression of this philosophy. Regrettably for marketing researchers, if general results do not exist in the world of potatoes for example, who are simple souls, then the chances are that they are at least as rare in the world of people – and it is high time that we took note of this and modified our research and published reporting procedures accordingly. While the pressure for generalisable results appears to have been the major cause of misapplication of these experimental techniques within academic marketing, it is also necessary to consider how these misapplications survived the review process to get into print in our leading journals and to influence matters to the degree that they did. In order to understand how this can happen it is necessary to study a discipline where, unlike our own, malpractice has immediate and devastating consequences. In the medical industry, hospitals usually have to perform a certain minimum number of procedures a year before they are allowed to offer certain services (e.g. cardio thoracic surgery). Such centres are also the subject of stringent minimum staffing requirements with respect to the number and relevant experience of personnel on hand. This system specifically recognises the key role of “method”, the verbal tradition described earlier in this paper. You can read the technique of how to undertake a heart transplant out of a book, but it helps if you have done it several times before, and if you have not done something like this before, then it is definitely better if you have immediate access to

colleagues who have, so that you can compare notes on the methods that work with respect to that specific technique! The fact that the most senior rank in the medical profession is “consultant” is not accidental. We have no such safeguards. Researchers in our discipline frequently undertake research “cold” by taking the procedures that they use out of a book, or even more dangerously, out of a computer programme’s instruction manual. The fate of the Latin Square makes the consequences of this clear. We now have simply thousands of research techniques in print, many published by researchers working in isolation both before and after the paper is submitted for publication – the anonymity of the peer review system does us a specific disservice in this respect. The sheer variety of techniques that are deployed in academic marketing is a contributory factor to the almost complete absence of true replication research in our literature. The absence of this key check on method, coupled with the abnormally long time that it takes to get into print in this discipline, only increases the chance of flawed research transferring itself, unchallenged, into accepted. As practitioners within industries that have much lower margins for error than ours know perfectly well, while the written word is sufficient for transferring technique, it is not adequate for transferring method. Within academic marketing, only viable research centres, that have sufficient application volume of a specific range of techniques, will allow method to be effectively developed and transferred both within and outside the group, and only such development and transfer of method will ensure that our quality of research and review improves rather than deteriorates, as it did with regard to the Latin Square between 1952 and 1972.While such centres of excellence have been mooted in the past, the proposals have simply lacked the detail with respect to specific techniques that is necessary to form any basis of action. More detailed discussion is overdue. Notes 1. The Latin Square did not completely disappear from the marketing literature, but it has become extremely rare. When it does appear it is usually now in a highly modified form. Regrettably the published examples do not usually give sufficient detail to allow an evaluation of their application of the technique. Recent examples include Malcolm (2002) and Stayman and Aaaker(1988). 2. The “efficiency” discussed here only accounts for the number of consumer groups, four in this case. The convention that the minimum number of subjects per group/cell should be 30 is rather more difficult to pin down. It is an ancient convention going back to the earliest works in the area. It appears to be related to the convention that for a population n $ 30, the binomial expansion, a discrete distribution based on individual events such as a consumer evaluation on a Likert scale, can be considered equivalent to the normal (continuous) distribution on which the analysis of variance is based (Albright, 1988). 3. The Latin Square is the only design that allows this approach. A 4 £ 4 £ 4 full factorial experiment could not use blocking on consumer groups, as each consumer would have to evaluate a very large number of products (16 in the case of a 4 £ 4 £ 4 full factorial). More importantly they would see individual vehicles with multiple levels of the independent variable, which would reveal the purpose of the research to them. A 4 £ 4 full factorial with country of origin as the independent variable and product as the single extraneous vehicle would need (16 £ 30 ¼ 480) consumers, and would not have the Latin Square’s ability to detect the main effects and interaction arising from variations within them. A 4 £ 4 full

The rise and fall of the Latin Square 347

EJM 39,3/4 4.

348

5. 6.

7.

8.

9.

factorial with consumer groups as the extraneous variable would only have one vehicle to carry the four levels of the independent variable, and would therefore immediately reveal its purpose to the 916 £ 30 ¼ 480) consumers participating in the study. For the purposes of this paper a “mock up” is a cue vehicle that is clearly constructed for research purposes only. A “fake” is a cue vehicle that is indistinguishable for a commercial product. A wine identified by the type of grapes used to produce it. While strong profiles for each of these four countries were obtained, this research indicated that the region had as much impact on the consumer evaluations as the country of origin. As these two important cues are presented in combination on any quality wine product. It is likely that the effect of country and region may be impossible to “deconfound” in research on bottled wine. Cask wine, by contrast, is frequently presented without regional cues, and may represent a better vehicle for future research of this type. Both of these techniques were tested using an automated dataset into which all types of main effect, interaction and levels of error could be introduced at will. Tukey’s technique never did yield a significant result for interaction. Youden and Hunters’ technique was extremely reliable. In only 5 per cent of tests did it fail to report a significant (p # 0:05) interaction before the significance of main-effects were destroyed. Relative empirical data are data based on cell and treatment deviations from the mean of the experimental population. Absolute empirical data is based on cell and treatment deviations from a fixed reference, usually a control, or “zero” point. The two books edited by Davies (1956, 1957) that were used as a source by Cochran and Cox were commissioned by ICI for use in the chemical industry. The researchers that this text was intended for would only have occasionally come across the problems associated with nominal and loosely defined ordinal data.

References Albright, C.S. (1988), Statistics for Business and Economics, Macmillan, New York, NY. Asam, E.H. and Bucklin, L.P. (1973), “Nutrition labeling for canned goods: a study of consumer response”, Journal of Marketing, Vol. 37, pp. 32-7. Banks, S. (1965), Experimentation in Marketing, McGraw Hill Book Company, New York, NY. Brown, S. (1996), “Art or science? Fifty years of marketing debate”, Journal of Marketing Management, Vol. 12 No. 4, pp. 243-68. Brown, S. (1997), “Marketing science in a postmodern world: introduction to the special issue”, European Journal of Marketing, Vol. 31 No. 3/4, pp. 167-82. Brunk, M.E. and Federer, W.T. (1953a), “How marketing problems of the apple industry were attacked, and the research results applied”, Methods of Research in Marketing, No. 4, Department of Agricultural Economics, Cornell University, Agricultural Experiment Station, New York State College of Agriculture, Ithaca, NY. Brunk, M.E. and Federer, W.T. (1953b), “Experimental designs in probability sampling in marketing research”, American Statistical Association Journal, September, pp. 440-52. Cox, K. (1964), “The responsiveness of food sales to shelf space changes in supermarkets”, Journal of Marketing Research, Vol. 30, pp. 63-7. Currim, I.S. and Sarin, R.K. (1983), “A procedure for measuring and estimating consumer preferences under uncertainty”, Journal of Marketing Research, Vol. 20, pp. 249-56. Darley, W.K. and Smith, R.E. (1993), “Advertising claim objectivity: antecedents and effects”, Journal of Marketing, Vol. 37, pp. 100-13.

Davies, O.L. (Ed.) (1956), The Design and Analysis of Industrial Experiments, 2nd ed., published for Imperial Chemical Industries, Oliver & Boyd, Edinburgh. Davies, O.L. (Ed.) (1957), Statistical Methods for Research and Production, 3rd ed., published for Imperial Chemical Industries, Oliver & Boyd, Edinburgh. Dean, M.L., Engel, F.E. and Telarzyk, W.W. (1972), “The influence of package copy claims on consumer evaluations”, Journal of Marketing, Vol. 36, pp. 34-9. Della-Bitta, A.J., Monroe, K.B. and McGinnis, J.M. (1981), “Consumer perceptions of comparative price advertisements”, Journal of Marketing Research, Vol. 18, pp. 416-27. Euler, L. (1782), Verh. Uitgegeven door het Zeeuwsh. Genootschap d. Watensch. te Vlissingen. Fishbein, M. and Ajzen, I. (1975), Beliefs, Attitude, Intention and Behavior: An Introduction to Theory and Research, Addison-Wesley, Reading, MA. Fisher, Sir R.A. (1937), The Design of Experiments, 2nd ed., Oliver & Boyd, Edinburgh. Fisher, Sir R.A. (1950), Contributions to Mathematical Statistics, John Wiley & Sons, New York, NY. Fisher, Sir R.A. (1970), Statistical Methods for Research Workers, 14th ed., Oliver & Boyd, Edinburgh. Hamlin, R.P. (1997), “The meat purchase decision”, PhD thesis, University of Otago, Dunedin. Hamlin, R.P. and Leith, K.J. (1995a), “The use of CADCAM and the Latin Square to isolate and investigate individual consumer cues in food products and their packaging”, The Food Technologist (NZ), Vol. 25 No. 1, pp. 32-8. Hamlin, R.P. and Leith, K.J. (1995b), “Isolating and investigating individual consumer cues in food products and their packaging – an example from the wine industry”, The Food Technologist (NZ), Vol. 25 No. 3, pp. 99-104. Han, C. (1989), “Country image, halo or summary construct?”, Journal of Marketing Research, Vol. 20, pp. 22-9. Han, C. (1990), “Testing the role of country image in consumer choice behaviour”, European Journal of Marketing, Vol. 24 No. 6, pp. 24-40. Holland, C.W. and Cravens, D.W. (1972), “Fractional factorial designs in marketing research”, Journal of Marketing Research, Vol. 10, pp. 270-6. Hoofnagle, W.S. (1965), “Experimental designs in measuring the effectiveness of promotion”, Journal of Marketing Research, Vol. 2, pp. 154-62. Kennedy, J.R. (1970), “The effect of display location on the sales and pilferage of cigarettes”, Journal of Marketing Research, Vol. 7, pp. 210-15. Kent, R.J. and Allen, C.T. (1994), “Competitive interference effects in consumer memory for advertising: the role of brand familiarity”, Journal of Marketing, Vol. 58, pp. 97-105. Kohli, R. (1988), “Assessing interaction effects in Latin Square type designs”, International Journal of Research in Marketing, Vol. 5, pp. 25-37. Leith, K.J. (1993), “The effect of country of origin on consumer attitudes towards domestic and imported wine”, MCom thesis, University of Otago, Dunedin. Malcolm, S. (2002), “The impact of pitch, volume and tempo on the atmospheric effects of music”, International Journal of Retail & Distribution Management, Vol. 30 No. 6/7, pp. 323-40. Piercy, N.F. (2002), “Research in marketing: teasing with trivia or risking relevance?”, European Journal of Marketing, Vol. 36 No. 3, pp. 350-63. Popper, K. (1935), Logik der Vorschung, Julius Springer Verlag, Vienna. Robinson, J. (1994), The Oxford Companion to Wine, Oxford University Press, New York, NY.

The rise and fall of the Latin Square 349

EJM 39,3/4

350

Sawyer, A.G. (1973), “The effects of repetition of refutational and supportive advertising appeals”, Journal of Marketing Research, Vol. 10, pp. 23-33. Sheluga, D.A., Jaccard, J. and Jacoby, J. (1979), “Preference search, and choice: an integrative approach”, Journal of Consumer Research, Vol. 6, pp. 166-76. Snedecor, G.W. (Ed.) (1955), “Queries”, Biometrics, Vol. 11, pp. 111-13. Stayman, D.M. and Aaaker, D.A. (1988), “Are all the effects of ad-induced feelings mediated by Aad”, Journal of Consumer Research, Vol. 15 No. 3, pp. 368-73. Tukey, W.J. (1949), “One degree of freedom for non additivity”, Biometrics, Vol. 5, pp. 232-42. Tull, D.S. and Hawkins, D.I. (1987), Marketing Research, Measurement and Method, Macmillan, New York, NY. Yadev, M.S. and Monroe, K.B. (1993), “How buyers perceive savings in a bundle price: an examination of a bundle’s transaction value”, Journal of Marketing Research, Vol. 30, pp. 350-8. Youden, W.J. and Hunter, J.S. (1955), “Partially replicated Latin Squares”, Biometrics, Vol. 7, pp. 399-405. Further reading Montgomery, D. (2003), “Challenges to academic marketing: towards the next decade, keynote address”, Proceedings of the Academy of Marketing Conference 2003, Aston Business School, Birmingham, published on CD-ROM, Academy of Marketing, Helensburgh.

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0309-0566.htm

Equivalence of survey data: relevance for international marketing Hester van Herk Vrije Universiteit, Amsterdam, The Netherlands

Ype H. Poortinga Tilburg University, Tilburg, The Netherlands and the University of Leuven, Leuven, Belgium, and

Equivalence of survey data

351 Received January 2003 Revised November 2003 Accepted January 2004

Theo M.M. Verhallen Tilburg University, Tilburg, The Netherlands Abstract Purpose – The paper presents a framework for establishing equivalence of international marketing data. The framework is meant to reduce confusion about equivalence issues, and guide the design of international studies and data analysis. Design/methodology/approach – A short overview is given of the two main approaches to equivalence in the literature. These are integrated and used to distinguish sources of cultural bias in the various stages of the research process. Findings – The highest levels of equivalence most often established are construct equivalence and partial measurement equivalence, implying that distributions of scores obtained in various countries cannot be interpreted at face value. To understand cross-cultural differences better, researchers should investigate why higher levels of equivalence could not be established; this can be done best by including elements from both the conceptual and the measurement approach to equivalence. Practical implications – This study can help marketing managers to establish the extent to which consumer perceptions can be considered equal across countries. Moreover, it helps researchers to determine causes of unequivalence and relate these to concrete stages in the research process. Originality/value – Integration of the two main approaches to equivalence will lead to a better understanding of the validity of cross-cultural differences and similarities. This should lead to improved decision making in international marketing. Keywords International marketing, Market research, Bias Paper type Conceptual paper

Introduction Faced with maturing markets and stiffening competition, industries are forced to rethink their strategies. In this, internationalisation of activities is a main strategy. Several multinationals have interests in at least 30 counties (Mitra and Golder, 2002), and a company such as Unilever sells its Lipton tea in as many as 110 countries (Unilever, 2003). As companies are increasingly engaging in global trade, global marketing has become vital. Cultural, economic, legal, and geographic differences between the home market and the markets of other countries have to be taken into account. Such differences also imply that people may react differently to marketing efforts.

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 351-364 q Emerald Group Publishing Limited 0309-0566 DOI 10.1108/03090560510581818

EJM 39,3/4

352

Unfortunately, little empirical research is available on customs, habits, attitudes, and reactions to marketing efforts in different regions. Therefore, companies tend to collect marketing information themselves (or have this done for them) in order to make well-founded decisions. The resulting growing need for international marketing research information is shown in the worldwide turnover for commercial opinion and market research; in 2001, this was e17 billion, up 5.8 per cent over the previous year (ESOMAR, 2002). In comparison with 1980, there was a seven-fold increase. When making international comparisons data should have the same meaning across those countries, because inequivalent or biased information leads to ambiguous or even erroneous conclusions. Therefore, the equivalence or comparability of data collected across countries is regarded as a key issue (e.g. Douglas and Craig, 1983; Hui and Triandis, 1985a; Sekaran, 1983; Singh, 1995; Van de Vijver and Leung, 1997a, b). Despite its importance, the equivalence of data is usually not examined (Aulakh and Kotabe, 1993; Malhotra et al., 1996; Sin et al., 1999, 2001) and most culture comparative studies do not address equivalence issues. This lack of attention for issues of culture is not limited to research on marketing; in commercial marketing studies, on which marketing decisions tend to be based, equivalence issues are also hardly addressed. The reasons for this negligence are not clear, but the analysis of equivalence in data is not a simple matter. In addition, lack of clarity in the literature has added to the complexity. In this article we try to present an integrated approach. We give an overview of terminology used in different publications and we distinguish two major approaches to equivalence. One approach focuses more on the whole research process, whereas the other approach focuses on data analysis. The objective of this article is to provide a framework for establishing equivalence that may help reduce the confusion, and better integrate measures that can be taken to avoid or deal with bias in data. First, a short overview is given of equivalence approaches in the literature. Second, we attempt to integrate the various approaches, introducing different levels of equivalence and linking these to sources of bias in the research process. Finally, we discuss what kinds of inferences are justified if there is evidence supporting various levels of equivalence. Approaches to equivalence In most general terms there are two approaches to equivalence. One is a psychometric approach in which characteristics of parameters in measurement models are tested for invariance across countries. If certain conditions of invariance are satisfied certain comparisons are deemed valid (e.g. Steenkamp and Baumgartner, 1998). The second approach, that started earlier in international marketing research, has been summarised by Douglas and Craig (1983). They started from a series of problems encountered in cross-cultural research for which convenient solutions were sought. In a recent edition of their well known handbook Craig and Douglas (2000, p. 141) define equivalence as: Data that have, as far as possible, the same meaning or interpretation, and the same level of accuracy, precision of measurement, or reliability in all countries and cultures.

Craig and Douglas (2000) address various issues that have to be taken into account if data are to be compared. They distinguish three forms of equivalence: construct equivalence, measurement equivalence, and equivalence in data collection techniques.

Within construct equivalence Craig and Douglas (2000) define three aspects: (1) Conceptual equivalence is “concerned with the interpretation that individuals place on objects, stimuli or behaviour, and whether these exist or are expressed in similar ways in different countries and cultures” (Craig and Douglas, 2000, p. 158). (2) Categorical equivalence “relates to the category in which objects or other stimuli are placed” (Craig and Douglas, 2000, p. 159). Categorical equivalence refers to comparability in product category definitions, and in background or socio-demographic classes that exist between countries. This definition by Craig and Douglas arises from the practice of marketing research. Product categories need not be similar across countries. For example, beer belongs to the category soft drinks in Southern Europe, whereas beer is considered to be an alcoholic beverage in Northern Europe. Moreover, category sizes may differ; in Greece spreading on bread or toast is common, making the category big, whereas spreading is hardly done in Italy, making the category small (Van Herk and Verhallen, 1995). (3) Functional equivalence relates to the question whether the concepts, objects or behaviours studied have the same role or function in all countries included in the analysis. It makes quite a difference whether a bicycle is considered mainly as a means of transport (such as in The Netherlands or India) or as a product for recreational purposes (as in the USA). Craig and Douglas (2000, p. 160) take examination of equivalence as a two-step procedure: “once construct equivalence has been examined, the next step is to consider measurement equivalence”. They distinguish three aspects of measurement equivalence: (1) Translation equivalence refers to the translation of the research instrument into another language so that it can be understood by respondents in different countries and has the same meaning in each research context. (2) Calibration equivalence refers to equivalence with regard to units of measurement, for example, monetary units and measures of weight used in questionnaires. Moreover, it refers to the use of colours and shapes in such a way that they are interpreted the same in different countries. (3) Finally, metric equivalence refers to the specific scale or scoring procedure used for assessment. In the approach by Craig and Douglas (2000) a solution is sought per problem, for example translation. There is little integration of conceptual and measurement issues. Other research in the same tradition as Craig and Douglas can be found in the management literature with authors like, for example, Sekaran (1983), Nasif et al. (1991), and Cavusgil and Das (1997). Sekaran (1983) links equivalence to various stages in the research process. She mentions equivalence issues related to function, instrumentation, data-collection methods, sampling design, and data-analysis. As in marketing, functional equivalence is associated with the role of objects or behaviours in different countries. Instrumentation equivalence includes equivalence in translation, syntax and concepts used. With data collection Sekaran mentions the importance of equivalence in response, timing, interviewer status, and type of research (longitudinal

Equivalence of survey data

353

EJM 39,3/4

354

or cross-sectional). Sampling equivalence covers issues such as representativeness, and matching of samples. Following Sekaran (1983), Nasif et al. (1991) identified methodological problems in the cross-cultural research process and gave suggestions for reducing those problems. They mention several issues like functional equivalence and equivalence of instrumentation and data collection, and per issue they indicate suggestions for improvement. For example, back-translation is recommended to increase translation equivalence. Building on this work Cavusgil and Das (1997) developed a “generic process model” for cross-cultural research. In this model, including seven steps, they thoroughly describe issues to be taken into account when doing a cross-national study. As in the studies already mentioned, issues of equivalence are linked to stages in the research process (for example, equivalence of administration and equivalence of responses are linked to the phase in the research process where the instrument is developed). We like to note that in this line of research data analysis is regarded important. However, it tends to be taken as one of several aspects in the cross-cultural research process. In the second line of research on equivalence the emphasis has been on data analysis, as the principal means of demonstrating whether or not cross-cultural data can be taken as equivalent. This research has its roots in psychology, specifically in literature on bias and measurement invariance (e.g. Horn et al., 1983; Meredith, 1993). In these studies psychometric procedures are defined for assessing whether (test) scores from different groups can be validly compared. In psychology (e.g. Little, 1997), marketing (e.g. Steenkamp and Baumgartner, 1998), as well as in management literature (e.g. Mullen, 1995; Vandenberg and Lance, 2000) the value of measurement invariance for cross-cultural research has been recognised. These authors argue that the equivalence of measures can be established by means of multi-group confirmatory factor analysis models. By adopting the procedures as outlined in, for example, Steenkamp and Baumgartner (1998) construct, metric or scalar invariance of measures can be established. That is, sequential steps in nested multi-group mean and covariance structure models can determine the extent to which constructs can be compared across groups. To distinguish the levels of invariance Van de Vijver and Leung (1997a, b) proposed three hierarchically ordered categories: construct equivalence, measurement unit equivalence, and scalar equivalence: (1) Construct equivalence (or structural equivalence) is the same as “configural invariance” a term also used (e.g. Horn and McArdle, 1992; Steenkamp and Baumgartner, 1998; Vandenberg and Lance, 2000). It refers to similarity of structural psychometric properties in data from different countries (Van de Vijver and Leung, 1997a). Construct equivalence exists if equal factor structures are obtained in different cultural populations. In terms of interpretation, construct equivalence implies that the same construct is being assessed. However, it is by no means certain that the scoring levels employed in one country are equivalent to those used elsewhere. (2) Measurement unit equivalence is also called “metric invariance” (Horn and McArdle, 1992; Vandenberg and Lance, 2000). It refers to a situation where the unit of measurement is equal across populations, but where the origin of the measurement scale may be different. An analogue is the measurement of temperature, where degrees Celsius and degrees Kelvin are measured in the

same units, but where the zero point (offset) differs. Thus, in terms of interpretation, measurement unit equivalence does not imply that scores on a single variable can be compared across countries; it implies that differences between scores (or patterns of scores) can be meaningfully compared across countries. (3) Scalar equivalence or full-score equivalence (also called “scalar invariance”) exists if the measurement scale in addition to having measurement unit equivalence also has an equal origin across countries. Scalar equivalence is the highest level of equivalence according to Van de Vijver and Leung (1997a, b). Comparisons of scores across countries on a single variable are only meaningful if this level of equivalence has been established. If there is scalar equivalence, it can be concluded that cross-national differences in score distributions on a variable correspond to differences in the underlying constructs. In this line of research, the level of equivalence that has been established determines which inferences can be made. For example, if a trait like innovativeness is the target of study it can be concluded that people in culture “A”, are less innovative than people in culture “B”, only if scalar equivalence has been established. On the other hand, if the (positive) evidence is limited to construct equivalence the only conclusion can be that the instrument used assesses innovativeness in both cultures, but it is unclear whether a higher mean score in “A” implies a higher level of innovativeness. It should be noted that in this line of research multi-item scales are needed; with single items multivariate procedures cannot be applied. Bias in the research process There can be sources of bias in every stage in the research process. To gain equivalent results in international marketing studies, attention has to be paid to a range of possible sources of bias and their impact. This implies an integration of the process-oriented approach (as by, e.g. Craig and Douglas, 2000), and the measurement-oriented approach (as by, e.g. Steenkamp and Baumgartner, 1998). The linch-pin between the two orientations in our opinion lies in the notion that sources of bias can affect (in)equivalence at different levels. In the psychological literature (e.g. Van de Vijver and Leung, 1997a, b; Berry et al., 2002) three kinds of bias are discussed, namely construct bias, method bias, and item bias: (1) Construct bias is likely to be present if the construct being studied differs across countries, or if the operationalisation does not fit cultural understanding. Construct bias can, for example, be induced if behaviours are sampled that are not associated with the construct studied. The use of butter for baking in one country cannot be compared with the use of butter for spreading in another country, and as a consequence, attitudes towards butter will reflect quite different notions about the use of butter (Van Herk et al., 1994). (2) Method bias refers to instances where all or most items in a questionnaire are equally affected by a factor that is independent of the construct studied (Berry et al., 2002). Method bias can be due to interviewers (interviewer-interviewee interaction), the research method (telephone, mail or personal interviewing), or

Equivalence of survey data

355

EJM 39,3/4

background characteristics of respondents, such as age or social class (Greenleaf, 1992a). (3) Item bias refers to distortions in specific items in the instrument (see Van de Vijver and Tanzer, 1997). Suppose, we employ a multi-item scale on “health consciousness” and an item is included on “visiting a fitness club at least once a week”. With an equal average concern about “health consciousness” in two groups, but differential availability of health clubs, the answer “no” obviously will have a different meaning. In such instances, we say that the item is biased.

356

At the beginning of a research process in marketing (stage I in Table I), the problem is formulated and the objectives of the study are defined. In a cross-national study, a common first check is to determine whether the issue to be studied is relevant across countries. This includes the concepts to be examined, and in commercial studies it also comprises the product category studied, and the function of products and consumer habits. Insight into foreign markets can be obtained from the literature, consultations with fellow researchers (see Craig and Douglas, 2000), colleagues who are nationals of target countries, and/or qualitative pre-studies (Malhotra et al., 1996), such as focus groups (Carson et al., 2001), and exploratory observation. For example, Barzilay et al. (1994) reported studies in Western Europe in which the behaviour of women during food preparation was videotaped to help marketers understand habits in other countries. Those habits turned out to be very different; for example, for frying potatoes a deep-frying pan with special fat was used in Germany, whereas women in Greece used a frying pan with olive oil. It turned out that the women used similar words, but the actual behaviour regarding frying was quite different. Such differences illustrate that concepts need not be equal in meaning and/or associated behaviours. Thus, country specific (or “emic”) practices are important to understand differences between

Stages in the marketing research process I

Table I. The research process and bias

Problem formulation

Source of bias

Issues

Concepts Purpose of the study Category Function II Research design Operationalisation Type of study Type of questions Instrument design Item selection Type of response format Translation Method Personal, mail, telephone III Sample selection Sampling Target population Sampling frame IV Data collection Fieldwork Procedures Interviewer selection Time frame V Data editing and coding Editing Data editing Coding Data coding Calibration VI Analysing and interpreting data Statistical procedures

Prevalent types of bias Construct Construct Item Method Item Method Method Method Item

countries. It is striking to note that about 80 per cent of studies in cross-cultural organisational research use an “etic” (culture-common) approach (Schaffer and Riordan, 2003), and less than 15 per cent include emic elements. In summary, during the first stage of a research project, the main way to minimise bias is through international collaboration; this provides important information on specific habits, and the suitability of methods. At stage II, the design stage, decisions are made concerning operationalisation of the constructs, the selection of items, and the response format. At this stage instruments (questionnaires, observation schedules) are developed and indications of construct, method, and item bias may emerge. For example, construct bias should be suspected, if a construct cannot be operationalised in a similar way in the countries studied. Again, collaboration with (preferably multi-lingual or bi-lingual) researchers across countries is vital. Another issue related to construct bias is the use of multi-item scales. Multi-item scales are required to be able to assess measurement invariance as outlined by, for example, Van de Vijver and Leung (1997a). In recent academic cross-national studies, the measurement of constructs using multi-item scales, needed for psychometric analysis of equivalence, seems more common (see, e.g. studies by Van Birgelen et al. (2002) and Athuahene-Gima and Li (2002)). However, for reasons of financial and time constraints, multi-items scales are scarce in commercial marketing research (Reynolds, 2000). Other decisions made at Stage II are decisions on what response scales to use. Method bias is introduced at this stage if there is any factor in the instructions, response format of the items, or administration procedure that elicits different reactions across countries. The format of response scales is a case in point. For example, in the USA a five-point or a seven-point rating scale is most common, whereas in France, a 20-point scale prevails (Kotabe and Helsen, 1998). To minimise a difference in familiarity with the response scales, a good introduction with some practice items can be provided. In addition to scale use, method bias can be introduced if respondents are unfamiliar with a particular data collection method. For example, in Western countries it is common to use computerised personal interviewing or computerised telephone interviewing (CAPI method and CATI method (e.g. Malhotra and Birks, 2003)), whereas this is still completely unknown in other parts of the world. Less familiarity with a research method is likely to affect results (see, e.g. Serpell, 1979). The use of different methods in various countries does not alleviate problems; cultural differences in the results can then still be real differences as well as differences due to the methods used, while several psychometric procedures for identifying method bias are no longer available. To minimise method bias, it is better to use the same method and the same response scales, and to give respondents the opportunity to practice. Another important issue at the design phase is the translation of the instrument into other languages. The translation of one or more items can be less than optimal, because of the absence of precisely equivalent terms in each language. To minimise bias, back-translation is often recommended (e.g. Craig and Douglas, 2000). Another common method to develop a translation is the committee approach (see, e.g. Van de Vijver and Tanzer, 1997); the strength of this approach is in the co-operative effort between people with different areas of expertise who together translate the instrument. Translation is paid attention to in more and more academic studies nowadays (see, e.g.

Equivalence of survey data

357

EJM 39,3/4

358

Sin et al., 2001), and also in commercial research it is an issue researchers are aware of (Reynolds, 2000). At stage III, the sample composition and the sampling frame are determined. The definition of the sample may introduce bias in various ways. One strategy is to work with samples that are representative of the target populations. In commercial surveys representative samples or samples specified by the client are preferred (Reynolds, 2000). Another strategy is to choose samples that are alike with respect to demographic characteristics. Such matched samples, for example students, can help to reduce bias. In academic studies, about half of the studies use matched samples (Schaffer and Riordan, 2003). To make between-country comparisons, samples should preferably show equal distribution on key demographic variables, such as age, education and income. This helps to determine whether differences found are real or measurement artefacts. If it is not possible to use similar samples, recording of background characteristics (e.g. age, education) is recommended to be able to control statistically for differences. This information can help detect differences in response styles (a type of method bias) such as yea-saying, that are known to be more prominent in people with a lower education, and a higher age (e.g. Greenleaf, 1992; Narayan and Krosnick, 1996). At the data collection phase (IV), virtually any procedure is vulnerable to method bias. To begin with, instructions to interviewers need not always be understood in the same way. Method bias can emerge during interviews if respondents are more willing to talk about sensitive issues with certain interviewers; women may be more willing to talk about violence to females than to males. Moreover, bias may be induced by different time frames. If data are collected in one country half a year or more before this is done elsewhere differences in fashions or in the eco-cultural environment (e.g. economic situation) may lead to different answers. This especially holds in a commercial setting, where the social context may affect variables such as buying intention. Again, method bias cannot be prevented; it can only be reduced. However, the researcher is not helpless; instructions can be tried out in pilot studies; and interviewer characteristics can be recorded. The latter should be standard practice, as it is known that interviewer effects can be non-negligible (Kumar, 2000). During stage V, coding and editing, item bias may be introduced. Coding refers to assigning answers to response categories if open-ended questions are used, and editing refers to correcting inconsistent answers in the questionnaires. Item bias is more likely if coding and editing are done separately in each country. Thus, item bias can be decreased if there is central coordination of research activities. At stage VI, the analysis phase, it is possible to assess the absence or presence of bias by means of statistical analysis. Procedures outlined by research on measurement invariance can be followed. In the preceding phases, one can be aware of bias (threats to equivalence), and try to minimise or avoid these, but the empirical proof of equivalence (i.e. absence of bias) usually has to come from analyses of equivalence after the data have been collected. An illustration In 1996, a pan-European analysis of the (male) shaving market was conducted. Countries included France, Germany, Italy, Spain and the UK. A main goal was to find similarities between these markets that could be used as a starting point for

pan-European product developments and pan-European product introductions. Therefore, the comparability of market information across countries was a main issue. As the company concerned had been active in the shaving market for almost a century, much information was available on the domain of study. For example, market shares in the various countries and shaving habits were known. In the past, qualitative exploratory studies in several countries had been done to determine the dimensions men use to describe their shaving experiences. At stage I, extensive information was available to the researchers. They knew that men use six dimensions to describe their shaving experience. The expected similarity of these dimensions across European countries made it worthwhile to examine whether they could be decomposed in much the same way. For example, it was expected that similar notions should exist to describe the dimension “shaving result”. In other words, it was expected that operationalisations of the dimensions should lead to (structurally) equivalent scales. As validated scales to measure these dimensions were not readily available, items had to be developed for each dimension (Stage II). In this process items from previous marketing research studies were used. After compiling the questionnaire items, it was decided that five-point rating scales, with the endpoints labeled 1 (“disagree strongly”) to 5 (“agree strongly”) should be employed. In addition, questions were developed on male shaving behaviour (e.g. shaving frequency, method used). As a next step, the questionnaires were translated from English into German, Spanish, Italian and French by bi-lingual researchers using the committee approach (see Van de Vijver and Tanzer, 1997). In all countries, the method chosen was a mail survey using a panel of a large marketing research agency. As in many commercial marketing research studies, the choice for this type of data collection method was driven by financial and time constraints. The sample sizes were fixed at about 1,000 in each country (Stage III). Representative samples of only male respondents were selected; ranging from 15 to 80 (mean 43) years of age in each country. It should be noted that this choice of representative samples might lead to method bias, because differences in demographic variables such as education and income level are known to exist between the countries studied. Stage IV entails the data collection. Instructions were given to the respondents, who were all members of established marketing research panels. The data collection was done in the same period in all countries studied. No special precautions were taken at this stage to avoid, or control possible sources of bias. Stage V does not apply in the present case, as there were no open-ended questions, and thus no special rules for editing and coding were needed. Whether the data collected in the way described were equivalent had to be established afterwards through data analysis. In a sense, the proof of the pudding had to be in the eating. This last stage (VI) was done in four steps. The first step was data cleaning. Respondents with missing values on items of interest were removed from the data set. Resulting sample sizes were 985 in Germany, 890 in France, 820 in the UK, 1062 in Italy, and 790 in Spain. The partial non-responders did not differ from the rest of the sample on demographic variables. The second step was equivalence assessment. For the sake of clarity, we focus here on one construct to assess equivalence in the shaving domain, namely “shaving result”. In the questionnaire six items were included

Equivalence of survey data

359

EJM 39,3/4

360

that together could be used to assess this construct. The items included, for example, “after shaving you can see there is not a single hair left uncut on your face” and “you are closely shaven from early morning till late at night”. Next, to test for equivalence in the five countries, the program Lisrel 8.5 (Jo¨reskog and So¨rbom, 2001) was used. It turned out that the six variables chosen to measure the construct “shaving result”, were not construct equivalent. Inspection showed that the poor fit was mainly due to a single item (“gives a very close shave”). After removal of this item the fit improved, and “shaving result” was construct equivalent However, further analyses showed that there was no measurement unit equivalence. Thus, the level of equivalence that could be established was construct equivalence. The third step in the analysis involved the interpretation of results on equivalence. The finding of construct equivalence justified the interpretation that men in all the five countries understand the same thing when “shaving result” is talked about. However, as there was no measurement unit equivalence, let alone scalar equivalence, we could not infer whether men in, for example, Italy do experience a better shaving result than men in Germany do. The fourth and final step concerned the substantive explanation of results. For international marketing purposes it was important to know why the item “gives a close shave” caused item bias, and what can be concluded on the basis of the construct “shaving result” across countries. For answers to these questions we used other information from the questionnaire, especially items on shaving behaviour. Regarding the first point it was found that “shaving result” is positively related to shaving frequency in all countries, whereas the item “gives a close shave” was not in some countries. Regarding the second point it could be concluded that the construct “shaving result” had the same meaning in all countries, but between country comparisons at levels of scores (e.g. means) were not allowed. However, investigating relations of the construct “shaving result” with other variables within each country was allowed. Such results can be valuable for marketing decision making. It was for example found that within all countries men scored higher on “shaving result” when they shaved with a blade as compared to an electric shaver. As this result was found in all countries, blade and electric shaving could be compared on “shaving result” in the same (qualitative) way in a pan-country communication strategy. However, it should be noted that this is no quantitative comparison. It remained unclear in this study why there was no measurement unit equivalence.

Discussion In analysing equivalence some researchers focus on the research process, while others are mainly concerned with analysis and interpretation of data. For greater clarity, we proposed here a differentiated view, distinguishing levels of equivalence and types of bias. Sources of bias in the research process are considered factors that decrease the level of equivalence that can be established. In our approach, equivalence is accepted if serious attempts to find inequivalence have been unsuccessful. In this we follow Van de Vijver and Leung (1997a, b), who reserve the use of the term “equivalence” for outcomes of formal statistical analyses. This makes the meaning of the term “equivalence” less extensive, and more a matter of measurement, than the way it is used by authors like Craig and Douglas (2000).

In this study we indicated sources of bias in the research process, and linked these to the level of equivalence that can be established. Construct bias is the prevalent type of bias in the first two stages in the marketing research process. This type of bias is the most serious one, because it precludes any form of comparison, making cross-national comparisons ambiguous or even erroneous. Item bias is less serious than construct bias as it only affects part of the items in the instrument. If various items are used to measure a construct, a biased item can be eliminated from the scale, and the resulting shortened scale can still be construct equivalent. Method bias affects the level of scores on all, or at least most items in a scale. If there is method bias, it is still possible to establish construct equivalence, or even measurement unit equivalence. However, scalar equivalence is ruled out. It is not easy to eliminate method bias, since separation of bias and real differences is not straightforward (Greenleaf, 1992). With respect to understanding method bias response styles offer an interesting avenue for further research (see, e.g. Baumgartner and Steenkamp, 2001; Smith, 2004). Equivalence cannot be assumed; construct, measurement unit, or scalar equivalence has to be established by means of explicit procedures. In our example from commercial research only construct equivalence was found, be it after elimination of one item. This is not exceptional. In research papers in international marketing, such those by Homburg et al. (2002), and by Van Birgelen et al. (2002) construct equivalence was found, but measurement unit equivalence was not. In other studies (Steenkamp and Baumgartner, 1998; Vandenberg and Lance, 2000) only partial measurement unit equivalence was found. Thus, even if the topic of research is well studied, the samples are matched, and questionnaires are carefully back-translated, equivalence is not guaranteed. In articles as mentioned, the use of equivalence testing is often seen as a pre-test: if a certain level of equivalence is attained certain comparisons can be made, if not, comparisons between countries are not allowed. However, if a next level of equivalence cannot be established, analyses to find out why this is so can still provide valuable information. Analyses should not stop; investigating which items are biased and why this is the case is an interesting avenue for further research. Our study has one important implication for the management of international companies. A large number of managerial decisions of companies is influenced by consumer perceptions and acceptance of a company’s products. The findings of our study can help managers to establish the extent to which such consumer perceptions are equal across countries. That is, if the perceptions are construct equivalent, it can be concluded that they have the same meaning for people in all countries studied. Building on this foundation, management can then take the decision to apply these concepts in the form of a multi-county communication strategy. Being able to establish the level of equivalence therefore provides business value, as the risk of making a wrong decision decreases. International marketing research studies are expensive, and cutback in expenditure is often looked for. However, this cutback in expenditure should not be in multi-item scales. They are worth the money, because they may provide valuable information on differences and similarities between countries. We like to conclude with a comment by Cavusgil and Das (1997, p. 74) who argued that it is: “easier to recover from lapses in data analysis than in specification error”. Minimising construct bias in the early stages of the research process is a basic prerequisite. At that stage collaboration with other researchers and marketeers can help define the marketing (research) problem. Later, to be certain that corresponding

Equivalence of survey data

361

EJM 39,3/4

362

constructs were measured in all countries, multi-item scales are required to establish equivalence. Best practices should include elements from both the conceptual and the measurement-oriented approach to equivalence. Analysis of cross-cultural differences is partly an art. A researcher needs a proper grasp of the various factors that can interfere with the interpretations of findings at face value. However, the analysis of cross-cultural differences is also a science; a conceptual approach aids in a systematic procedure for assessing equivalence. References Athuahene-Gima, K. and Li, H. (2002), “When does trust matter? Antecedents and contingent effects of supervisee trust on performance in selling new products in China and the United States”, Journal of Marketing, Vol. 66 No. 3, pp. 61-81. Aulakh, P.S. and Kotabe, M. (1993), “An assessment of theoretical and methodological development in international marketing: 1980-1990”, Journal of International Marketing, Vol. 1 No. 2, pp. 5-28. Barzilay, J., Chada, P.M., Van Herk, H. and Verhallen, T.M.M. (1994), “International segmentation”, in Lambrinopoulos, C. (Ed.), Marketing Review 1992, Athens, pp. 207-8. Baumgartner, H. and Steenkamp, J.B.E.M. (2001), “Response styles in marketing research: a cross-national investigation”, Journal of Marketing Research, Vol. 38 No. 2, pp. 143-56. Berry, J.W., Poortinga, Y.H., Segall, M.H. and Dasen, P.R. (2002), Cross-cultural Psychology. Research and Applications, 2nd ed., Cambridge University Press, Cambridge. Carson, D., Gilmore, A., Perry, C. and Gronhaug, K. (2001), Qualitative Marketing Research, Sage Publications, Beverly Hills, CA. Cavusgil, S.T. and Das, A. (1997), “Methodological issues in empirical cross-cultural research: a survey of the management literature and a framework”, Management International Review, Vol. 37 No. 1, pp. 71-96. Craig, C.S. and Douglas, S.P. (2000), International Marketing Research, 2nd ed., Wiley, New York, NY. Douglas, S.P. and Craig, C.S. (1983), International Marketing Research, Prentice-Hall, Englewood Cliffs, NJ. European Society for Opinion and Marketing Research (ESOMAR) (2002), ESOMAR Annual Study on the Market Research Industry 2001, ESOMAR, Amsterdam. Greenleaf, E.A. (1992), “Improving rating scale measures by detecting and correcting bias components in some response styles”, Journal of Marketing Research, Vol. 29 No. 2, pp. 176-88. Homburg, C., Krohmer, H., Cannon, J.P. and Kiedaisch, I. (2002), “Customer satisfaction in transnational buyer-supplier relationships”, Journal of International Marketing, Vol. 10 No. 4, pp. 1-29. Horn, J.L. and McArdle, J.J. (1992), “A practical and theoretical guide to measurement invariance in aging research”, Experimental Aging Research, Vol. 105, pp. 117-44. Horn, J.L., McArdle, J.J. and Mason, R. (1983), “When invariance is not invariant: a practical scientist’s look at the ethereal concept of factor invariance”, The Southern Psychologist, Vol. 1, pp. 179-88. Hui, C.H. and Triandis, H.C. (1985), “Measurement in cross-cultural psychology. A review and comparison of strategies”, Journal of Cross-cultural Psychology, Vol. 16 No. 2, pp. 131-52. Jo¨reskog, K.G. and So¨rbom, D. (2001), Lisrel 8 Users’ Reference Guide, SSI, Chicago, IL.

Kotabe, M. and Helsen, K. (1998), Global Marketing Management, John Wiley & Sons, New York, NY. Kumar, V. (2000), International Marketing Research, Prentice-Hall, Upper Saddle River, NJ. Little, T.D. (1997), “Mean and covariance structures (MACS) analyses of cross-cultural data: practical and theoretical issues”, Multivariate Behavioral Research, Vol. 32 No. 1, pp. 53-76. Malhotra, N.K. and Birks, D.F. (2003), Marketing Research: An Applied Orientation, 2nd European ed., Pearson Education, London. Malhotra, N.K., Agarwal, J. and Peterson, M. (1996), “Methodological issues in cross-cultural marketing research”, International Marketing Review, Vol. 13 No. 5, pp. 7-43. Meredith, W. (1993), “Measurement invariance, factor analysis and factorial invariance”, Psychometrika, Vol. 58 No. 4, pp. 525-43. Mitra, D. and Golder, P.N. (2002), “Whose culture matters? Near-market knowledge and its impact on foreign market entry timing”, Journal of Marketing Research, Vol. 39 No. 3, pp. 350-65. Mullen, M.R. (1995), “Diagnosing measurement equivalence in cross-national research”, Journal of International Business Studies., 3rd quarter, pp. 573-96. Narayan, S. and Krosnick, J.A. (1996), “Education moderates some response effects in attitude measurement”, Public Opinion Quarterly, Vol. 60, pp. 58-88. Nasif, E.G., Al-Daeaj, H., Ebrahimi, B. and Thibodeaux, M.S. (1991), “Methodological problems in cross-cultural research: an updated review”, Management International Review, Vol. 31 No. 1, pp. 79-91. Reynolds, N.L. (2000), “Benchmarking international marketing research practice in UK agencies – preliminary evidence”, Benchmarking: An International Journal, Vol. 7 No. 5, pp. 343-59. Schaffer, B.S. and Riordan, C.M. (2003), “A review of cross-cultural methodologies for organizational research: a best-practices approach”, Organizational Research Methods, Vol. 6 No. 2, pp. 169-215. Sekaran, U. (1983), “Methodological and theoretical issues and advancements in cross-national research”, Journal of International Business Studies, Fall, pp. 61-73. Serpell, R. (1979), “How specific are perceptual skills?”, British Journal of Psychology, Vol. 70, pp. 365-80. Sin, L.Y.M., Cheung, G.W.H. and Lee, R. (1999), “Methodology in cross-cultural consumer research: a review and critical assessment”, Journal of International Consumer Marketing, Vol. 11 No. 4, pp. 75-96. Sin, L.Y.M., Hung, K. and Cheung, G.W.H. (2001), “An assessment of methodological development in cross-cultural advertising research: a 20-year review”, Journal of International Consumer Marketing, Vol. 14 No. 2/3, pp. 153-92. Singh, J. (1995), “Measurement issues in cross-national research”, Journal of International Business Studies, 3rd quarter, pp. 597-619. Smith, P.B. (2004), “Acquiescent response bias as an aspect of cultural communication style”, Journal of Cross-cultural Psychology, Vol. 35 No. 1, pp. 50-61. Steenkamp, J.B.E.M. and Baumgartner, H. (1998), “Assessing measurement invariance in cross-national consumer research”, Journal of Consumer Research, Vol. 25, June, pp. 78-90. Unilever (2003), available at: www.unilever.com Van Birgelen, M., De Ruyter, K., De Jong, A. and Wetzels, M. (2002), “Customer evaluations of after-sales service contact modes: an empirical analysis of national culture’s consequences”, International Journal of Research in Marketing, Vol. 19 No. 1, pp. 43-64.

Equivalence of survey data

363

EJM 39,3/4

364

Van de Vijver, F.J.R. and Leung, K. (1997a), Methods and Data Analysis for Cross-cultural Research, Sage, Beverly Hills, CA. Van de Vijver, F.J.R. and Leung, K. (1997b), “Methods and data analysis for cross-cultural research”, in Berry, J.W., Poortinga, Y.H. and Pandey, J. (Eds), Handbook of Cross-cultural Psychology. Vol. I. Theory and Method, 2nd ed., Allyn & Bacon, Boston, MA, pp. 257-300. Van de Vijver, F.J.R. and Tanzer, N. (1997), “Bias and equivalence in cross-cultural assessment: an overview”, European Review of Applied Psychology, Vol. 47 No. 4, pp. 263-79. Vandenberg, R.J. and Lance, C.E. (2000), “A review and synthesis of the measurement invariance literature: suggestions, practices, and recommendations for organizational research”, Organizational Research Methods, Vol. 3 No. 1, pp. 4-69. Van Herk, H. and Verhallen, T.M.M. (1995), “Response effects in international research in the food area”, Proceedings of the 2nd International Conference on The Cultural Dimension of International Marketing, Odense, pp. 392-402. Van Herk, H., Verhallen, T.M.M. and Barzilay, J. (1994), “Methodological issues in international segmentation”, in Bloemer, J., Lemmink, J. and Kasper, J. (Eds), Marketing: Its Dynamics and Challenges, EMAC, Maastricht, pp. 1311-14.

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0309-0566.htm

The evolution of “classical mythology” within marketing measure development Nick Lee and Graham Hooley Aston Business School, Aston University, Birmingham, UK

Marketing measure development 365 Received November 2002 Revised June 2003 Accepted January 2004

Abstract Purpose – This paper provides a deeper examination of the fundamentals of commonly-used techniques – such as coefficient alpha and factor analysis – in order to more strongly link the techniques used by marketing and social researchers to their underlying psychometric and statistical rationale. Design/methodology approach – A wide-ranging review and synthesis of psychometric and other measurement literature both within and outside the marketing field is used to illuminate and reconsider a number of misconceptions which seem to have evolved in marketing research. Findings – The research finds that marketing scholars have generally concentrated on reporting what are essentially arbitrary figures such as coefficient alpha, without fully understanding what these figures imply. It is argued that, if the link between theory and technique is not clearly understood, use of psychometric measure development tools actually runs the risk of detracting from the validity of the measures rather than enhancing it. Research limitations/implications – The focus on one stage of a particular form of measure development could be seen as rather specialised. The paper also runs the risk of increasing the amount of dogma surrounding measurement, which runs contrary to the spirit of this paper. Practical implications – This paper shows that researchers may need to spend more time interpreting measurement results. Rather than simply referring to precedence, one needs to understand the link between measurement theory and actual technique. Originality/value – This paper presents psychometric measurement and item analysis theory in easily understandable format, and offers an important set of conceptual tools for researchers in many fields. Keywords Market research, Psychometric tests, Statistics Paper type Research paper

Introduction Since 1979, when Churchill’s seminal article “A paradigm for measure development in marketing research” (Churchill, 1979) was published in the Journal of Marketing Research, the use of complex statistical tools in the development of reflective, multi-item scales to measure latent constructs within marketing research has flourished. In the decade between 1980 and 1989 Bruner and Hensel (1993) reported 750 instances of multi-item scales in the six leading marketing journals. It cannot be denied that the widespread adoption of this “classical” measure development paradigm has enriched marketing science immeasurably as a discipline, over and above its rather unrefined state prior to 1979 (see Peter, 1979). More recently, the level of sophistication of marketing measure development has been further enhanced by the introduction into non-technical literature of confirmatory factor analysis (CFA) methodology by authors such as Gerbing and Anderson (1988), Gerbing and Hamilton (1996), Bagozzi and

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 365-385 q Emerald Group Publishing Limited 0309-0566 DOI 10.1108/03090560510581827

EJM 39,3/4

366

Yi (1988), and Fornell and Larcker (1981). The great benefit to a field such as marketing of articles such as those referred to above, is that they distill critical methodological knowledge in such a way as to enable its widespread adoption by those not inclined to delve into complex, obscure, and often impenetrable technical resources. In such ways is the evolution of marketing as a science assisted since, as an applied discipline, marketing scholars (if most of us are honest) are primarily interested in the results of studies, not the intricacies of analysis and mathematics involved in obtaining those results. However, the enormous benefits of following paradigms such as that proposed by Churchill (1979) have also masked some of their dangers to marketing scholarship, which are inherent to the latter’s applied nature. In particular, the great efforts expended by authors such as Churchill (1979), Gerbing and Anderson (1988), Bagozzi and Yi (1988), and Diamantopoulos (1999) to make difficult and rigorous methods accessible to general marketing scholarship, seem to have resulted in the development of a number of misconceptions, or myths, around the application of these complex methods. It could be argued that these misconceptions have evolved as a result of our increasing ability to use complex methods, without the requirement of a fundamental understanding of their mathematical and – crucially – conceptual underpinnings. Furthermore, these myths and misconceptions tend to be perpetuated within marketing research since it is a field where one of the primary justifications for any method’s use is simple precedent (Stewart, 1981). To put things more succinctly, the labours of the distinguished authors already quoted (and many others) have allowed marketing academics to avoid dealing with a number of critical considerations simply by quoting or referring back to their work. Furthermore, the continued lack of attention to these issues has led to a state where analysis techniques are applied slavishly to a given task to which they may be inappropriate, or the results may lack sufficient explanation to assure readers of their appropriateness. As a consequence there may be serious problems with the measures we develop, and ultimately the theories we aim to test. In fact, this has been stated even more strongly by one of the “fathers” of psychometric testing, Paul Kline: “to follow the precepts of [test construction and/or use] without understanding their psychometric rationale is to accept dogma, the antithesis of a scientific approach to . . . testing, a cast of mind which in earlier times led inevitably to the dark ages” (Kline, 2000, p. 32). This paper is an attempt to go some way towards addressing the above concerns within the context of the evolution of marketing research. More specifically, the aim of this paper is to present some fundamental issues concerning classical measure development techniques in a non-technical manner and to a wide audience of general marketing academics. The issues we detail are concerned solely with the classical psychometric paradigm for measure development as defined by authors such as Churchill (1979), DeVellis (1991) and Spector (1992), and the techniques we discuss are based around scale purification, reliability and validity assessment. While we recognise that contemporary work has expanded the canon of techniques available to marketing research (e.g. Diamantopoulos and Winklhofer, 2001; Gerbing and Anderson, 1988; Rossiter, 2002), we also recognise that a great many of our existing measures were based upon Churchill’s (1979) construction method, and indeed many still continue to be. Furthermore, by discussing methods that have had a long history within marketing, we hope to illuminate critical evolutionary issues concerning our use

of classical measure development methods, which may not be as prevalent when discussing newer methods without such a history of usage. Nevertheless, it should be clear that our aim is not to present a technical explication but a conceptual one, and to utilise well-founded marketing methodologies to illustrate how the evolution of marketing research has proceeded to a state whereby it is arguable that our use of classical measure development methods may in fact have detracted from our goal of providing valid, reliable measures. We begin with a brief reiteration of the “classical” paradigm for measure development introduced into marketing by Churchill (1979), with particular focus on the purification stage of the process. Following this, the rest of the paper discusses issues related to two key methods of measure purification. First, we discuss some key potential problems with the concept of internal consistency in purifying measures, and of Cronbach’s (1951) coefficient alpha to measure internal consistency. We also suggest some areas where we could improve our use of internal consistency and alpha criteria. Following this, we then turn to issues concerned with the usage of exploratory factor analysis (EFA) in measure purification, and suggest areas for particular caution for marketing scholars. Finally, we provide some conclusions and suggest some critical ways in which marketing researchers can advance the use of existing measure development tools in achieving the goals of scientific measure development. The “classical” paradigm for the development of measures of marketing constructs The publication of Churchill’s seminal paper on construct development marks a watershed in the evolution of marketing science. Prior to the 1970s many marketing measures were poorly developed and poorly tested (Jacoby, 1978; Peter, 1979), and Churchill’s paper was essentially the first to apply classical psychometric theory (which was then already well-accepted in other disciplines) to the marketing discipline. In the introduction to his paper, Churchill (1979) quoted a then-recent paper by Jacoby (1978, p. 91): More stupefying than the sheer number of our measures is the ease with which they are proposed and the uncritical manner in which they are accepted. In point of fact, most of our measures are only measures because someone says they are, not because they have been shown to satisfy standard measurement criteria (validity, reliability and sensitivity).

In proposing a more systematic and rigorous approach to measure development, Churchill laid the groundwork and procedures for the next generation of marketing scholars. Specifically, Churchill argued for the development of multi-item measures as more robust ways to measure marketing phenomena, and put forward an eight-step procedure to achieve that. This classical paradigm is presented in Figure 1. While few would argue with the stages proposed, the issue of how each stage is undertaken, and what techniques are appropriate for each, is open to question (summarised in the right-hand column in Figure 1). This is because the stages are conceptual, yet we must use various specific operational techniques to achieve their goals. If the link between the operationalisation and the conceptual goal is broken, we run the risk of hampering rather than helping our tasks. In particular, the central stage of “measure purification” is an area where technical dogma and tradition have dominated marketing academia, to the potential detriment of better measures.

Marketing measure development 367

EJM 39,3/4

368

Figure 1. Churchill’s paradigm for developing better measures of marketing constructs

Measure purification Classical psychometrics (see Nunnally, 1967), as advocated by Churchill (1979), recommends that marketing measures be developed through the systematic generation and subsequent “purification” of a large bank of scale items (DeVellis, 1991; Spector, 1992). Under this approach, the underlying construct that the researcher seeks to measure is represented by a “sample” of the hypothetical, and usually infinite, set of all the items which could possibly be created to measure the construct (see Kline, 2000). The critical issues in identifying an appropriate sample of items are: . to ensure that the final set of items used captures or reflects the underlying construct as fully as possible, but without redundancy; and . to ensure that the items really do reflect the underlying construct and not some other, similar but distinct, construct. The scale-purification process attempts to “purify” the scale (after the initial item-generation and data-collection phases) to remove redundant or non-reflective items. Two statistical techniques have traditionally been recommended for scale purification, coefficient alpha and EFA (Churchill, 1979). The balance of the present

paper discusses a number of concerns with both techniques, which the evolution of marketing research appears to have marginalised, to the detriment of the conceptual goals of the classical measure development paradigm. A glossary in the Appendix at the end of this paper provides definitions of terms used. Cronbach’s coefficient alpha and the reliability of multi-item scales When developing multi-item scales according to the classical paradigm, Cronbach’s (1951) coefficient alpha had been the measure of choice within marketing and many other fields for assessing the reliability of a given scale (Peter, 1979; Peterson, 1994). While the “composite reliability” coefficient (Fornell and Larcker, 1981; Gerbing and Anderson, 1988) has received some use recently, coefficient alpha still appears to be the primary foundation upon which the reliability assessment of marketing scales is based. However, while coefficient alpha has close to an unassailable position within marketing research, scholars in other disciplines have raised a number of concerns over the use and misuse of alpha, even prior to Churchill’s (1979) work (e.g. Boyle, 1991; Cattell, 1973; Cattell and Kline, 1977; Cortina, 1993; Green et al., 1977; Hattie, 1985; Rogers et al., 2002). Indeed, even within marketing it has been suggested that there is some need for care when using coefficient alpha to assess reliability, although no real criticisms appear to have been levelled until recently (e.g. Rossiter, 2002). However, before discussing the possible limitations of coefficient alpha as a reliability indicator, and any subsequent caveats for its usage, a brief re-examination of both the concept of measure reliability, and the coefficient alpha statistic itself is presented. Reliability and validity The reliability of a measure is high when “independent but comparable measures of the same trait or construct of a given object agree” (Churchill, 1979, p. 65). Reliability is unimportant in and of itself, but is nevertheless vital since in order to be valid (i.e. to measure what it claims to measure) a given measure must be reliable (Kline, 2000; Peter, 1979; Spector, 1992). It is important to note, however, that a measure may be reliable without being valid (Kline, 2000). Reliability is concerned with the proportion of variance in a measure, which is attributable to the true score on the latent construct that is being measured (DeVellis, 1991). For example, if variance in a measure of say, job satisfaction, is comprised of a high proportion attributable to the true score on the latent job satisfaction construct, then that measure of job satisfaction is reliable. Conversely, if a subject’s score on the job satisfaction measure is comprised of a high percentage of random error, then the job satisfaction measure is not reliable. Thus, reliability is essentially a theoretical concept, in that (practically speaking) we can never be certain of the true score of a subject on a given latent construct (Kerlinger and Lee, 2000) – and if we could then we would measure the latent construct directly, and it would no longer be latent. Furthermore, we can never be certain that the other variance in a measure not attributable to the latent construct is in fact all random error and not some part of systematic error (such as respondent biases). As a consequence of this definition, it is clear that reliability can never be exactly determined, but it can be approximated in a number of ways (see Kerlinger and Lee, 2000; Kline, 2000). Drawing from classical measurement theory, it has been suggested that if the scores of the items of a reflective scale are highly influenced by the level of the latent construct

Marketing measure development 369

EJM 39,3/4

they are designed to measure (in other words not influenced by error), then the items should also have strong relationships (i.e. high intercorrelations) with each other (DeVellis, 1991). Thus, high intercorrelations between scale items should indicate that the items are also strongly related to the latent construct, and thus not highly influenced by random error. Scales that have high intercorrelations between the items are known as internally consistent (see Kline, 2000; DeVellis, 1991; Churchill, 1979).

370 Internal consistency The internal consistency of a scale is generally measured by coefficient alpha, although this is by no means the only workable approach (Kline, 2000; Spector, 1992). Coefficient alpha is, practically speaking, “the proportion of a scale’s total variance that is attributable to a common source” (DeVellis, 1991, p. 27), nothing more, nothing less. By definition, this means that alpha is also a measure of the unique variance in a scale. However, it should be clear from the above discussion that coefficient alpha is most definitely not an actual measure of reliability. Furthermore, while the “common source” referred to by DeVellis (1991) is assumed to be the latent construct we are trying to measure, this is only theoretically the case, and the link between the latent construct and the actual scale items is therefore purely theoretical. However, coefficient alpha can be utilised in estimating the reliability of a multi-item reflective scale, by providing an indication of a scale’s internal consistency. Nevertheless, the characteristics of both internal consistency, and coefficient alpha, necessitate a number of caveats. While most psychometricians are in no doubt that internal consistency is of considerable utility in estimating the reliability of a measure (see Kline, 2000), a number of dissenters do not share this confidence. In particular, concern has been heard for a significant amount of time from Cattell (e.g. Cattell, 1973; Cattell and Kline, 1977), while more recent caution has come from Boyle (1991) and Rossiter (2002) among others. The root source of this concern appears to be the possible negative impact of high levels of internal consistency on the validity of a measure. Drawing from the earlier discussion, if reliability is primarily important because of its impact on measure validity, then for internal consistency to be a good indicator of reliability, logically speaking it must not negatively impact a measure’s validity. There are two key premises to the argument against internal consistency. First, high internal consistency is strongly related to high intercorrelations between items. Second, for a scale to be a valid and useful measure of a latent construct, the scale scores must be able to predict the latent true scores to a high degr1e (Kline, 2000). Of course, according to classical reflective scale theory, technically speaking it is the latent true score that should predict the scale item scores. However, since we can never actually measure the true score of a latent construct, this distinction is arguably purely academic. Furthermore, if the true score is in theory able to predict the scale scores accurately, then in practical terms the reverse should also be true. Notwithstanding the above technicalities, drawing from multicolinnearity theory (which for marketing researchers has previously been primarily concerned with multiple regression), it can be shown that if we wish to predict the value of any variable from a set of independent items, then the items should not be correlated with each other (see Kleinbaum et al., 1998). As the correlation between the items increases, the explanatory power of any single item decreases (e.g. Kleinbaum et al., 1998; Hair et al., 1998). While this point has received some discussion in literature relating to construction of formative indices

within marketing (e.g. Diamantopoulos and Winklhofer, 2001), marketing-focused theory pertaining to classical reflective scales has not dealt with the issue. Nevertheless, Kline (2000, p. 12) suggests that “a test [i.e. a reflective scale] can be seen as a set of items with which we intend to predict the criterion [i.e. the latent construct] test score”. In other words for a scale to be valid, a subject’s score on a multi-item scale should be closely predictive of the subject’s actual level of the unobservable construct the scale is designed to measure. Thus, it would follow that the scale would be best able to predict this latent construct score if each individual item was highly correlated with the true latent construct score but correlated zero with all the other items. While such a scale would have a high level of validity (since it is able to measure what it claims to measure with a high accuracy), it would have a very low level of internal consistency, and thus a low coefficient alpha. By contrast, a scale with a high level of internal consistency is likely to have a high level of intercorrelation between its items, which would suggest that a number of those items may in fact be redundant in predicting the subject’s true score on the latent construct (see Boyle, 1991; Cattell, 1973). In other words, scales high in internal consistency may be comprised of items that are essentially paraphrases or repetitions of each other (Kline, 2000; Smith, 1999)[1]. Put crudely, a scale consisting of the same item repeated six times will give an excellent alpha (of 1.0 if respondents are consistent in their answers) but will have low validity in measuring the latent construct. This can lead to a situation where a scale has significant item redundancy, and may actually tap only a narrow area of a latent construct, leaving it low in validity (Boyle, 1991). Therefore, researchers should clearly be cautious when using internal consistency as a proxy for reliability in their scale development activities. Dogmatic adherence to high internal consistency can lead to high levels of item redundancy and measures which tap only limited aspects of a latent construct, ultimately leaving them of low validity. Notwithstanding the above issues, internal consistency remains a valuable source of information about a scale, whether that information suggests high reliability or a worrying level of item redundancy. It is the researcher’s job to interpret that information. However, if one wishes to evaluate internal consistency in scale development, then the characteristics of coefficient alpha must be understood also, in order for scale developers and users to interpret the results of a coefficient alpha test. The effects of scale length on coefficient alpha Coefficient alpha also has some relatively well-known caveats on its usage, perhaps the best known of which concerns the number of items in the scale to be tested. Specifically, as an artefact of the alpha formula itself, coefficient alpha will increase as the number of scale items increases ceteris paribus (Hair et al., 1998). So in other words, regardless of the level of internal consistency achieved by a scale, if enough items are present the alpha will be high. In theory this is not a problem, since it should be expected that as the number of items increases, so should the reliability of the scale as it gets closer to including the theoretical “universe” of all possible items (Kline, 2000). Despite the latter, in practical terms a problem may exist since the universe of all possible items is usually infinite (Kline, 2000). Thus, researchers must be wary if measures of reliability such as alpha are influenced by scale lengths, which in no way approach this universe.

Marketing measure development 371

EJM 39,3/4

372

Peterson’s (1994) meta-analysis discovered that scale length had a significant impact on coefficient alpha (R 2 of 0.1), a result also achieved by Churchill and Peter’s (1984) marketing-based research. While in absolute terms Peterson (1994, p. 386) suggested that this impact was not “especially strong”, it should be considered that a systematic bias of around 10 per cent (i.e. an R 2 of 0.1) is likely to have a substantive impact on the assumed reliability and validity of a measure. Unfortunately, further research outside of marketing has found that scale length has an even more pervasive effect on coefficient alpha. For example Cortina (1993) found that he could achieve levels of alpha commonly regarded as sufficient (i.e. over 0.7) with scales of 12 items and above even when the scale consisted of two or more completely uncorrelated dimensions. Thus, it can be seen that coefficient alpha can be problematic as a measure of internal consistency when scales consist of practically large numbers of items (i.e. well before such numbers approach the theoretical levels where reliability should significantly increase). In such situations, large values of alpha do not necessarily indicate either the presence of a single, homogenous factor (i.e. internal consistency), or high reliability, as is commonly assumed (see DeVellis, 1991). Furthermore, as might be expected from the previous discussion, coefficient alpha is downwardly biased when few items are present in a scale. In fact, the relationship between item numbers and alpha score is curvilinear (Komorita and Graham, 1965), which perhaps led to Peterson’s (1994) suggestion that the bias at low item numbers is more pronounced than at high. More specifically, it has been previously argued that, with very small numbers of scale items, coefficient alpha scores will underestimate internal consistency of the scale (see Gerbing and Anderson, 1988). Within marketing research this is a significant issue. Specifically, respondents to large-scale marketing surveys will not in general tolerate large questionnaires, and thus marketing researchers should try to reduce the length of their instruments wherever possible (see Churchill, 1999). Furthermore, increased use of covariance structure analysis methods seems likely to increase reliance on short scale lengths in order to reduce the number of estimated parameters (see Diamantopoulos and Siguaw, 2000; Kelloway, 1998; Sharma, 1996). As a result of this, the use of coefficient alpha under conditions of sub-optimal scale length should be considered very carefully. However, the use of the composite reliability formula (e.g. Gerbing and Anderson, 1988) can avoid problems regarding scale length, among other issues (Fornell and Larcker, 1981). Multidimensionality of scales A further issue concerning coefficient alpha, which was implied above, is its performance under conditions of multidimensionality. From the theories discussed so far, it is clear that coefficient alpha should be a measure of the extent to which a hypothesised common factor is present in all items. Furthermore, if there is one common factor to a scale, alpha should be a measure of the strength of that factor (Cortina, 1993). Therefore, if a scale is comprised of more than a single dimension, then to be a useful measure of internal consistency, alpha should provide some indication of this fact. Unfortunately, a significant amount of research has provided evidence that this is not the case (e.g. Cortina, 1993; Gerbing and Anderson, 1988; Green et al., 1977). In fact if enough items are present, a scale can contain even five separate dimensions of which each item loads on two separate dimensions and still return an alpha greater

than 0.8 (Green et al., 1977). The upshot of this is that high coefficient alphas cannot be taken as indicators of unidimensionality in a given scale (e.g. Gerbing and Anderson, 1988). This is because as mentioned earlier, high coefficient alphas simply measure the amount of unique and common item variance in a scale, and not what that common variance consists of. As alluded to earlier, the make-up of the common variance in a given scale (i.e. exactly what construct that scale measures) is a conceptual decision made by the researcher, not something that can be drawn from the results of a mathematical operation. The fact that high alphas do not indicate unidimensionality is unfortunate in the context of marketing research. In particular, Churchill (1979, p. 68, emphasis in original) explicitly stated that “[c]oefficient alpha absolutely should be the first measure one calculates to assess the quality of the [scale]”. This statement seems to imply that, if coefficient alpha reaches an acceptable level, one should simply take a measure to be of high quality. However, the theory presented above suggests clearly that this should not be the case. Nevertheless, despite the introduction of a number of sophisticated measure development techniques which do provide excellent indications of measure dimensionality (e.g. Gerbing and Anderson, 1988, etc.), it has been quite acceptable to utilise Churchill’s (1979) process even in the recent past. Furthermore, whether or not this changes in the future, it is undeniable that many of our key measures have been developed using the latter methodology, and are often repeatedly used by marketing scholars in a relatively unquestioning manner. Summary If one were to reflect on the arguments presented so far, one could rather easily get the impression from the above that coefficient alpha should be avoided at all costs. However, this is not our recommendation. Instead, we strongly advocate the interpretation of coefficient alpha values for scales, rather than the unquestioning reporting of scale alphas, usually coupled with the rigid acceptance of anything over 0.7. It is apparent that an alpha of 0.7 may technically mean exactly the same thing across different scales (i.e. the proportion of unique item variance), but in practical terms mean wildly different things, and in some cases mean exactly nothing of practical use, such as when item numbers are very large. Thus, marketing researchers should, and should have in the past, taken into account a number of factors when interpreting coefficient alpha: (1) Most important, researchers need to understand exactly what they are looking at, in terms of both internal consistency and alpha itself. In other words, does one actually want to measure internal consistency, and should it be high or low, and what do high/low alphas actually mean in the context of the study? (2) If alpha is used, researchers should be aware of the specific factors which can influence alpha, such as item numbers. (3) Researchers need to realise that if dimensionality is an issue (and according to classical theory it always is), one cannot rely solely on an alpha value, but must at minimum compute a precision estimate (see Cortina, 1993), and ideally utilise FA to examine dimensionality more fully. (4) Finally, researchers should interpret alpha scores in light of the factors above, and also in light of alphas which have been reported for the scale when it has

Marketing measure development 373

EJM 39,3/4

374

been used in other studies. For example an alpha of 0.7 on a job satisfaction scale may require different levels of concern depending on alphas reported in previous uses of the scale. The use of EFA in scale development While internal consistency and coefficient alpha are often primary tools in the early development of measures, EFA also assumes a prominent place in the measure purification process (Churchill, 1979; Gerbing and Anderson, 1988). However, it has been argued that there are many misconceptions about FA which have developed within marketing research, and that these myths have become so integrated into marketing research’s imaginary that such misinformation is perpetuated even by “[factor analysis’] defenders and some prominent reviewers” (Stewart, 1981, p. 51). Particular misapplications of EFA are many and varied, and indeed some marketing-focused literature has made attempts to provide discussion of a number of misconceptions (e.g. Flynn and Pearcy, 2001; Stewart, 1981). However, there remain important issues in need of substantive explication and justification. Some of these issues are relevant to marketing research as it has evolved, and others are especially pertinent due to newer developments in marketing science such as the introduction of alternative theories of measurement (e.g. Diamantopoulos and Winklhofer, 2001; Rossiter, 2002). Principal components analysis (PCA) and FA The first issue is the potential confusion which may arise between PCA and FA. Scholars outside of marketing have noted that there is considerable misunderstanding over the differences between PCA and FA, altho ugh they are conceptually and technically quite distinct (e.g. Sharma, 1996). However, the critical issues regarding these differences do not appear to have had significant discussion within marketing research literature. This has led to a situation where marketing researchers may employ PCA to do the (quite different) job of FA for a number of reasons. First, PCA and FA are both techniques of data reduction in that, ostensibly at least, they both reduce large numbers of items down to a smaller number of components or factors. Second (perhaps due to this similarity), many popular statistical software packages allow one to perform both PCA and FA from the same sub-menu, and indeed in some, PCA is the default option. The confusion is further exacerbated by the existence of principal components factoring, which is essentially a special case of PCA and arguably not actually FA at all (Sharma, 1996). Dealing with the conceptual differences first, it should be clearly stated at the outset that the aims of PCA and FA are quite different. The objective of PCA is to utilise the observed variance in the data set to create new variables which are composed of the original items. The objective of FA is to identify an underlying or latent factor which is responsible for observed correlations among the original items (see Kline, 2000; Sharma, 1996). Drawing from this, it can be seen that components extracted from a PCA do not mean anything conceptually, other than what they are composed of. In other words, they are an index of the original items, and a given component is simply a linear combination of the relevant items (see Bollen and Lennox, 1991). By contrast, a factor extracted from a FA is by definition responsible for the correlation between the relevant items, and thus does represent an underlying common or latent factor. Therefore, if one is to return to classical scale development theory, the use of PCA to

determine the underlying factor structure (i.e. the dimensionality) of a multi-item reflective scale is conceptually inappropriate, since the scale items in fact form a principal component, rather than reflecting the presence of an underlying factor (as is the case with a FA). Regardless of the conceptual differences between PCA and FA, it has been suggested that their actual results may be relatively similar within a given data set, particularly with large numbers of cases (see Stewart, 1981; Sharma, 1996), and thus it has been implied that in practical terms researchers should not be too concerned which technique they use (Stewart, 1981). However, without going into technical detail it can be seen that the variables that together make up a component formed by PCA are not assumed to have any unique variance (i.e. random error). This is due to the fact that the component is a straightforward linear combination of the items. However, within FA, each item is assumed to consist of a segment of shared variance, and a segment of unique variance (Sharma, 1996), with the shared variance representing the presence of the latent common factor. Moving back to the principles of classical measure development (see Churchill, 1979), it can be seen that an assumption is made that any single scale item’s observed score is comprised of a component attributable to the latent construct, and also of a unique component (perhaps caused by measurement error, or other factors). Therefore, utilising PCA to determine the underlying dimensionality of a reflective scale is fundamentally misleading and incorrect. In fact, in purely statistical terms, since error or other unique variance is not factored in, we should expect the loadings of individual items on the principal components in PCA to be inflated when compared with their counterparts in FA (Stewart, 1981), which could lead to errors and inconsistencies in scale construction when PCA is used for the latter purpose. Furthermore since, by using PCA to develop reflective scales, we are not taking into account the measurement or other error which classical theory tells us always exists in a given set of items, there is no reason to expect the “dimensions” uncovered by a PCA to be generalisable to other samples. This is because the components formed by a PCA are completely specified by the values of the individual items, and can be argued to be simply artefacts of the data set used to create them (see Kline, 2000; Nunnally, 1967). Ejection of items due to low communality Communality is the amount of variance an original variable shares with all other variables included in a FA (Hair et al., 1998). A communality of under 0.5 signifies that less than half of the variance in the item has been taken into account in identifying the latent construct (Hair et al., 1998). In psychometric terms, we consider the remaining variance to be made up of specific (also known as unique) variance and error variance. Specific variance is that variance associated with another variable, while error variance is that variance due to unreliability in the data-gathering process, measurement error or a random component in the measured phenomenon (Hair et al., 1998). It is common practice in scale purification through FA automatically to eject items from the analysis which exhibit low communality (e.g. Savery, 1993; Singh, 2000). However, drawing from the earlier discussion of internal consistency, it can be seen that the practice of ejecting items exhibiting low communality should be done with care. For example, a highly internally consistent measure containing a set of essentially redundant items will likely have high communalities for the items. By contrast, the

Marketing measure development 375

EJM 39,3/4

376

more valid yet less internally consistent measures advocated by authors such as Cattell (1973) are likely to have significantly lower item communalities. As a result, the rejection of items with low communalities can only be conceptually justified where it is either suspected that the error variance is high, or that the item may be tapping a substantively different construct or dimension. However, if the researcher suspects that this is not the case, a situation where common variance is low, but specific, or unique variance is high this may mean that the item is the sole measure of a particular aspect or nuance of a construct. To eject such an item could result in the loss of an important, and unique, contributing item. Indeed, presence of low communality in an item generated through sound prior research (literature based or empirical) should signal the possibility of under-specification of the particular aspect of the overall conceptual domain tapped by the item. Unfortunately there are no simple, statistical means for differentiating unique variance from error variance. It is therefore incumbent on the analyst to go back through the previous stages of the research (i.e. the item generation process in the Churchill model) to ensure that the item was well specified. Only where there can be confidence that either the error variance is high, or the item does not tap the appropriate conceptual domain, should the item be ejected from further analysis. Selection of number of factors A related issue concerns the number of factors to extract in the FA. In EFAs in marketing the common practice is to use the “Kaiser criterion” to make this decision (e.g. Alexander and Colgate, 2000; Urban, 2001). Under this approach factors are extracted only if they explain more of the total variance than a single item is capable of explaining. Each variable contributes a value of 1 to the total eigenvalue, hence any factors with an eigenvalue of less than 1 are discounted (Kim and Mueller, 1978). However, it is interesting to note that – technically speaking – the Kaiser criterion is only applicable when using PCA, although it has been argued that it is “applicable in spirit” to other types of FA (DeVellis, 1991, p. 97). In conceptual terms this is because the Kaiser criterion is solely based on maximising the variance explained of each factor (the goal of PCA), rather than their conceptual value (more in tune with the goal of FA). Thus researchers using FA should naturally be cautious when applying the Kaiser criterion in all situations. However in more specific terms, as Hair et al. (1998) point out, the use of the Kaiser criterion as a cut-off is most reliable when the number of variables is between 20 and 50. With less than 20 variables there is a tendency for this method to extract a conservative number of factors, whereas if the number of variables is greater than 50 too many factors may be extracted. Furthermore, it can be suggested that items which exhibit low levels of communality are also unlikely to appear in factors with eigenvalues greater than 1, since they exhibit low levels of correlations with all of the other items. In fact, what is likely to happen is that such items will again be candidates for ejection. Again, it is incumbent on the analyst to interpret this result carefully by returning to the earlier specification stages of the research to ensure that the items are well specified. Taken together, the issues raised concerning items with low communalities, or which load only on factors with eigenvalues less than 1, suggest the need to re-visit the specification stages of the research, rather than attempt to “correct” problems in the purification stage (one could perhaps refer to the old adage “garbage in – garbage out”

at this point). These might be poorly specified items with high error variance that deserve to be ejected if other items really are measuring the latent construct more effectively. Alternatively they may be important items that have been well specified, but are essentially single item scales, rather than the multi-item scales the research is attempting to develop. In this case the only alternative is to develop further items to better measure this poorly represented construct if the benefits of multi-item scales are to be realised. Because of the problems inherent in selecting the number of factors purely on the basis of statistical considerations (i.e. the Kaiser criterion), it has been suggested that researchers need to compare different solutions. For example, Hair et al. (1998) suggest the use of the Kaiser criterion to approximate the number of factors, then running the analysis with one less factors, with one more factor and with two more factors, to compare the interpretability of results. More specifically, “by examining a number of different factor structures derived from several trial solutions, the researcher can compare and contrast to arrive at the best representation of the data. An exact quantitative basis for deciding the numbers of factors to extract has not been developed” (Hair et al., 1998, p. 103). A further issue concerns the actual scale items that are used to describe the latent construct. To aid the selection of items, factor loadings (essentially the correlation between the item and the corresponding factor) are generally used. The squared factor loadings indicate the proportion of the variance in an original item that is explained by a particular factor. In common with most statistical techniques, the significance of a factor loading is dependent on sample size. However, the evolution of marketing research has resulted in the situation where a factor loading of 0.3 is generally assumed sufficient without consideration of the sample size (see Chae and Hill, 2000; Stewart and McAuley, 2000). Of course, it is actually the case that the lower the sample size, the higher the loading must be to indicate an item significantly loading on a given factor. More specifically, guidelines in Hair et al. (1998) suggest that with samples of 350 or more, the traditionally-used factor loading of 0.3 or greater is significant. However, with samples of 200 a factor loading of 0.4 or greater is needed for the same level of significance, while samples of 100 require loadings of 0.55 or greater for this level of significance. Thus, researchers would be advised to take sample size into account when evaluating the factor loadings of individual items. Additionally, difficulties of interpretation can arise where an item loads significantly on more than one factor (which has been termed “cross-loading”). Again, the normal practice within marketing research has been to eject such variables since they do not tap only a single construct. However, what significant cross-loading may imply is that the two factors are, to some degree at least, correlated with each other. As will be discussed subsequently, this is hardly surprising in many marketing situations where the items are all related to some degree. Thus, blanket ejection of cross-loading items may again detract from the overall goal of producing robust measures of the underlying, latent constructs. Interpretation of results through factor rotation Related to the issue of cross-loading is that of factor rotation. While it is beyond the scope of this paper to give a full reiteration of the relevant theory, factor rotation is simply a mathematical attempt to simplify the interpretation of the factor loadings of

Marketing measure development 377

EJM 39,3/4

378

individual variables (Sharma, 1996). From the first introduction of classical psychometry into marketing (e.g. Churchill, 1979; Peter, 1979), researchers have generally had the opportunity to choose between two main methods of factor rotation, oblique and orthogonal (see Cattell, 1978). The main difference between the two methods is that orthogonal rotations constrain the “rotation axes” to be at right angles to each other (i.e. factors are uncorrelated with each other), whereas oblique rotations allow the axes to vary in angle (i.e. factors may be correlated with each other). Despite the fact that the choice is available, it is clearly evident that orthogonal rotations are the rule within marketing research (see Stewart, 1981). However, it could be argued that the popularity of orthogonal rotation methods may stem mainly from non-theoretical reasons, such as convenience and precedent. Without detailing the geometrics of the issue, factor rotation is performed in order to allow the researcher to interpret the structure of the data more meaningfully, i.e. which items load on which latent constructs (Sharma, 1996). This is helpful since, when there are two or more factors, the variance in each item which is attributable to each common factor can infinitely vary (Sharma, 1996). Rotation can help the researcher to identify the most theoretically plausible factor structure. Thus, a rotated factor solution is by no means the only structure, but merely the one which maximises “simple structure” (i.e. items loading highly on only one factor, thus minimising the incidence of cross-loading). Since there are an infinite number of possible factor solutions obtainable by various rotations, the theoretical assumptions of the competing factor rotation methods are critical. In the present context the assumption of factor uncorrelation (i.e. orthogonal rotation) or correlation (i.e. oblique rotation) is particularly important. Specifically, a number of psychometricians have argued that uncorrelated factors are hardly ever likely to be the case (e.g. Cattell, 1978; Cattell and Kline, 1977; Hair et al., 1998; Kline, 2000). In terms of pure practicality, this is perhaps best argued by Cattell (1978, p. 104, emphasis in the original) when he states: “it makes sense for factors to be correlated rather than represented artificially in rigid orthogonality, because influences in the real world do get correlated”, and further “we should not expect influences in a common universe to remain mutually uninfluenced and uncorrelated” (Cattell, 1978, p. 128). Notwithstanding the “common-sense” logic of Cattell’s quotes, it is still possible to argue that a certain set of constructs should in fact be uncorrelated in the real world. However, in statistical terms there is still reason to argue that any uncorrelated set of constructs in the real world will in fact be correlated in any sample taken (Cattell, 1978). Furthermore, even if factors were, by some chance, uncorrelated in a sample (although this would naturally be unknown), one should still use an oblique rotation for a very simple reason. Namely, if the situation is such that it is appropriate to rotate using orthogonal axes, then an oblique rotation will use orthogonal axes since the angle between the axes is unconstrained (see Kline, 2000). Therefore, there appears little reason in theory to use an orthogonal rotation, which naturally begs the question as to why orthogonal rotation is so prevalent in marketing, and indeed seems to also be so within other disciplines (Stewart, 1981; Sharma, 1996). One reason may be that of pure statistical complexity (which may also be a reason for the use of the more mathematically elegant PCA rather than FA). At the time of his writing, Cattell (1978) implied that the physical effort and computational complication of hand-rotating oblique factor structures could be behind the prevalence of orthogonal

rotations. Furthermore, it is also clear that orthogonal rotational methods were available in computer form before oblique methods (Cattell, 1978). This may explain the lack of explicit consideration given to oblique methods in the early years of marketing psychometrics (e.g. Churchill, 1979). However, today we have no such excuses to fall back on, and it would seem that marketers have done themselves a significant disservice by relying on simple precedent in choosing to use orthogonal rotation, rather than examining the options in more detail. A further reason why we have continued to use orthogonal rotations is that they may offer statistical advantages which go beyond mere simplicity. Specifically, the creation of correlated factors by use of oblique rotation has potential to cause substantial problems in later model testing due to multicollinearity of the predictor variables. With the prevalence of multiple regression methods (and now structural equation modelling) in marketing research, it seems possible that many more informed researchers have avoided the use of oblique rotations when the latter methods may have in fact been more theoretically appropriate. Thus the rotation decision seems to boil down to a trade-off between theoretical rigour (which would suggest oblique rotation) and statistical simplicity (which would suggest orthogonal rotation). As a result, it seems that researchers would be advised to reverse the standard procedure, and instead beginning with oblique rotations by default, and only using orthogonal rotations when they were appropriate or necessary. In most cases, there is no theoretical reason to suggest uncorrelated factors in any situation in which FA is used, thus oblique rotations should be employed in the first instance. As Cattell (1978, p. 128) argues; “one does not need to confound logical, ideational independence with statistical independence in the domain of empirical observation . . . [t]he sad fact for research is that constraint to an artificial orthogonality destroys both the correctness of the pattern discovered and its constancy from one research to another”. As a result, marketing theories may become at worst sample-specific, and not generalisable to other contexts, a situation which seems to go against the very idea of scientific measure development in marketing. Summary Thus, when subjected to a more rigorous scrutiny, it can be seen that many of the unquestioned doctrines regarding EFA in marketing research should be viewed not as laws, but instead as suggested starting points for more detailed and thoughtful analysis. Furthermore, it would also seem that some of the procedures we commonly utilise are potentially conceptually incorrect, regardless of the fact that their practical differences may be minor in many cases (see Stewart, 1981). Thus, from the above a number of issues can be identified that researchers need to be aware of and take into account when using FA to purify scales: . The conceptual choice between PCA and FA must be explicitly addressed. PCA may be suitable for some purposes such as creating indices (see Sharma, 1996), while FA is more appropriate in the case of scales designed to reflect latent constructs. . Items with low communality should not be automatically ejected from the analysis. Further work should be undertaken to determine whether the low communality was due to error variance or specific variance. In the case of the latter the evidence may point towards the need to revisit earlier development

Marketing measure development 379

EJM 39,3/4 .

380 .

stages, to develop further items to measure the unique latent construct represented by the single item. The number of factors selected to represent the domain of interest should not be slavishly identified using the Kaiser criterion. Rather alternative solutions and dimensionalities should be examined and compared for interpretability. Where prior theory or empirical research suggests a particular factor structure confirmatory approaches are to be preferred over EFA (Gerbing and Hamilton, 1996). In rotating factor solutions explicit attention should be given to whether orthogonal or oblique rotation fits best with the theory being tested. Whatever rotation method is selected (e.g. VARIMAX, PROMAX, OBLIMIN etc.) should be justified conceptually rather than just through citation of precedent.

Conclusions, caveats, and limitations This paper has focused squarely on a single aspect of marketing research, measure development, and further on a single aspect of that – the item purification process in the classical psychometric paradigm. As a result, we recognise that one of the primary limitations of this paper is that very focus. In particular, we offer no specific insights into the wider context of measure development (e.g. formative measures), or regarding other stages of the classical paradigm (such as item generation and data collection). However, we do feel that the general points argued in this paper do have a wider relevance to scientific marketing research than just classical reflective scale development. That resonance concerns our generally high regard for the technical, operational and procedural aspects of research methodology, to the exclusion of conceptual foundations of those methods. With the exception of a minority of scholars, it is repeatedly the case that new methodologies are introduced into marketing, relatively unquestioningly accepted, then rigidly applied until the next methodological advance. This has been observed with, for example, classical psychometrics (e.g. Churchill 1979), CFA methodology (e.g. Gerbing and Anderson, 1988), and also with structural equation modelling (e.g. Bagozzi and Yi, 1988). We strongly believe that scientific marketing scholarship should value a critical conceptual approach to methodology, not an operational or procedural one. However, we recognise that, in writing such a paper as this, we run the risk of influencing and increasing further the dogma surrounding classical measure development, rather than reducing it. We have no wish that the ideas and concepts presented herein replace those “rules” already associated with the classical paradigm with simply another set of rules to abide by. Our arguments should be viewed as supplements to the tenets of psychometric marketing measure development, not replacements. In conclusion, the discussion of factor rotation above presents in a microcosm the entire point of this paper. Simply put, marketing researchers who subscribe to the idea of generalisable theories need reliable, valid, and ultimately generalisable measure development methods with which to measure the constructs which our theories are built around. However, we have tried to show that many of the methods which have been used in the past to create our most vital construct measures (and indeed are still being used today) may actually detract from the validity and generalisability of our

measures. Even worse, the reliance on rigid adherence to precedent has led to a situation where our use of such methods is hardly questioned within marketing. Nevertheless, this paper should not be seen as a diatribe against the classical model of measurement theory, but more as a defence of the theory itself. More specifically, we have tried to show that it is not the theory which is at fault, but that sometimes the link between the theory and techniques we use is broken, which leads us to employ inappropriate methods which do not create measures in line with classical theory (although those same methods may be appropriate in other measure development situations). What this paper should be seen as a criticism of is dogma. For too long have marketing academics (ourselves included – mea culpa!) been content to merely quote precedence in the use of techniques to develop measures, without spending time examining what these techniques were actually doing. Furthermore, even though newer methods have arrived to largely solve a number of the technical problems we discuss (such as assessing dimensionality), including CFA (e.g. Gerbing and Anderson, 1988; Gerbing and Hamilton, 1996), it should not become the case that these newer methods are used in the same way as the old, i.e. simply by referring to precedence. Instead, we strongly believe that it is part of our remit as scholars to understand the link between our theories and our techniques, and while it is unnecessary to continue to restate such things in research outputs (which is why the use of precedence indeed evolved), we at least need to understand them ourselves. In other words, we should be able to interpret results of our measure development tools, not simply report them. Note 1. In fact, it can be argued that a number of well-established marketing scales are excellent examples of this (R). References Alexander, N. and Colgate, M. (2000), “Retail financial services: transaction to relationship marketing”, European Journal of Marketing, Vol. 34 No. 8, pp. 938-53. Bagozzi, R.P. and Yi, Y. (1988), “On the evaluation of structural equation models”, Journal of the Academy of Marketing Science, Vol. 16 No. 1, pp. 33-46. Bollen, K.A. (1989), Structural Equations with Latent Variables, John Wiley & Sons, New York, NY. Bollen, K. and Lennox, R. (1991), “Conventional wisdom in measurement: a structural equations perspective”, Psychological Bulletin, Vol. 110 No. 2, pp. 305-14. Boyle, G.J. (1991), “Does item homogeneity indicate internal consistency or item redundancy in psychometric scales?”, Journal of Personality and Individual Differences, Vol. 12, pp. 291-4. Bruner, G.C. and Hensel, P.J. (1993), “Multi-item scale usage in marketing journals: 1980 to 1989”, Journal of the Academy of Marketing Science, Vol. 21, Fall, pp. 339-44. Cattell, R.B. (1973), Personality and Mood by Questionnaire, Jossey-Bass, San Francisco, CA. Cattell, R.B. (1978), The Scientific Use of Factor Analysis in Behavioral and Life Sciences, Plenum Press, New York, NY. Cattell, R.B. and Kline, P. (1977), The Scientific Analysis of Personality and Motivation, Academic Press, London.

Marketing measure development 381

EJM 39,3/4

382

Chae, M.-S. and Hill, J.S. (2000), “Determinants and benefits of global strategic marketing planning formality”, International Marketing Review, Vol. 17 No. 6, pp. 538-62. Churchill, G.A. Jr (1979), “Paradigm for developing better measures of marketing constructs”, Journal of Marketing Research, Vol. 16, February, pp. 64-73. Churchill, G.A. Jr (1999), Marketing Research: Methodological Foundations, Dryden, Orlando, FL. Churchill, G.A. Jr and Peter, J.P. (1984), “Research design effects on the reliability of rating scales: a meta-analysis”, Journal of Marketing Research, Vol. 21 No. 4, pp. 360-75. Cortina, J.M. (1993), “What is coefficient alpha? An examination of theory and applications”, Journal of Applied Psychology, Vol. 78, pp. 98-104. Cronbach, L.J. (1951), “Coefficient alpha and the internal structure of tests”, Psychometrika, Vol. 16, pp. 297-334. DeVellis, R.F. (1991), Scale Development: Theory and Applications, Sage, London. Diamantopoulos, A. (1999), “Modeling with LISREL: a guide for the uninitiated”, in Hooley, G. and Hussey, M. (Eds), Quantitative Methods in Marketing, 2nd ed., International Thomson Business Press, London, pp. 222-60. Diamantopoulos, A. and Siguaw, J.A. (2000), Introducing LISREL, Sage, London. Diamantopoulos, A. and Winklhofer, H.M. (2001), “Index construction with formative indicators: an alternative to scale development”, Journal of Marketing Research, Vol. 38, May, pp. 269-77. Flynn, L.R. and Pearcy, D. (2001), “Four subtle sins in scale development: some suggestions for strengthening the current paradigm”, International Journal of Market Research, Vol. 43 No. 4, pp. 409-23. Fornell, C. and Larcker, D.F. (1981), “Evaluating structural equation models with unobservable variables and measurement error”, Journal of Marketing Research, Vol. 18, February, pp. 39-50. Gerbing, D.W. and Anderson, J.C. (1988), “An updated paradigm for scale development incorporating unidimensionality and its assessment”, Journal of Marketing Research, Vol. 25, May, pp. 186-92. Gerbing, D.W. and Hamilton, J.G. (1996), “Viability of exploratory factor analysis as a precursor to confirmatory factor analysis”, Structural Equation Modeling, Vol. 3 No. 1, pp. 62-72. Green, S.B., Lissitz, R.W. and Mulaik, S.A. (1977), “Limitations of coefficient alpha as an index of test unidimensionality”, Educational and Psychological Measurement, Vol. 37, pp. 827-8. Hair, J.F. Jr, Anderson, R.E., Tatham, R.L. and Black, W.C. (1998), Multivariate Data Analysis, Prentice-Hall International, London. Hattie, J. (1985), “Methodology review: assessing unidimensionality of tests and items”, Applied Psychological Measurement, Vol. 9 No. 2, pp. 139-64. Jacoby, J. (1978), “Consumer research: a state-of-the-arts review”, Journal of Marketing, Vol. 42, April, pp. 87-96. Jo¨reskog, K.G. (1971), “Statistical analysis of sets of congeneric tests”, Psychometrika, Vol. 36 No. 2, pp. 1293-305. Kelloway, E.K. (1998), Using LISREL for Structural Equation Modelling, Sage, Thousand Oaks, CA. Kerlinger, F.N. and Lee, H.B. (2000), Foundations of Behavioural Research, Harcourt College, London. Kim, J.-O. and Mueller, C.W. (1978), Factor Analysis: Statistical Methods and Practical Issues, Sage, Beverly Hills, CA.

Kleinbaum, D.G., Kupper, L.L., Muller, K.E. and Nizham, A. (1998), Applied Regression Analysis and Other Multivariate Methods, Duxbury, Pacific Grove, CA. Kline, P. (2000), The Handbook of Psychological Testing, Routledge, London. Komorita, S.S. and Graham, W.K. (1965), “Number of scale points and the reliability of scales”, Educational and Psychological Measurement, Vol. 15, pp. 987-95. Nunnally, J.C. (1967), Psychometric Theory, McGraw-Hill, New York, NY. Peter, J.P. (1979), “Reliability: a review of psychometric basics and recent marketing practices”, Journal of Marketing Research, Vol. 16 May, pp. 6-17. Peterson, R.A. (1994), “A meta-analysis of Cronbach’s coefficient alpha”, Journal of Consumer Research, Vol. 21, pp. 381-91. Rogers, W.M., Schmitt, N. and Mullins, M.E. (2002), “Correction of unreliability of multifactor measures: comparison of alpha and parallel forms approaches”, Organizational Research Methods, Vol. 5 No. 2, pp. 184-99. Rossiter, J.R. (2002), “The C-OAR-SE procedure for scale development in marketing”, International Journal of Research in Marketing, Vol. 19 No. 4, pp. 305-36. Savery, L.K. (1993), “Difference between perceived and desired leadership styles: the effect on employee attitudes”, Journal of Managerial Psychology, Vol. 8, pp. 28-33. Sharma, S. (1996), Applied Multivariate Techniques, John Wiley & Sons, New York, NY. Singh, J. (2000), “Performance productivity and quality of frontline employees in service organizations”, Journal of Marketing, Vol. 64 No. 2, pp. 15-34. Smith, A.M. (1999), “Some problems when adopting Churchill’s paradigm for the development of service quality measurement scales”, Journal of Business Research, Vol. 46 No. 2, pp. 109-22. Spector, P.E. (1992), Summated Rating Scale Construction: An Introduction, Sage, London. Stewart, D.B. and McAuley, A. (2000), “Congruence of domestic and export marketing strategies – an empirical investigation of its performance implications”, International Marketing Review, Vol. 17 No. 6, pp. 563-85. Stewart, D.W. (1981), “The application and misapplication of factor analysis in marketing research”, Journal of Marketing Research, Vol. 18 No. 1, pp. 51-62. Urban, D.J. (2001), “Perceptions of advertising in the market for used vehicles”, Journal of Advertising Research, Vol. 41 No. 1, pp. 66-78.

Further reading Finn, A. and Kayande´, U. (1997), “Assessment and optimization of marketing measurement”, Journal of Marketing Research, Vol. 34, May, pp. 262-75.

Appendix. Glossary of terms Coefficient alpha. An indicator of the internal consistency of a multi-item scale. Often used as an indicator of reliability (Hair et al., 1998). Essentially a summary measure of the intercorrelations that exist among a set of items (see Churchill, 1999; Kerlinger and Lee, 2000). Composite reliability formula. A more recent approach to estimating reliability of a multi-item scale seemingly first articulated by Jo¨reskog (1971). This formula does not assume equal item reliabilities, and appears to have been introduced into mainstream marketing in the 1980s (see Fornell and Larcker, 1981; Gerbing and Anderson, 1988).

Marketing measure development 383

EJM 39,3/4

384

Confirmatory factor analysis (CFA). A factor analysis where the precise structure of the model is hypothesised, based on some theory. Thus true CFA is a theory testing approach (Sharma, 1996). However, CFA in marketing appears to have come to mean any factor analysis using covariance structure analysis techniques. Eigenvalue (also known as the Latent Root). The column sum of the squared loadings for any given factor. In essence, it is the amount of total variance in the data set accounted for by the factor (Hair et al., 1998). Exploratory factor analysis (EFA). A factor analysis where no structure is pre-specified, and the data are used to help reveal or suggest the structure of the model. EFA is therefore a theory building approach (Sharma, 1996). However, within marketing EFA appears to have come to mean only analyses conducted using “traditional” factoring techniques such as principal axis or principal components factoring. Factor loading. The correlations between the original variables and the extracted factors, vital to understanding the factor itself (Hair et al., 1998). Factor rotation. The manipulation of the factor axes in order to achieve a simpler and more meaningful solution (Hair et al., 1998). Formative index. A measure where the observed indicators are assumed to cause a latent construct rather than be caused by (as in reflective scales) it (Bollen, 1989; Diamantopoulos and Winklhofer, 2001). An example of a formative index (sometimes erroneously called a formative scale) is socioeconomic status (SES), which is a combination of four indicators. If any one of these increases, SES also increases, which would not necessarily be accompanied by an increase in the other three (Diamantopoulos and Winklhofer, 2001). Thus the indicators give rise to the construct. Internal consistency. The extent to which the items in a scale are inter-correlated (DeVellis, 1991). Kaiser criterion (also known as the Latent Roots criterion). A rule for deciding the number of factors to extract from a factor analysis. The Kaiser criterion states that only factors which explain more variance in the data set than one of the individual items should be retained (DeVellis, 1991; Sharma, 1996). This is usually operationalised by only retaining factors with an eigenvalue of greater than one (see Sharma, 1996). Latent construct. The underlying phenomenon which a scale (or sometimes an index) is intended to measure (DeVellis, 1991). Latent constructs are not directly observable (Nunnally, 1967), but can be measured (within definable limits of validity) by observing subjects’ responses to (for example) a multi-item scale. Measurement error. General term for the inaccuracies in measuring a subject’s “true” score on a latent construct, due to the shortcomings of the measuring instrument (Hair et al., 1998). Multicollinearity. The extent to which one variable in a model can be explained by the other variables (Hair et al., 1998). Multi-item scales (also called summated scales or composite measures). The combining of several indicators that measure the same (generally latent) construct into a single variable in order to reap the benefits (e.g. increased reliability) of multivariate measurement (Hair et al., 1998).

Oblique factor rotation. A factor rotation in which the factors are extracted in such a way that their axes are free to vary in angle, rather than restricted to orthogonality (Hair et al., 1998). This means that the factors are allowed (but are not necessarily, depending on the angle of the axes which results) to correlate (Kline, 2000).

Marketing measure development

Orthogonal factor rotation. A factor rotation in which the factors are extracted in such a way that their axes remain at 90 degrees to each other (Hair et al., 1998). This means that the factors are independent of one another and must have a correlation of 0 (Kline, 2000).

385

PCA. A data reduction technique where the sole aim is to reduce the set of items to a smaller number of factors which explain the maximum amount of variance. Variables are not composed of unique and common variance, and each factor is simply a linear combination of variables (Hair et al., 1998; Sharma, 1996). Precision estimate. The precision of Cronbach’s (1951) coefficient alpha is measured in terms of the standard error of the intercorrelations between the items in the scale, which is a function of the variance of those intercorrelations. The precision of a perfectly unidimensional scale will be 0. However, as the scale departs from unidimensionality, precision estimates become larger than 0. Thus, the precision estimate is a symptom of multidimensionality, although is not proof of the latter (see Cortina (1993) for a fuller explication of the precision of alpha). Reflective scales. A measure where the underlying latent construct is assumed to give rise to the observed indicators (Diamantopoulos and Winklhofer, 2001). This model is underpinned by much psychometric theory (see Kline, 2000) and within marketing appears to be the rule rather than the exception, although scholars have argued against its unthinking application (e.g. Bollen, 1989; Bollen and Lennox, 1991; Diamantopoulos and Winklhofer, 2001). Reliability. The amount to which a variable (or set of variables such as a multi-item scale) is consistent in what it intends to measure. This means that multiple administrations of the measure should return consistent results (Hair et al., 1998). In conceptual terms, if a measure is perfectly reliable, random error is 0. However, systematic error may still exist. Structural equation modelling. A term used to refer mainly to covariance structure analysis, where hypothesised relationships among latent constructs, and also latent constructs and their observed indicators, are tested using the covariance matrix derived from the observed data (see Diamantopoulos and Siguaw, 2000). Validity. The extent to which a variable (or set of variables such as a multi-item scale) correctly represents the construct it is designed to measure (Hair et al., 1998). A valid measure should have little systematic as well as little random error.

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

EJM 39,3/4

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0309-0566.htm

Long life to marketing research: a postmodern view Michela Addis and Stefano Podesta`

386 Received January 2003 Revised October 2003 Accepted January 2004

Institute of Management, Bocconi University, Milan, Italy and Marketing Department, SDA Bocconi Business School, Milan, Italy Abstract Purpose – This paper aims at interpreting the epistemology of marketing. The paper investigates several research questions, proposing some initial reflections concerning their impact on marketing. Design/methodology/approach – The paper addresses the research questions by conducting an analysis of the marketing literature. An analysis of philosophical postmodern literature is also carried out. The paper’s attempt constantly to create links between the level of philosophical elaboration and that of marketing research leads to a proposal of new approach to marketing research: experiential research. Findings – In the paper’s review of the marketing literature the traditional pragmatic approach of marketing as a discipline is highlighted. Its strong managerial perspective has partly diverted researchers’ attention from the theory, and focused it mainly on the method. This has created an increasingly marked distinction between the marketing literature aimed at management, and that aimed at the academic community. The postmodern perspective on marketing calls for a rethinking of the “scientific nature” of marketing as an investigative field. Research limitations/implications – The main point is that marketing cannot be a scientific discipline only by adopting a scientific method. Marketing research is by definition different in nature: it cannot generate better but only different knowledge. This perspective shift has an impact on all research components. First, the field of research widens enormously, because researchers can deal with everything arousing their interest and to which their accumulated knowledge can be applied. Since the discipline does not become scientific, the researcher can use any method. All methods can originate scientific theories, and therefore incremental knowledge. Hence science is neither objective nor absolute. Originality/value – This paper analyses the philosophical roots of postmodernism, in order to understand its impact on postmodern marketing better. It also focuses on the impact of postmodernism on marketing research, and proposes a new approach. This paper then explores the features of the experiential research in marketing, and its effect on the processes of generating knowledge. Keywords Market research, Postmodernism, Sciences Paper type Conceptual paper

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 386-412 q Emerald Group Publishing Limited 0309-0566 DOI 10.1108/03090560510581836

Postmodernism Postmodernism is a system of thought that became well established in the nineteenth and twentieth centuries. Postmodernism has been applied in a variety of disciplines, and, in all of them, it denies rationality and any kind of rationalisation, calling for fragmentation and multiplicity. Postmodernism can be defined as “a loss of faith in metanarratives”, according to Lyotard’s original idea. By metanarrative we mean any transcendental theory or a reference frame, which is used to evaluate and judge any other theory or reference frame (Lynch, 2001). The authors thank Antonella Caru`, Paola Cillo, Michele Costabile, Bernard Cova, and Francesca Golfetto for their useful comments on an earlier draft of the paper.

Introduction Since “postmodernism” has begun to spread among academics as a new philosophical and scientific concept, management theory has also witnessed, although with different tones in some of its components, a debate concerning a new interpretation of issues as well as discipline. This debate offers new horizons to academics, and could bring about some interesting developments like those in other disciplinary fields. In an extreme hypothesis, the impact of this debate could be so remarkable that not only the contents of management theory in their academic domain, but also the theory and research methods it is based on would have to be reconsidered. The adoption of a postmodern worldview, in fact, questions the concept of truth, as well as any other certainty that “modernism” has previously been based on, thus requiring a complex redefinition of the research process. In this paper, we try to interpret the epistemology of marketing, a specific part of management theory, by conducting an analysis of the literature as it has developed so far, and constantly creating links between the level of philosophical elaboration and that of marketing research. The review of marketing literature, and its evolution since its birth in the 1950s, highlights a clear “modern” approach – where the term modern refers to a specific vision of the world and lacks any positive meanings which are traditionally attributed to it – from which derive the principles and the concepts of marketing, commonly shared today (Brown, 1998). Although some authors have started to criticise the main concepts of marketing in the course of the years, and the first to do so were the academics of relationship marketing and those of experiential marketing, their observations are still mainly characterised by a “modern” vision of the world. The analysis of marketing epistemology, in fact, allows an in-depth reflection on the prevailing thought today, i.e. “modernism”, as well as the new “postmodern” thought. The critical digression from these issues allows a better insight into their epistemological basis, and makes it possible to suggest some initial reflections concerning their impact on marketing, in the light of the consideration that the adoption of a system of thought, no matter if modern or postmodern, implies the adoption of its correlated epistemology. If postmodernism was to establish itself, marketing could no longer retain the current features and contents, nor defend its approach, which have been consolidating in the last 50 years. The development of the discipline is based on the attempt constantly to bring suitable contributions to companies in order to develop successful market approaches. The contents of marketing have been the object of a constant process of refining aimed, in particular, at providing support to management. The analysis in the next pages will highlight how the extreme finalisation of marketing has partly diverted researchers’ attention from the theory, and focused it mainly on the method: a distorted mechanism was created which guaranteed the scientific nature of the discipline by using scientific methods considered universal and immutable. The focus on the method, derived from the need to make marketing a discipline with an academic status, has created an increasingly marked distinction between the marketing literature aimed at management, and that aimed at the academic community. While the former tends to stress the managerial implications of the contribution (they are, therefore, operative and tangible implications), the latter mainly insists on the adoption of a scientific method, which, in very complex contexts, is often coupled with excessive specialisation, although supported by sophisticated modelling.

Long life to marketing research 387

EJM 39,3/4

388

In an extreme hypothesis, the adoption of postmodernism could therefore result in a considerably important criticism of the “scientific nature” of marketing as an investigative field, at least in the sense so far accepted (Brown, 1998). Such criticism, far from threatening the existence of marketing, could nevertheless give rise to a revolutionary rethinking of the discipline. As a matter of fact, the postmodern thought highlights the role of experience in the construction of theory. Once again, the method becomes a means of support for the theory, not the opposite. This reconsideration of roles means that researchers will lose many long-established criteria for evaluating and judging a theory, and implies the need to rethink the sense of their work and, last, but not least, to redefine their responsibilities. By changing its face, marketing loses the aura of science and becomes a body of knowledge created by the individual for the individual. Both the researcher and the discipline are now faced with a very risky and, at the same time, challenging task. The removal of any reference to the method’s objectivity – or pseudo-objectivity – from the frame of reference that the researcher uses to evaluate any research, implies the removal of any kind of standardised and external support from the research activity. The researcher, thus deprived of any adherence to a standard, is faced with a risky challenge. As a matter of fact, adherence to a method, to its “scientific nature”, to its rigour, has traditionally constituted a safety net for the researcher: the respect for standardised channels, shared by the whole academic community, provided with a guarantee for the quality of its work and its acceptance. At the same time, such a safety net would also become the cage of the researcher who was not allowed to leave without being condemned to the deprivation of the academic status. Instead, embracing postmodernism completely means leaving any net or cage, with much greater risks for the researcher, but also more freedom: freedom from every scheme, from every dichotomy, even from the choice between true or false and allows the researcher to make the experience he/she chooses, and how to make it. There is, therefore, only one criterion to evaluate research. It is enrichment that the research experiences, and not only that, witnessed by the researcher, bring to the knowledge of the individual and of the community. It is clear that the process of creating knowledge is endless and unstoppable since the more the individual (and the community) is enriched, and therefore learns, the more he/she realises his/her lack of knowledge. Knowledge generates knowledge. 1. The body of marketing The origins of marketing take their roots in American management literature in the late 1950s and the beginning of the 1960s, when some researchers started to investigate some management practices and, above all, the origin of market success. Those articles are now considered the landmark of marketing and have established the main concepts upon which this field of human knowledge has been developing for many decades, and is still accepted today. Felton (1959), for example, reported the lack of awareness of the processes that allow American companies to reach pre-defined market targets. In the light of the failures of companies to approach the market, the author gave some suggestions to create a successful marketing concept, defined as a mental approach, based on co-ordination as well as on integration among activities and functions, and aimed at obtaining long-term profits. Borden (1964) consequently operationalised this concept by transforming it from a mental approach into an ensemble of elements – tools – at the marketing manager’s disposal to create marketing strategies. It is a set of tools in which every marketing action program can be dismantled in the final analysis. The concept of marketing mix was then defined. This concept has since been

considered a pillar of this discipline[1] thanks to its simplicity and the possibility it offered of liaising the concept of economic value (which is crucial in marketing) to the actual managerial action. Over the years, these approaches have been the object of a constant process of systematisation and refining. The activities that can be ascribed to the market domain have been evolving in the course of time, drawing frequently from other disciplines; let us think, for example, of the market segmentation which draws its technique from psychology and statistics, or of the study of purchaser and customer behaviour which availed itself on psychological, sociological, and economic principles and models, and so on[2]. Besides the specific cases, the whole body of the discipline has been the object of constant refining in order to define tools and technique to meet the needs of the market and, at the same time, those of the manager. This process has strongly contributed to the spreading of the discipline, thus creating a language universally shared and recognised as principles of marketing management by the academic community, the practitioners and, last but not least, by undergraduate and postgraduate students. The origins of these principles should obviously be traced in the “marketing concept”, which has been itself the object of a process of constant refining. From the first basic definitions of the 1960s, the concept of marketing has evolved and become an ensemble of fundamental principles, which can be summed up in the expression “customer orientation”. Customer orientation or market orientation makes the customer the main and fundamental point of reference for the company: the latter plans and realises its offer, which is aimed at satisfying the customer in the best possible way, only in relation to its needs. The way to the customer’s satisfaction is, in fact, successful in the long term, as it is able to lead the company in the uncertain future (Olivier, 1996; Reichheld, 1996). Customer orientation is thus the management philosophy, which legitimises marketing actions, and makes the whole coherent and harmonious. Within this philosophy the process of marketing acquires a sense, starting from the analysis of the environment and its components (competition and market included) to the adoption of strategic measures – segmentation and positioning – which become effective in operative decisions, the so-called “marketing mix”. Since the 1970s, this definition of marketing has met a growing, rapid and general consensus, thus transforming it to an evergreen or, in a “marketing megalomania” or even in a “Kotlerite” (Brown, 2002) as defined by the critics. The considerable simplicity of use of marketing mix, which was initially created to translate the marketing concept into operative terms, has been one of the main reasons of the great diffusion and credibility of marketing as a discipline on a global scale[3]. However, this has caused a stagnation in research and studies concerning marketing. As a matter of fact, research projects have almost exclusively been conducted within the frame of reference, which has been described above, and theoretical models and operative references allowing businesses better economic performances have been developed only in that context. In the following decades, the frame of reference remained the same, unchanged; for so long such was the interpretation of marketing mix as a revealing and unquestionable reality that no criticism concerning its effectiveness has been put forward. A clue to this phenomenon of generalised uniformisation could certainly be that the principle of marketing management retains a high degree of clarity, simplicity, usefulness and applicability. Besides this explanation, however, there are others deriving from the sociological interpretation of knowledge (Bloor, 1976) known otherwise as social epistemology (Schmitt, F., 1999). In this sense, knowledge would be “what is collectively sanctioned as such” (Bloor, 1976, p. 9). Science would not therefore be really objective, but

Long life to marketing research 389

EJM 39,3/4

390

inter-subjective (Toraldo di Francia, 1993). According to this perspective, the constant process of improvement in the discipline becomes a declaration of belonging to a social group, the international academy of marketing, and a process of social recognition of its members, which is, in the final analysis, nothing but an example of isomorphism. The perceived need of improving the effect of marketing policies, in a context whose evolution is increasingly faster and less intelligible, have necessarily induced researchers to increase their specialisation, thus becoming great experts of single tools, and marketing aspects[4]. The evolutionary trend of marketing contributions has been fostered by the editorial choices of A journals which tend to publish very specialised papers, supported by solid empirical analysis, but which turn out to be scarcely comprehensive. For this reason, they have also been critically labelled as journals of marketing obscurity (Piercy, 2000; Baker, 2001). If on the one hand this trend tends towards a specialisation of competencies and allows the discipline to progress in a “scientific”[5] way, on the other hand there is a considerable risk of losing sight of the conceptual frame of reference of the domain in which such competencies are applicable. Also owing to this tendency, marketing has been the object of strong criticism in the course of time, which nevertheless has not changed its initial general approach and has only affected the following phase of its evolution. This relatively simplistic approach to marketing has transformed both the general theory of marketing and the consumers themselves into victims. Consumers have been reduced to mere numbers (that is quantitative data) by marketing operators, not to mention university students who have been forced to learn the principles of marketing as a simple recipe book made of ingredients that can be mixed, and formulae (which most of the time are quite rigid) that can be applied according to the circumstance. In most businesses, many marketing activities are still today an exclusive task of some specialists who are considered the only persons responsible for the firm orientation, and who have only to apply formulae and recipes learned in their education and refined by practice (whose partial dynamic and heterogeneous matrix has been recognised)[6]. 2. Criticism directed at marketing 2.1. Two important kinds of criticism directed at marketing The scientific advancement proceeds in the domain of human knowledge through what is now a standard process: theory, criticism and new theory. Marketing follows this general approach as well. In practical terms, this means that every marketing contribution – as for every kind of discipline – starts with the analysis of literature, pinpoints a critical point due to a poor correspondence between theory and reality, and continues the reconstruction of knowledge for that specific area, thus contributing to the improvement of the knowledge of society. In marketing literature, it is possible to trace some critical approaches that share common features and give rise to actual movements for the refounding of the discipline. In particular, the main trends, which have addressed marketing with strong criticism, are two: relationship marketing and experiential marketing. They both have accused the discipline of involution, and of being devoted only to the modelling of interpretation schemes of reality which have proved to be too far from it, and were therefore not suitable to provide an exhaustive and explanation, which can be generalised. Chronologically, the first trend to cause a crisis within the discipline was later called relationship marketing. During the 1970s, a part of marketing literature started to question the object of the discipline and its extendibility to other realities. In particular,

the Swedish School of Industrial Marketing and the Nordic School of Services have contemporarily criticised marketing by maintaining that it adjusted well to the exchange relations of the mass consumption goods market, for which it was initially studied, but lost analytical and interpretative effectiveness when used exactly in the same way in other kinds of situations, especially in the industrial goods and service industry. The mass consumption goods market is characterised by a strongly atomistic demand in which the personal features of the purchaser lose relevance, giving space to anonymous and homogenous expectations. In the most sophisticated cases these can be analysed, through segmentation techniques that are sometimes quite refined. According to this scheme, the consumer is clearly passive and subjected to the company policy without the possibility of affecting it in any way. The only possible action is the choice among predetermined supply alternatives. Exchanging power is therefore asymmetric and unbalanced; the single purchaser does not have decisional weight since his/her contractual force is proportional to the percentage of his/her purchase in relation to the total turnover of the company and is, therefore, almost nil. According to the representatives of the “relationship” vision, the situation described above is considerably different in the industrial goods and service market. The peculiar features of this industry make it a different kind of market altogether in which the customer retains a particular and active participation and emerges as a consumer, producer and production resource. This calls for a reconsideration of marketing. As far as the market of mass consumption goods is concerned, the literature had put the exchange at the centre of the relation between demand and supply, and consequently at the centre of the analysis too, the new reflections. They had been developing between the 1970s and the 1980s, and they replaced the concept of exchange with that of relationship, which is established (in a more or less continuous way) between the purchaser and the seller; in analysis this is what really counts, not the single exchange act (which is often sporadic). Both the Swedish School of Industrial Marketing and the Nordic School of Services stressed how crucial the role played by the long-term perspective is in the management of these markets. From this common point, both schools have independently developed their line of thought: industrial markets researches have focused mainly on the relations among companies, in particular on the role of trust and on the concept of relationship network (Ha˚kansson and O¨stberg, 1975; Ha˚kansson, 1982; Jackson, 1985; Halle´n and Sandstro¨m, 1991; Ganesan, 1994; Morgan and Hunt, 1994; Doney and Cannon, 1997; Smith and Barclay, 1997; Duncan and Moriarty, 1998). Services researchers have concentrated on the differences in services in relation to goods and, in particular, on the continuous and necessary interaction between producer and consumer (Berry, 1980; Normann, 1985; Turnbull and Valla, 1985; Gro¨nroos, 1994; Vavra, 1995). Moreover, in the course of the time, the importance of the relationship approach has spread in the consumer markets as well, thus making it necessary to consider the consumer perspective in all marketing choices. In the light of the specificity of the new analysed contexts, these authors have highlighted the weak points of the traditional approach -to marketing, and defined it “traditional marketing” or the “traditional paradigm” of marketing, which is not suitable in the contexts in which the firm can pinpoint the counterpart and treat it individually. The second and important criticism to “universal” marketing was put forward by the trend of experiential marketing, some years later. The experiential interpretation of consumer behaviour started at the beginning of the 1980s in contrast with the traditional and prevailing view of studies on consumer behaviour, whose first contributions date back to the 1960s, and constitute what the experiential authors

Long life to marketing research 391

EJM 39,3/4

392

consider an utilitarian view (which is still today the major research trend in consumer behaviour). Since the middle of the 1980s, some researchers have started to suggest an extension of the consumer behaviour interpretation, highlighting some limits of the utilitarian view of thought, such as the thesis of univocal rationality of the individual (Hirschman and Holbrook, 1982; Holbrook and Hirschman, 1982). By focusing on the mere act of purchasing, the utilitarian view has highlighted the rational component that leads the purchaser toward the resolution of the decisional problem he/she is faced with – a problem of choice among product alternatives. Resolution of the decisional problem is, in fact, an area, which can easily be the object of a rationalistic interpretation of consumption, especially of a sophisticated modelling which becomes sometimes exasperated. Therefore, if on the one hand consumer behaviour researchers have acquired a considerable store of knowledge concerning the issue, on the other they have almost completely neglected all the other aspects of consumption which do not have a rationalistic component, especially the interaction between consumer – and non-purchaser – and product. This is the real experience of consumption whose definition is, by nature, elusive and difficult. The necessity of combining different disciplines, which, in the course of time, have been consolidating as incompatible, in order to come to a more comprehensive idea of the individual and of its choices, has recently been recognised with the 2002 Nobel Prize in Economics being awarded to Daniel Kahneman for having introduced elements of psychology research in economics sciences. Although the experiential view openly criticised traditional marketing only in 1999 (Schmitt, B., 1999), the first critical attacks could actually be dated back to 1982 when Hirschman and Holbrook carried out an initial comparison between the traditional and the experiential approach to the study of consumer behaviour. The two researchers, who are the pioneers of this trend of study, which has slowly gained agreement and support in the course of time, have ascribed the differences of the two approaches to the mental construction used, the categories of the analysed products, the use of product and finally the consideration of individual differences among individuals. They have carried on this comparison by defining the essential features of the experiential interpretation of consumer behaviour. Criticism of the traditional approach concerns, in particular, the thesis of customer rationality and utilitarianism. According to the traditional theorists of consumer behaviour, founders of the traditional approach (Nicosia, 1966; Engel et al., 1968; Howard and Sheth, 1969; Holloway et al., 1971; Walters and Paul, 1970), the behaviour of the consumer is regulated by a general rationality which allows every decisional problem to be resolved easily, in particular, the purchasing decision, in order to pinpoint the supply which maximises the utility for the consumer. In this sense, the object of these researchers is the decisional process, which leads an individual to make a specific purchasing choice, with the final objective of creating, with the same process, a universal model of reference (Zaltman and Wallendorf, 1979). It is clear that the origins of such interpretation of consumer behaviour could be traced to the utilitarian view of the general economics theory (Sherry, 1991). Bernd Schmitt (1999) resumes the same process of comparison analysis by slightly modifying the categories compared and, especially, by highlighting once more the contrast between the traditional and the emerging experiential view[7]. In Schmitt’s contribution, however, the comparison analysis regards traditional marketing, an expression that seems to define the ensemble of principles, models and marketing management tools. In any case, the experiential view develops initially in an

antithetical way in relation to the prevailing trend, thus constituting a real reaction to the traditional consumer behaviour model, and aimed at a revision of models and tools in order to improve adherence to reality. It is, in fact, with the objective of studying the consumption behaviour of hedonistic products (considered as not strictly “rational”) that the concept of experience is defined, giving rise to the importance of individual emotions (Caru` and Cova, 2003).

Long life to marketing research 393

2.2. Criticism to marketing: an analysis At first sight, the two main critical ideas that have been described do not seem to have anything in common. As a matter of fact, their differences are undoubtedly quite consistent: first of all, the cultures in which they were born are different – if relationship marketing comes from Northern Europe, the experiential marketing comes from North America. Second, the object of their criticism, and therefore of separation from the traditional approach, is different. Relationship marketing criticises the extendibility of traditional marketing to all industries. On the other hand, experiential marketing’s attack regards the issue of the consumer’s rationality on which the development of consumer behaviour has traditionally been based. Finally, their terminology is different, and it is clearly the case of the traditional marketing concept that we have mentioned. Despite these clear differences, relationship marketing and experiential marketing share some common elements. Both trends of study have attacked marketing by following very similar lines of reasoning. Relationship marketing has developed from the analysis of the applicability of traditional marketing models in new realities, such as service and industrial goods sectors. Similarly, the first studies that can be attributed to experiential marketing have noted that long-established models of analysis of consumer behaviour could not be suitably applied to the study of particular situations, especially as far as hedonistic products – which are characterised by a strong emotive component – were concerned. In the course of time, the attention of researchers has grown and included a wide range of products, even those that were not typically hedonistic (Addis and Holbrook, 2001). In any case, services, industrial goods and hedonistic products (and not only these) account for specific realities, which cannot be analysed by using long-established models and tools (which had universalistic vocation and rationalistic applicability). In other words, these realities behave differently compared to what is coded, described and prescribed by the “general” models of marketing. The scientist, therefore, cannot neglect this lack of correspondence between reality and theory, and should intervene to fill the gap since “specific realities”, unusual, are now clearly more consistent that those that have been so far considered in literature. This first common element can be added to a second one, which is no less important: the scheme of reasoning. As a matter of fact, both trends start by reporting the heterogeneous nature of reality and consequently suggest the enlargement of models and principles, therefore theory, as the only possible way of allowing marketing to be sufficiently adherent to the different situations studied. Only by intervening on the theory can we fill the gap: only with a rethinking of theory is it possible to widen the range of models and tools to make theory a valid means of understanding and interpreting reality. In this perspective, reality is considered a fact, which is external to science and on which science cannot intervene. The task of science is only to analyse reality in order to interpret it and to know it. The process of knowledge in this case can be related to the process of appropriation: knowing something means taking possession of it and therefore “owning” it. Through the scientific process, theory takes

EJM 39,3/4

394

possession of the surrounding reality and dominates it, which means that science owns reality through knowledge. These first common elements do not refer to the contents of the discipline, but to the scientific process of discovery. As a matter of fact both currents of thought adopt, similarly to traditional marketing which is the object of their attacks, the process of the traditional scientific discovery. A third and last common element of the two trends is given by the kind of criticism addressed to the discipline. Both of them, in fact, criticise specific and particular aspects of the discipline: relationship marketing attacks the use of the model of marketing mix, which is uncritical, lacks any logic, and has become a simplistic recipe to manage a reality which is far more complex than what is assumed by theory. Experiential marketing attacks the development of the traditional model of marketing in general and of consumer behaviour in particular by reporting different weaknesses and especially some unrealistic theses[8]. They both have criticised the implementation and development of the discipline – in the first case the implementation of the marketing concept has been attacked, whereas in the case of experiential marketing the implementation of consumer behaviour study was criticised. However, neither of them has questioned the element that actually creates a difference in the approach: the underlying system of thought. Traditional marketing originates from a neopositivist view of the world, whereas relationship and experiential marketing are, more or less emphasised, postmodern (perhaps without being aware of it). Indeed, relationship marketing deprived the company of the immeasurable power that mass marketing had attributed to it. The rebalancing of the power between company and consumer agrees in fact with the postmodern interpretation, which states the absence of any hierarchy. Experiential marketing is even more postmodern since, besides considering consumer and companies as equal counterparts of an interaction, it strongly attacks the thesis of consumer rationality and univocity. Both relationship marketing and experiential marketing are, therefore, in clear contrast with the prevailing trend as their general approach is postmodern. At the same time, however, despite the basic differences, they all follow a similar scientific method. At this point, one should question the role played by the research method in marketing, and its contribution to the “scientific nature” of the discipline, and find out if it is possible to claim that the body of knowledge of marketing is scientific. The question that the marketing researcher is faced with can have destructive effects on his/her identity (Piercy, 2002) but, at the same time, offers freedom of thought and action. To try to give an answer to this question, we need to analyse what postmodernism and modernism are and what postmodern marketing can be. 3. Modernism and postmodernism 3.1. Modernism Although the term modernism refers to a system of thought that has developed in the course of the last four centuries, its actual definition can be traced to the last decades, when postmodernism emerged, thus contrasting with the previous one[9]. In order to understand the postmodern system of thought, whose unsettling effects are having an impact on every aspect of human knowledge, it is necessary to start with analysing modernism. Modernism is the view of the world, which made its mark on human action in the modern era. The latter is conventionally said to have started with the Industrial Revolution and experienced its highest moment between the nineteenth and the

twentieth centuries. In Europe, the first Industrial Revolution, and even more the second one, gave rise to a process of well-being and improvement in general living standards, which seemed unstoppable. This process was led by reason. Human thought was obviously affected by this advent, which resulted in “modernism”, gathering together the philosophical currents of Neo-positivism, Logical Empiricism, Logical Positivism and Neo-empiricism thinking; dating back to Descartes and Kant, Smith, Locke and Hume, the members of the Positivism movement are generally considered the pioneers of modernism which received a considerable contribution from Newton’s research (Cobb, 1990; Abbagnano, 1995). In the modern perspective, the recognised ability of the individual to understand nature, reality and its truths, allowed him to intervene on the state of things and to guide and improve them. Thinkers and researchers therefore aimed at defining the laws regulating economic and scientific phenomena in order to understand their applications, allowing, above all, their replication and improvement (Chiurazzi, 1999). A finalised pragmatism dominated scientific analysis and its disciplines. Knowledge advanced toward reality and truth. Knowledge was aimed at the “good” as it was fostered by the certain truth of the real. The will to reach increasingly higher levels of well-being necessarily extended the concept of science to every discipline of knowledge: the mere application of the “scientific” method transformed every discipline into science. At the beginning of the twentieth century, in fact, with the birth of psychology, sociology and psychoanalysis, the rationality of the individual was further valued and emphasised. In the light of the breakthrough obtained by mankind, the modern term has acquired strongly positive meanings, thus coinciding with the term “advanced”. Today, instead, the term modern indicates a past era that has ended, at least for those who are more sensitive to social change (Cobb, 1990). 3.2. From modernism to postmodernism In the second half of the nineteenth century, some philosophers – Kierkegaard, Nietzsche and Heidegger above all – started to doubt the inflexible faith of their contemporaries in rationality, and in the ability of defining, circumscribing and knowing the truth (Jackson, 1996; Best and Kellner, 1997). Their thought emerged and were developed by a group of French philosophers connected with Poststructuralism – Derrida, Foucault, Lyotard and Baudrillard to mention the most famous ones – who are today known to have been the first postmodernism theorists (Williams, 1998; Chiurazzi, 1999). It was only during the 1980s, however, that their thought began to spread all over the world. In Kierkegaard’s thought there was already strong criticism regarding the faith in human rationality, as well as every reflection aimed at knowing what is real and what is true. These concepts, according to Kierkegaard, imprison mankind and trick it into possessing certainties, thus destroying feelings, inspiration and spontaneity. Those make up the essential part of human beings and their inclination toward God. Nietzsche’s criticism concerning modern thought is even stronger as it lacks any religious reference. Nietzsche extols individuality, its power and autonomy in strong contrast with any form of aprioristic, immanent, rational, definitive, and in any case salvific, ideology. According to the philosopher, true knowledge is the simultaneous existence of a multiplicity of interpretations, each of which is the result of a particular perspective that is essential, and should therefore be valued; this manifold knowledge leads the individual to appreciate difference. However, it is also the result of a long process requiring considerable effort and a will to learn, an unappeasable and humble desire to know, where the knowable is endless as all knowledge is the

Long life to marketing research 395

EJM 39,3/4

396

source of other consequent researches and knowledge. Knowledge generates knowledge. Through the knowledge of the new, the individual, who is free and not afraid of not possessing a truth, questions him/herself, and encounters risks, but can find fulfilment. The third father of postmodernism is Heidegger. In his opinion, truth is not what corresponds to reality but it is freedom. And freedom is, first of all, freedom in language, which is a privileged manifestation of being. As a consequence, Heidegger gave to art and poetry, neglected by modern view, a particular meaning as he considered them an alternative way of learning, and therefore of knowing. During modern times, truth and knowledge are searched for at the expense of the individual attributes, which are, instead, his/her expression and richness. 3.3. Postmodernism The thoughts of those philosophers, sometimes fragmented, that constituted an opposing trend during the nineteenth century, became an actual system of thought during the following century, when historical events highlighted the frailty of human certainties. They have started to revise the idea of progress, according to the evidence that the objective of a minimal level of wellbeing, common to all social classes, cannot be achieved, which is, on the contrary, the result of a concept that can be relatively historicised and questionable, certainly not absolute. During the 1960s, the voice of Gilles Deleuze, Roland Barthes, Jean Baudrillard and Julia Kristeva, rose in unison against the certainties of rationality. A few years later, the poststructuralists’ thought became part of postmodernism. The word “postmodernism” does not contain a precise meaning, and refers to many fragmented cultural phenomena, to the extent that some have suggested the need of using the plural and therefore of referring to “postmodernisms” in line with the postmodern spirit (Featherstone, 1991; Brown, 1994, 1995, 1997; Chiurazzi, 1999)[10]. In spite of that, it is possible to recognise in this complexity, fragmentation and even unknowability of reality, which was so far defied by modernism, the central element of the new philosophy. The very same concept of reality is then questioned together with that of truth and any other certainty of modernism (Cobb, 1990). As a matter of fact, each philosopher has developed his/her own thought in a specific way. Deconstructionists, in particular Derrida and Lyotard, emphasised the concept of difference, and highlighted its link with language (Chiurazzi, 1999; Best and Kellner, 1997). Vattimo and Rovatti focused their attention on the critique of rationality, and opted for a celebration of difference and tolerance that was a “week thought”, in fact. Bocchi, Ceruti, Morin and others developed the theory of complexity that is the celebration of multiformity as the basis of the world. Foucault concentrated on the subject and denounced its submission to society and to its false constructions. Rorty analysed western philosophy and defied its role in the political life of society, which results as lacking in its own critical interpretative conscience. All these trends of thought, although with their own peculiarities, claimed the validity of the differences between historic periods, geographic places and single individuals. There is no core, no structure to be known. Every single thing cohabits with the other, without a precise aprioristic and absolutist meaning. They are consequently the object of interpretations that are sometimes opposing, but opposition and therefore diversity is the real richness of the individual. If on the one hand, modernism gave some answers to the important issues of mankind, on the other, not only did postmodernism fail to offer any explanation but it also created new questions. Questioning leads to knowledge. Knowledge is the richness

that generates more richness through creating new questions. This upgrades new research, and therefore new knowledge. As the knowable is endless, the possibility of enrichment does not have any limits apart from the ability and will to know. There is no beginning since there is no end. The concept of the end is a limit that the individual cannot afford to define. The context in which these reflections are present is uncertain, but is also productive as far as knowledge results are concerned and therefore offers real richness. The implications of postmodernism over the disciplines are different and unsettling (Featherstone, 1991), but those concerning the concept of science account for the most interesting and relevant issues. 4. What possibility is there for a postmodern science? 4.1. From the scientific nature of method to the scientific nature of discipline The philosophical thought that modernism refers to is mainly that of logical empiricism, which has its foundations in positivism and in the support given to scientific research as the method to be used. As a matter of fact, according to thinkers who expressed these positions better (in primis Carnap and Hempel), logical analysis and method are the foundations on which knowledge is built, while physics is the science par excellance. Two main consequences derive from this concept. First, all other sciences, even the social ones – marketing included – can and must adopt the physics method because it enhances them (Rosenberg, 2000; Salmon, 2000). The slogan associated to logical positivism called Principle of Verification is revealing: “The meaning of a statement is its method of verification” (Ray, 2000). Second, all sciences share a method, which seems enough to state the unity of science (Hooker, 2000). The principle of linear and progressive evolution and the advance of science imply that knowledge develops by subsequent achievements and stratifications. New knowledge combines with the previous one, it improves it and makes it truer but not different. No different knowledge referring to the same reality can exist. As a matter of fact, a good researcher observes reality – namely all phenomena – and compares it with the interpretative models already codified in knowledge in order to verify if they are valid to understand the phenomenon he/she is analysing. Infringement of those patterns allows the researcher to look for new ones, which, after having been verified, are seen to be hypothetically more suitable in analysing reality. This is called the Correspondence Theory of Truth, which dates back to Aristotle and has now become the common theory in modern epistemology (Lynch, 2001). Thus, a scientific discovery is the product of human desire to know and dominate his/her environment, and “to go beyond” him/herself in order to achieve an unrestrainable – but hypothetic – progress in knowledge, which is based on approximations to the Truth. This is the principle of Verisimilitude, well defined by Popper (Brink, 2000; Shapere, 2000). Two consequences derive from the acceptance of the method described above: (1) The employment of the scientific method draws the difference between science and non-science. (2) New theories are by their nature better than the old ones, in that they go beyond the old ones by denying them. The first implication has its origin in the rationality category, which is the foundation for the development of modernism. The use of reason, and the use of the scientific method, draws the line between science and non-science; in other words it is able to transform any discipline – even marketing – into science.

Long life to marketing research 397

EJM 39,3/4

398

The second implication deriving from the scientific method is related to the linear interpretation of the development of knowledge. If the development of new knowledge must start from the one that has already been codified in order to go beyond it, it goes without saying that the new knowledge will be better since it can explain something more, even if just a little bit more, about reality, and thus it can explain it in a “better way”. What is new is by definition more advanced, and therefore goes beyond the quality of what had already been acquired. 4.2. The deconstruction of the concept of scientific nature In the postmodern interpretation of epistemology, both the concept of scientific method and its two connections are attacked. Obviously, the criticism of modern epistemology expressed by postmodern philosophers is related to their worldviews. Among the main concepts of postmodernism there is the criticism of the modern outlook on the world, which is considered an objective reality, something that is “out there”, patiently and passively waiting for individuals to get to know and consequently possess it. Postmodern reality is something else; it is nothing but a context in which individuals, together with millions of other creatures, act. Furthermore, by acting and using both their rationality and intuition, and at the same time exploiting chance events, they actively contribute creating and changing the so-called reality, namely to change the way they look at it. Therefore, modern attempt to identify absolute and ideal truths and values, which are effective no matter when and where, is considered absurd as everything is part of a given context. And since knowledge belongs to the world, the world is changed by knowing the world. Consequently, the modern system of thought needs to be “deconstructed”, that is to say, that its dangerous incongruities and false contradictions have to be exposed in order to create new concepts. If there is no absolute truth to be understood and uncovered, science cannot aim at objectivity, nor at the complete understanding of phenomena, but it has to proceed through experiments and attempts and be content with partial and transitory knowledge. The concept of the stratification of knowledge loses meaning; past knowledge is not strictly necessary for future knowledge and, indeed, it could hinder it. Therefore, the concept of truth cannot have a transcendental origin. On the contrary it is necessarily related to contingency. The kind of contingency that can be referred to changes from observer to observer, from analyst to analyst, from “scientist” to “scientist”. Foucault considers contingency as the socio-political group in which knowledge has developed. When an individual judges a phenomenon to be true or false s/he simply expresses his/her belonging to a social group, therefore his/her political interest. On the contrary, Lyotard considers language as the context in which knowledge develops. Different languages are indeed preserved through dissent, difference and contradictions, while the continuous process of innovation creates new knowledge that add to the previous one and does not replace it tout court. Finally, according to Rorty, knowledge must be contextualised to place and time: truths are valid in their places and at their times. Rorty criticises any theory that claims the right to use a language affirming that it describes reality better than others. If language performs the most important role, literature becomes the favourite field for this experimentation[11]. By seriously attacking the concept of absolute truth, postmodernists attack Popper’s epistemology too, which today is common ground for all sciences. According to Popper, the absolute truth exists but cannot be known completely by the individual who must accept a continuous and ameliorative process that brings him/her closer to an asymptote of knowledge. On the contrary, according to postmodernists the absolute truth does not exist and cannot exist because it would constrain knowledge to a limited

dimension, which is the most evident contradiction of the concept of knowledge itself, and of learning. Implications for “modern” sciences arising from all this are destructive. First of all, the modern criterion to evaluate and assess theories can no longer be used. The complexity of reality is so paradoxical and controversial – but true and authentic too – that a hierarchy in the validity of scientific theories cannot be determined[12]. Each scientific theory aims at knowing only one part, or one moment of reality, which anyway is highly dynamic, complex and furthermore built by the researcher’s actions and, thus, the knowledge of that part or moment is in any case always limited. Namely, the idea of scientific progress is denied in the “modernist” meaning of the term. The consequences following the denial of the process of stratification add to those deriving from the criticism of rationality. As a matter of fact, the concept of rationality loses all absolutism and priority, and consequently the definition of science must be revised. No verifiable predominance of human rationality on emotionality exists, let alone on the dark side of mankind, and on chaos. The rationality of modern thought has been put in the foreground only because of its relatively easy intelligibility, and because the individual aims at going beyond his/her limits and getting closer to the transcendent, which is supposed to be “perfect”. In any case, the postmodern individual has no interest in understanding which tool, or method, makes it possible for him/her to acquire knowledge, because he/she sets a totally different value on knowledge compared to the modern individual anyway. Since the human being is one of many creatures living on the Earth, he/she has no right, and probably, no possibility and no interest, to dominate nature. In the same way he/she has no right, and no possibility, to dominate reality. Furthermore, this kind of knowledge is not a dominion but a relationship (Zagzebski, 1999), and in this relationship the researcher is not the only active subject. He/she is an active subject together with reality itself, of which the researcher is an integral and irreplaceable part, because reality “is born” with the researcher’s essence. That is to say, that a postmodern science can exist, but with a different meaning than a modern one. First of all, science is not only led by rationality. Furthermore, since rationality is a limited cognitive tool, science deriving from it cannot have absolute value but can only refer to a certain context, which is in any case relative. Consequently, postmodernism does not lead to the feared death of science or the “disaster of hope” as someone called (Petitot, 1993). On the contrary, postmodernism allows individuals to be potentially free. However, it also places all the responsibility of their freedom on them, and for the first time, only on them. 5. The modern past of marketing The modern and postmodern outlook on reality, whose main features have been defined above, can also be applied to marketing. As a matter of fact, an analysis of the existing marketing literature allows the “modern” features of this discipline to be highlighted, features deriving from its deep finalism. In fact, the current development of marketing is based on settled and highly modern foundations. The debate on the opposition between modern and postmodern marketing becomes part of the previous debate on the nature of marketing – between art and science (Hunt, 1976; Anderson, 1983; Brown, 1997). In this work, however, that debate has a secondary role because the main aim is the analysis of the impact of the postmodern outlook on marketing, with particular care given to the implications for marketing. Although, for about ten years, renowned researchers have been calling for a complete rethinking of the assumptions of marketing from a postmodern point of view

Long life to marketing research 399

EJM 39,3/4

400

(Firat and Venkatesh, 1993; Firat and Shultz, 1997; Brown, 1999), the whole marketing system of thought is still clearly modern, where modern refers to the historical and philosophical assumptions widely explained above – in opposition to modern too. In its managerial function, marketing helps companies to define and implement their approach to the market by offering tools to manage the value of their offer for their clients, and the value of their clients for themselves. In other words, at the core of this discipline there are companies and their actions and contacts, regardless of their kind, with a market characterised by demand and competitors. Therefore, since its general foundation marketing has been characterised by a clear opposition between a main subject – the company – that must act in a certain way in the market, and an object of context – the market itself – to whose dynamics the company must react by trying to anticipate, determine and finally dominate its evolution. All marketing topics studied in literature – namely the buying behaviour of the consumer, the analysis of competitors, the product or distribution policies, and others – are based on the binomial company-market[13]. In fact this binomial is evidently dual and modern, and its first term, the company, has an individual role and a positive value, while its second term, the market, has a complex role, which is often negative. The aim of marketing is to help the company’s “marketing” function in particular, and the whole company in general, to interact with the market by supplying concepts from which operative tools can be derived. The knowledge produced by marketing researchers aims at enhancing the dominion of the company on the market: even in modern marketing knowledge and supremacy are highly interrelated concepts. The assumption is the identity between supremacy, market power and profit. As a result, marketing becomes an economic discipline, and this feature can be “forced” on the company, but it becomes extremely restrictive for the consumer whose multi-dimensionality is essential to understand his/her behaviour. If both pure research and applied research are generally considered two correlated research categories, marketing seems to lack the phase of “pure” research[14]. At the most, seeing the concept of finalisation with a sense of proportion, without eliminating it, in the managerial sciences a distinction can be made between research aiming at generating knowledge regardless of its application to the managerial world, and research clearly aiming at the development of managerial implications. In any case, if the finalisation of the discipline is too extreme, to the extent that the discipline ends up justifying its own existence, from a postmodern point of view its “scientific nature” could therefore be questioned. In order to avoid a contradiction in terms – a discipline claiming to be a science and in the meantime existing only as a tool for companies; large investments have traditionally been made to refine research methods by making them as sophisticated as possible and similar to those of natural sciences (the problem of the uniqueness of science remains in the background, unsolved and generally untackled). The attention of marketing researchers gradually shifted toward the method itself, in an attempt to make this discipline “scientific”, thus acquiring that academic and social legitimacy that its mere content does not confer. In literature, on the contrary, there are some suggestions that marketing is a science that can produce “knowledge for the sake of knowledge”, and therefore, that it has no practical and useful implications (Hunt, 2002, p. 306). Such suggestions come from the supporters of the scientific nature of marketing, those who have frequently defended this thesis with the attempt of deleting the finalisation of marketing by simply hiding it. One can say that the reason for this defence is to be found in the identification between marketing and the researcher; elevating the former means criticising the latter.

The basic assumption is that the adoption of a rigorous and scientific method, like those of experimental sciences, would automatically make the contribution scientific. In other words, attention has been shifted from the contents to the method in a likely unconscious way. The scientific nature of a rational and “aseptic” method has been used to try to make up for the difficulty of labelling marketing as scientific in relation to the contents discussed. A research of a distinctive identity, that Levy (2002) states is common to other disciplines, is being carried out, and in marketing has become a distorted mechanism subduing the subject of the research and its features to methodological choices, when on the contrary, the opposite should happen (Piercy, 2002). The attention devoted to the research method has been so stimulated by the finalisation of marketing that it has been transformed into excessive modelling. In fact, the need to provide a constant support to companies in defining and implementing their approach to the market has forced marketing to be well ingrained into reality and, therefore, to cope with an environmental complexity which is constantly growing (knowledge generates knowledge, but also questions). As a matter of fact, in order to face the growing complexity of the company-market relations, which can be seen both at competitor and demand levels, marketing researchers have drawn from other disciplines trying, at least, to consider the basic elements of the dynamism of markets. It is in this way that marketing, in different ways and measures, has traditionally referred to political economy, psychology, statistics, anthropology, mathematics and so on. Obviously, however, the widening of the market’s borders and the introduction of many different complex phenomena in marketing have hardly proved coherent with the need for concrete and immediate support to companies: answering questions or preparing hurried recipes? With such strong contradiction, the direction chosen has been to try to analyse complexity, necessarily searching for some logic and breaking it into pieces that, if considered individually, were easier to manage. In other words, there has been an attempt to rationalise the complexity by reducing it to its essential terms. As a result, there has been a gain in terms of management possibility – the theme of knowledge being supreme emerges again – and a loss in terms of richness of the context considered. The advantage obtained for companies was clearly considered to be greater than the impoverishment of the research for the researcher. Obviously it is a short-term policy aiming at bringing up good managers and obsequious researchers, rather than developing the critical ability of reasoning and thinking (Burton, 2001). The attempt to reduce complexity to identified and, therefore, manageable, elements has turned into a project of sophisticated mathematical and statistical models so that the marketing researcher will remain useful to companies, notwithstanding the intensification of the complexity of the context. Therefore, being forced to analyse a more and more complex and dynamic reality with the aim of supplying useful practices to companies, marketing has little by little got closer to abstract models which, paradoxically, turn out to be more suitable for an academic exercise than for a study on reality. In this way, even if researchers want to study complex phenomena, they solve complexity in a set of variables of which they study correlations and impacts. This is the case, for example, when emotions are introduced in consumption (Bagozzi et al., 1999), which has provided a sophisticated rationalisation of their role in consumption through ad hoc statistical models: paradoxically emotions have been rationalised by classifying them like objects. The possibility to apply models and therefore manage their variables is anyway obtained at the cost of an extreme focalisation of the research field, and often unrealistic assumptions. Since the possibility of managing variables relates to patterns, which are so deterministic as to be unrealistic, the aim pursued to

Long life to marketing research 401

EJM 39,3/4

402

serve companies turned into a truly paradoxical abstraction from reality. In other words, the pattern seems to have lost its role as a tool serving the representation of concepts in order to become itself a concept, pleased with itself. That is to say, that a distorted mechanism has been put on which has forced marketing researchers to create more and more rigorous patterns, from the statistical and mathematical aspects, which are in the meantime more and more abstract and theoretical, transforming researchers into “measurement technicians”, to use Gummesson’s (2001, p. 44) terminology. International business schools, marketing reviews and journals, spurred by a paradoxical competitive battle aimed at verifying hypotheses more than understanding the nature and meaning of those hypotheses, have evolved coherently towards these new marketing trends, thus making them grow. Therefore, during the development of marketing, quantitative methods have been preferred to qualitative ones, which gained a supporting role, typically in a phase preceding the actual test of every theory. The level of abstraction reached by this evolution – or, rather, involution – of marketing does not fit with the purely instrumental purpose of “modern” marketing. Some even think that, the separation between theory and practice has consequently widened, and researchers have reached an extreme specialisation “knowing everything about nothing” (Wensley, 2002, p. 392), thus closing themselves even further inside their ivory tower. Therefore, in marketing a narrow-minded approach has become common, which not only fails to relate to the other fields of human knowledge, as was feared for all specialised disciplines (Geymonat, 1972), but also implies a non-acceptance of other positions within the same discipline. The need for studies to become again applicable to the managerial practices of companies has stimulated a process of “translation” of those concepts with difficult mathematical formulae in academic literature into easier and, therefore, more approachable managerial implications that can be used by company managers. From this process, therefore, a duplication of this discipline and its language seems to be derived: on the one hand, the mathematical and statistical language created for and in the international community of researchers, and on the other a managerial language expressively developed for managers and their companies. As a result, there is a duplication of language due to the need for combining two different objectives: (1) The acquisition of the academic and social legitimation of marketing as a science. (2) The development of proposals that support companies in their approach to the market. The double soul of marketing can be also found in the diversity among marketing reviews. On the one hand, there is “A journal” (scientific reviews expressively dealing with research in marketing), on the other, there are more managerial reviews for a larger public that is more interested in practical implications. This duplicity can also be found often with little success among published articles which appeal to both researchers and managers by dealing with matters that could interest the two groups (for researchers a wide analysis of the literature and a clear research project; for managers practical managerial implications at the end of the article). It has to be pointed out that the evolution of the discipline described above has reached extreme positions both in the method, where sophisticated models and software have been created to try desperately to reach a phase of pure research through method, which cannot be reached by contents, and in its implications, by offering lists of “ingredients” to be mixed according to the instructions of a simple perceptive “recipe”. However, both results create the same dilemma: is it a theory? As a matter of

fact, it seems that in both cases no new theory is developed, but only some knowledge which can be valued according to its functioning. In fact, it is not a question whether the knowledge produced is true or false, but only if it works according to the model created (is the model well developed and coherent in itself?) or according to reality (are the managerial implications applicable to reality with positive results?). Enjoying research for the sake of research, stimulated by the capacity of critical reasoning, is therefore overshadowed by an extreme technicality, almost engineering, or by a research aiming at company interests. If this is the current marketing situation, then one must wonder what the scientific nature of marketing and its future will be. And it is to answer this question that one can address to postmodernism.

6. Considerations on the postmodern future of marketing The vicious circle in which marketing has fallen into seems to lead it towards an involution with no a way out. If it is true, as Brown (1997) states, that with postmodernism we entered the era of anti-science, the future of marketing is obscure and difficult to see. Only if marketing researchers acquire responsible awareness will be possible to have a brave re-directing of the discipline. Researchers who tried to deal with postmodern marketing (among whom Sherry, 1991; Brown, 1993, 1994, 1995, 1997, 1998, 1999; Holbrook, 1993; Thompson, 1993; Bouchet, 1994; Elliott, 1994; Firat and Venkatesh, 1995; Uusitalo, 1998; Cova and Cova, 2001) concluded their works by inviting marketing researchers to consider the limits of the modern marketing philosophy, thus joining the “marketing-is-not-working manifesto” (Brown, 2002). Furthermore, they recommended paying attention to the new marketing issues related to the postmodern vision of the world. The most radical among them announced the death of Kotler (Smithee, 1997) and his marketing model, which was declared to have failed (Brown, 2002). A postmodern version of reality is needed to put the individual, both as a consumer and as a researcher, at the centre of marketing, and to give space to imagination in all its forms (Brown, 2002). Perhaps, this is the more likely direction for the future development of marketing. Nevertheless, the individual is no longer the typical individual of modern marketing: their new perspective has radically changed the roles of consumers and researchers. On the one hand, the consumer does not end his/her relations with the company, with the simple purchasing act, nor with the product when it is consumed. On the other, the marketing researcher does not give up his/her study of reality after his/her observations, nor after the knowledge generated with models and implications. Both relate to the context by experiencing it; knowledge becomes relation, not supremacy. “Interactive research” (Gummesson, 2002) is also mentioned to underline the interaction between the researcher and the object of his/her and others’ studies, and all the other components of the research. Anyway, in this work the concept of experience is preferred to the concept of interaction, not because its validity is denied, but because its meaning is widened. In fact, the concept of interaction seems to be related with a temporary, almost static, idea of research. On the contrary, research is a continuous learning process starting from interaction but not ending with it. In the same way, interaction has at least two interacting entities. Instead, even if the concept of experience is more vague, it draws better to multiple shades of knowledge. In fact, the perspective of experience is focused on the single individual and is related to him/her as a parameter to evaluate experience. This is clearly in contrast with the modern interpretation, which considers the group, as the most important referent, a concept that often degenerates into the concept of mass. This does not mean that the social context in which the experience develops is

Long life to marketing research 403

EJM 39,3/4

404

not important. That context is a component influencing the present and future individual’s experience according to the concept of learning. In fact, experience is the core of consumption, and at the same time the core of marketing for one simple reason: experience is the decomposition of the individual’s life. In other words, the experiential perspective is totalising that cannot be applied to everything. For example, the experiential perspective is used in pedagogy, too, where the current of experiential learning theory has developed. The latter has its origin in Dewey, Lewin and Piaget (Walter and Marks, 1981; Kolb, 1984; Merriam and Caffarella, 1991; Frontczak and Kelley, 2000). As consumption is the research field of marketing, the experiential has been studied referring to this moment, but only because the research field is relatively limited. Experience is also important in all other fields of life because the individual’s many experiences, and in each one of them he/she tells his/her history, made up of previous experiences, and thus what he/she has learned. Limiting the analysis to marketing, the value of consumption comes from the consumer’s experience, both for the company and the consumer. And the value of marketing as a discipline also originates from experience, both for the single researcher and for society. The experience of the individual as a consumer conveys to consumption certain symbols and individual meanings which are also created by society and the context in which the experience is lived. Such an interpretation allows the consumer to express his/her personality and mood, which are partly products of his/her past, and partly products of his/her personal creativity and mood at that moment. The diversity and heterogeneity of consumption experiences come potentially out of the studies on postmodern marketing not only because the researcher attaches importance to the diversity of the individual, and therefore to consumption experience, but also because at the same time s/he attaches importance to his/her own individuality. It is therefore a twofold source of diversity: the consumption and the researcher him/herself. Both are products of single individualities created by social interactions, and both are involved in deep experiences: the consumer experiences consumptions, the researcher experiences research. The study of marketing and research are also experiences. Postmodernism allows the researcher to be aware of the human nature of his/her research. And by recognising that, research is experience, and the need to reconsider the “scientific nature” of marketing emerges. As a matter of fact, the experiential nature of research seems to suggest a new way to consider what is scientific. In fact, if one assumes that there is no reality to be discovered “out there” by subsequent approximations, then it is not possible to use the “true/false” principle to judge the scientific nature of a theory, because the reference parameter of this evaluation – reality – has disappeared and is anyway constantly changing, not only for itself but also for the observer. In the meantime, the principle of “usefulness/un-usefulness” cannot be applied to the evaluation of the scientific nature of a theory either, because its use would seem to consider all that is pragmatic as “scientific” and all that does not have immediate implications as “non-scientific”. And even the opposition between true/false is not scientific. What does true mean? What does false mean? How should the partially true be considered? The world has many shades, like the human mind. The adoption of a particular method is the product of the researcher’s choice and it always generates only one part of the possible knowledge. The frailty of every dichotomy becomes clear, which is by definition very easy to apply, but at the same time it does not capture the diversity (Baker, 2002). On the contrary, by interpreting marketing research according to the postmodern logic of experience, it is possible to evaluate the scientific nature of a theory according

to the richness produced by this experience: if the research experience enriches, even in a small quantity, the body of knowledge already codified, the research can be considered scientific. Now the problem is to understand the meaning of enrichment. In order to try to give an answer, one can say that a research enriches the previous codified knowledge when it adds something different or something new, in any case, when it modifies it and makes it more articulate. The comparison between two theories (usually the old and the new) in order to establish which is the best has no longer value: the only evaluation criterion of a theory is the enrichment of the knowledge already codified. The difference between two sciences, therefore, is not in the method, but again in the ability of enriching knowledge. At this point it is clear that every reference to the strictness of research as opposed to its practical importance is no longer valid. As a matter of fact, it is very dangerous to use the strictness of a theory to evaluate its quality, and above all its scientific nature. Again, in fact, in this way the scientific nature depends on the method used, in this case on the method’s strictness, while the real subject which is potentially “scientific”, theory, is left aside. Piercy (2002) describes the wrong use of this reference well, stating that there is no compatibility between the concept of strictness and the importance of research. Focusing on the strictness of research means focusing on the cohesion between the work and the scheme recommended for the employment of the method used. And again, these considerations block any possibility for the researcher to give space to his/her humanity, experience and knowledge, which are essential for generating knowledge. Strictness can contribute to generate knowledge only if it is based on three features of the researcher: intellectual integrity, curiosity and humility. The enrichment of experience celebrates the difference, the heterogeneity, the multi-facets of theories and thus it is the opposite of the impoverishment of knowledge, typical of modernism, which means reduction of the variety of reality and experience to a set of constants and variables to which everything must be reduced. The postmodern perspective of marketing plays down the crucial importance of every research to generalise and create patterns, since specifics and contingencies are more varied and valid than any abstraction. Generalisation is useful but it is never “the whole”. Consequences for marketing could be disruptive: if on the one hand the creation of patterns would lose all value, on the other marketing could lose any perceptive possibility because reality is so heterogeneous that it is not possible to suggest easy univocal recipes, which are good for all occasions and table-companion. This highly increases the professional responsibility both of the researcher and manager. The experiential interpretation of consumption and research redirects the attention of researchers on theory, and therefore on the real essence of marketing, by moving it from method (which is only a tool of theory) and from managerial implications (namely, the operative consequences of theory itself). In this way, postmodernism restores the importance of theory, and therefore of knowledge, by rebalancing the roles of the components. It is now clear that the celebration of differences characterising every vision of the world is the source of the scientific nature of marketing, whilst according to modern perspective rules and repetitiveness originate science. This perspective shift has an impact on all research components. First of all, the field of research widens enormously. In the modern vision, the researcher can enquire into any field where s/he envisages something new or uncovered – either because reality has changed or because new tools allow him/her to see a new reality. On the contrary, in postmodern vision the marketing researcher can deal with everything arousing his/her interest, and to which his/her accumulated knowledge can be applied. As a matter of

Long life to marketing research 405

EJM 39,3/4

406

fact, a new theory is not originated, at least potentially, by the choice of the research field but by the product of the research itself, which must be considered in order to evaluate its scientific nature. This means that a researcher can also study fields, which had already previously been studied. If the researcher adopts a different perspective, or anyway draws different or more in-depth conclusions, then he/she has produced new scientific knowledge. This is a revolutionary interpretation: modernism urged researchers to investigate only what was not already known; on the contrary, postmodernism does not care for research but only for the theory generated. It is a matter of choice between an order without knowledge or chaos full of creative hints. Postmodernism accepts this dichotomy as a challenge, but with humility: truth belongs to nobody, reality cannot be known at a collective level. This implication relating to the field of research is closely related to the effects of postmodernism on the various methods adopted to investigate. Since the discipline does not become scientific thanks to the employment of a particular method, but to the theory generated, the researcher can use any method. Obviously, quantitative methods remain valid, but qualitative methods also can be employed, from ethnography to fiction, discourse analysis and so on, including the most instinctive and restructured and the least rational ones. Science is not necessarily rational (rational referred to what?). In fact, a theory does not become scientific thanks to the employment of a method itself but to the enrichment obtained with the new knowledge. No method is more scientific than others (including Popper’s “scientific” method). On the contrary, all methods can originate scientific theories, and therefore incremental knowledge. Therefore, it is clear that the concept of scientific nature is relative and not absolute. In fact, if the label “scientific” is given to knowledge originating in a research experience, it is clear that knowledge considered scientific is the parameter used to evaluate, and that knowledge itself had enriched a previous knowledge. Hence science is neither objective nor absolute. Every absolute certainty is lost in this conception that does not turn into nihilism of science (the so-called “disaster of hope” highly feared by those rejecting postmodernism), but enormously increases the possibilities of generating theory, in a never-ending learning process. The theory created will never be absolutely scientific, but only in relation to the context in which it is considered and gains value, and in relation to those who get to know it. The peculiarities of every research experience are also expressed in the impossibility to define scientific nature once and for all, since this concept depends on a particular situation of knowledge, on a context, on a society, on time and space. In fact, if the conditions of this evaluation change, the label “scientific” would not be applicable. In order to face the issue of the scientific nature of theory correctly, the conditions defining contingency must be made clear. Despite physical sciences, in social sciences this condition of relativity is emphasised: in social sciences the context conditions in which knowledge develops are subject to continuous changes, whilst in physical sciences the (relative) stability of the context makes the idea of a principle of a (anyway conventional) scientific nature with absolute value more acceptable. The considerations about the generation of theory and its scientific nature are only a first, temporary and relatively imprecise, meditation on the implications of postmodernism in science. Nevertheless, they are enough to reject the accusation of postmodernism as nihilism. Nature is not order, and nobody can create it. In the meantime everybody should cast light on its complexity: perhaps chaos is not accidental. Brown (1999) had already wondered about the possibility for postmodernism to represent the end of marketing and concluded his analysis

pointing out the need to consider the meaning of knowledge of marketing and to reconsider marketing. On the contrary, this paper wants to look further. In fact, even though we know that these considerations have many limits and that it is necessary to study the issues in-depth, nevertheless they are alternative considerations trying to adopt a constructive approach. In fact, one has to remember that the use of postmodernism in marketing is a very difficult exercise, which is neither simple nor automatic, because it attacks any apriorism of marketing. In particular, the researcher assumes more responsibility that can only be tackled with both a curious and humble attitude. Only if research is led by curiosity and is moderated by humbleness, can the experience be enrichment. In this way postmodernism could kill the machine and give freedom – and responsibility – to individuals again.

Notes 1. The term, originally suggested by J.W. Culliton, has been further expanded and introduced in literature for the first time by Borden and Marshall (1959). 2. Burton (2001) has denounced that the interdisciplinarity is more a letter of intents than an actual behaviour of marketing researchers. This is due to both the difficulty of penetrating a vast and varied literature with undefined boundaries and to the absence of incentives. 3. According to Brown (2002), the success of the marketing concept is given to the necessity perceived by marketing individuals to believe in something regardless of its contents. 4. The tendency toward specialisation of the sciences started to spread in France at the beginning of the eighteenth century and later extended to the advanced economies (Geymonat, 1972). 5. Let us accept for the moment the traditional meaning of science; the matter will be tackled more in-depth in the next pages. 6. This phenomenon is clear also from the analysis of text books suggesting lengthy checklist (Gummesson, 2002) of a generally prescriptive nature (Burton, 2001). 7. The considerable timing distance between the above mentioned articles (17 years) and the strong similarities of their contents (comparison between the two trends) show the difficulties encountered by the experiential thought in the process of establishing within the international academic world. 8. For a more in-depth analysis of the comparison between traditional and experiential approach, we refer once more to Holbrook and Hirschman (1982) and Bernd Schmitt (1999). 9. The reader should take into consideration that there is no consistency in the meanings of the terms of modernism and postmodernism and its features. Sometimes they refer to historical periods and sometimes to the system of thought they are correlated with. 10. Some other authors, instead, have disagreed with the clamour surrounding postmodernism, and considered it only a rediscovery of some aspects of modernism (Bouchet, 1994). 11. The social constructionist theory, which interprets the role of language as a creator of the historically determined world and not as a descriptor of reality, derives from Rorty. 12. In particular, first embryonic observations on this matter can already be found in Nietzsche and Heidegger (Vattimo, 1991). 13. It is clear that many researches have tried to bridge the gap between theory and practice, or according to another point of view, the gap between the producer and the consumer (for further in-depth analysis see the special issue of European Journal of Marketing, Vol. 36

Long life to marketing research 407

EJM 39,3/4

No. 3, dedicated to “Bridging the divide”). It is, anyway, a marginal debate of the problem here discussed. 14. The traditionally and commonly accepted definition of pure and applied research is used here: Pure research “aims at discovering nature’s secrets”, while applied, or technical, research “uses such discoveries to solve everyday problems” (Geymonat, 1971, p. 9).

408 References Abbagnano, N. (1995), La Filosofia dei Secoli XIX e XX. Dallo Spiritualismo all’Esistenzialismo, TEA, Milan. Addis, M. and Holbrook, M.B. (2001), “On the conceptual link between mass customisation and experiential consumption: an explosion of subjectivity”, Journal of Consumer Behaviour. An International Research Review, Vol. 1 No. 1, pp. 50-66. Anderson, P.F. (1983), “Marketing, scientific progress, and scientific method”, Journal of Marketing, Vol. 47, Fall, pp. 18-31. Bagozzi, R.P., Gopinath, M. and Nyer, P.U. (1999), “The role of emotions in marketing”, Journal of Academy of Marketing Science, Vol. 27 No. 2, pp. 184-206. Baker, M.J. (2001), “Commentary: bridging the divide”, European Journal of Marketing, Vol. 35 No. 1/2, pp. 24-6. Baker, M.J. (2002), “Guest editorial”, European Journal of Marketing, Vol. 36 No. 3, pp. 293-8. Berry, L.L. (1980), “Services marketing is different”, Business, Vol. 30, May-June, pp. 24-9. Best, S. and Kellner, D. (1997), The Postmodern Turn, Guilford Press, London and New York, NY. Bloor, D. (1976), Knowledge and Social Imagery, The University of Chicago Press, Chicago, IL. Borden, N. (1964), “The concept of the marketing mix”, Journal of Advertising Research, June, pp. 2-7. Borden, N.H. and Marshall, M.V. (1959), Advertising Management. Text and Cases, rev. ed., Irwin, Homewood, IL (first published in 1949). Bouchet, D. (1994), “Rails without ties. The social imaginary and postmodern culture. Can postmodern consumption replace modern questioning?”, International Journal of Research in Marketing, Vol. 11 No. 4, pp. 405-22. Brink, C. (2000), “Verisimilitude”, in Newton-Smith, W.H. (Ed.), A Companion to the Philosophy of Science, Blackwell, Malden, MA, pp. 561-3. Brown, S. (1993), “Postmodern marketing?”, European Journal of Marketing, Vol. 27 No. 4, pp. 19-34. Brown, S. (1994), “Marketing as multiplex: screening postmodernism”, European Journal of Marketing, Vol. 28 No. 8/9, pp. 27-51. Brown, S. (1995), Postmodern Marketing, Routledge, London and New York, NY. Brown, S. (1997), “Marketing science in a postmodern world: introduction to the special issue”, European Journal of Marketing, Vol. 31 No. 3/4, pp. 167-82. Brown, S. (1998), Postmodern Marketing 2: Telling Tales, International Thomson Business Press, London and Boston, MA. Brown, S. (1999), “Postmodernism: the end of marketing?”, in Brownlie, D., Saren, M., Wensley, R. and Whittington, R. (Eds), Rethinking Marketing: Towards Critical Marketing Accountings, Sage Publications, London, pp. 27-57.

Brown, S. (2002), “Vote, vote, vote for Philip Kotler”, European Journal of Marketing, Vol. 36 No. 3, pp. 313-24. Burton, D. (2001), “Critical marketing theory: the blueprint?”, European Journal of Marketing, Vol. 35 No. 5/6, pp. 722-43. ` Caru, A. and Cova, B. (2003), “Revisiting consumption experience: a more humble but complete view of the concept”, Marketing Theory, Vol. 3 No. 2, pp. 259-78. Chiurazzi, G. (1999), Il Postmoderno. Il Pensiero nella Societa` della Comunicazione, Paravia, Turin. Cobb, J.B. (1990), “Two types of postmodernism: deconstruction and process”, Theology Today, Vol. 47 No. 2, July, pp. 149-58. Cova, B. and Cova, V. (2001), Alternatives Marketing. Re´ponses Marketing aux Evolutions Re´centes des Consommateurs, Dunod, Paris. Doney, P.M. and Cannon, J.P. (1997), “An examination of the nature of trust in buyer-seller relationships”, Journal of Marketing, Vol. 61, April, pp. 35-51. Duncan, T. and Moriarty, S.E. (1998), “A communication-based marketing model for managing relationships”, Journal of Marketing, Vol. 62, April, pp. 1-13. Elliott, R. (1994), “Addictive consumption: function and fragmentation in postmodernity”, Journal of Consumer Policy, Vol. 17 No. 2, pp. 159-79. Engel, J.F., Blackwell, R.D. and Kollat, D.T. (1968), Consumer Behavior, 2nd ed., Holt, Rhinehart & Winston, New York, NY. Featherstone, M. (1991), Consumer Culture and Postmodernism, Sage Publications, London. Felton, A.P. (1959), “Making the marketing concept work”, Harvard Business Review, Vol. 37, July-August, pp. 55-65. Firat, A.F. and Shultz, C.J. II (1997), “From segmentation to fragmentation. markets and marketing strategy in the postmodern era”, European Journal of Marketing, Vol. 31 No. 3/4, pp. 183-207. Firat, A.F. and Venkatesh, A. (1993), “Postmodernity: the age of marketing”, International Journal of Research in Marketing, Vol. 19 No. 3, August, pp. 227-49. Firat, A.F. and Venkatesh, A. (1995), “Liberatory postmodernism and the reenchantment of consumption”, Journal of Consumer Research, Vol. 22, December, pp. 239-67. Frontczak, N.T. and Kelley, C.A. (2000), “Special issue on experiential learning in marketing education”, Journal of Marketing Education, Vol. 22 No. 1, April, pp. 3-4. Ganesan, S. (1994), “Determinants of long-term orientation in buyer-seller relationships”, Journal of Marketing, Vol. 58, April, pp. 1-19. Geymonat, L. (1971), “Caratteri positivi e negativi della nuova epoca”, in Geymonat, L. (Ed.), Storia del Pensiero Filosofico e Scientifico: Dall’ Otocento al Novecento, Vol. V, Garzanti, Milan, Ch. 1, pp. 7-16. Geymonat, L. (1972), “L’esigenza di una nuova concezione del mondo e il problema di una nuova cultura”, in Geymonat, L. (Ed.), Storia del Pensiero Filosofico e Scientifico: Il Novecento, Garzanti, Milan, VI, Ch. 21, pp. 1039-67. Gro¨nroos, C. (1994), “Quo vadis marketing? Toward a relationship paradigm”, Journal of Marketing Management, Vol. 10 No. 5, pp. 347-60. Gummesson, E. (2001), “Are current research approaches in marketing leading us astray?”, Marketing Theory, Vol. 1 No. 1, pp. 27-48. Gummesson, E. (2002), “Practical value of adequate marketing management theory”, European Journal of Marketing, Vol. 36 No. 3, pp. 325-49.

Long life to marketing research 409

EJM 39,3/4

410

Ha˚kansson, H. (1982), International Marketing and Purchasing of Industrial Goods, John Wiley & Sons, London. Ha˚kansson, H. and O¨stberg, C. (1975), “Industrial marketing: an organizational problem?”, Industrial Marketing Management, Vol. 4, pp. 113-23. Halle´n, L. and Sandstro¨m, M. (1991), “Relationship atmosphere in international business”, in Paliwoda, S.J. (Ed.), New Perspectives on International Marketing, Routledge, London and New York, NY, pp. 108-25. Hirschman, E.C. and Holbrook, M.B. (1982), “Hedonic consumption: emerging concept, methods and propositions”, Journal of Marketing, Vol. 46, Summer, pp. 92-101. Holbrook, M.B. (1993), “Postmodernism and social theory”, Journal of Macromarketing, Vol. 13, Fall, pp. 69-75. Holbrook, M.B. and Hirschman, E.C. (1982), “The experiential aspects of consumption: consumer fantasies, feelings, and fun”, Journal of Consumer Research, Vol. 9 No. 2, September, pp. 132-40. Holloway, R.J., Mittelstaedt, R.A. and Venkatesan, M. (1971), Consumer Behavior. Contemporary Research in Action, Houghton Mifflin, Boston, MA. Hooker, C.A. (2000), “The unity of science”, in Newton-Smith, W.H. (Ed.), A Companion to the Philosophy of Science, Blackwell, Malden, MA, pp. 540-9. Howard, J.A. and Sheth, J.N. (1969), The Theory of Consumer Behavior, John Wiley & Sons, New York, NY. Hunt, D.J. (1976), “The nature and the scope of marketing”, Journal of Marketing, Vol. 40, July, pp. 17-28. Hunt, D.J. (2002), “Marketing as a profession: on closing stakeholder gaps”, European Journal of Marketing, Vol. 36 No. 3, pp. 305-12. Jackson, B.B. (1985), “Build customer relationships that last”, Harvard Business Review, Vol. 63, November-December, pp. 120-8. Jackson, F.L. (1996), “Postmodernism and the recovery of the philosophical tradition”, Animus. A Philosophical Journal of Our Time, Vol. 1, 23 December. Kolb, D.A. (1984), Experiential Learning. Experience as the Source of Learning and Development, Prentice-Hall, Englewood Cliffs, NJ. Levy, S. (2002), “Revisiting the marketing domain”, European Journal of Marketing, Vol. 36 No. 3, pp. 299-304. Lynch, M.P. (2001), The Nature of Truth. Classic and Contemporary Perspectives, The MIT Press, Cambridge, MA. Merriam, S.B. and Caffarella, R.S. (1991), Learning in Adulthood. A Comprehensive Guide, Jossey-Bass, San Francisco, CA. Morgan, R.M. and Hunt, S.D. (1994), “The commitment-trust theory of relationship marketing”, Journal of Marketing, Vol. 58, July, pp. 20-38. Nicosia, F.M. (1966), Consumer Decision Processes. Marketing and Advertising Implications, Prentice-Hall, Englewood Cliffs, NY. Normann, R. (1985), Service Management: Strategy and Leadership in Service Business, John Wiley & Sons, Chichester. Oliver, R.L. (1996), Satisfaction: A Behavioral Perspective on the Consumer, Irwin/McGraw-Hill, Boston, MA.

Petitot, J. (1993), “Attualita` scientifica ed etica del razionalismo critico”, in Mangione, C. (Ed.), Pensiero Scientifico e Pensiero Filosofico. Conflitto, Alleanza o Reciproco Sospetto?, Franco Muzzio Editore, Padova, pp. 65-83. Piercy, N.F. (2000), “Why it is fundamentally stupid for a business school to try to improve its research assessment score”, European Journal of Marketing, Vol. 34 No. 1/2, pp. 27-35. Piercy, N.F. (2002), “Research in marketing: teasing with trivia or risking relevance?”, European Journal of Marketing, Vol. 36 No. 3, pp. 350-63. Ray, C. (2000), “Logical positivism”, in Newton-Smith, W.H. (Ed.), A Companion to the Philosophy of Science, Blackwell, Malden, MA, pp. 243-51. Reichheld, F.F. (1996), The Loyalty Effect, Harvard Business School Press, Boston, MA. Rosenberg, A. (2000), “Social science, philosophy of”, in Newton-Smith, W.H. (Ed.), A Companion to the Philosophy of Science, Blackwell, Malden, MA, pp. 451-60. Salmon, W.C. (2000), “Logical empiricism”, in Newton-Smith, W.H. (Ed.), A Companion to the Philosophy of Science, Blackwell, Malden, MA, pp. 233-42. Schmitt, B. (1999), Experiential Marketing. How to Get Customers to Sense, Feel, Think, Act and Relate to Your Company Brands, The Free Press, New York, NY. Schmitt, F. (1999), “Social epistemology”, in Greco, J. and Sosa, E. (Eds), The Blackwell Guide to Epistemology, Blackwell, Malden, MA, pp. 354-82. Shapere, D. (2000), “Scientific change”, in Newton-Smith, W.H. (Ed.), A Companion to the Philosophy of Science, Blackwell, Malden, MA, pp. 413-22. Sherry, J.F. Jr (1991), “Postmodern alternatives: the interpretive turn in consumer research”, in Robertson, T.D. and Kassarjian, H.H. (Eds), Handbook of Consumer Behavior, Prentice-Hall, Englewood Cliffs, NJ, pp. 548-91. Smith, J.B. and Barclay, D.W. (1997), “The effects of organizational differences and trust on the effectiveness of selling partner relationships”, Journal of Marketing, Vol. 61, January, pp. 3-21. Smithee, A. (1997), “Kotler is dead!”, European Journal of Marketing, Vol. 31 No. 3/4, pp. 317-25. Thompson, C.J. (1993), “Modern truth and postmodern incredulity: a hermeneutic deconstruction of the metanarrative of ‘scientific thruth’ in marketing research”, International Journal of Research in Marketing, Vol. 10 No. 3, pp. 325-38. Toraldo di Francia, G. (1993), “Errori e miti nel concetto comune di scienza”, in Mangione, C. (Ed.), Pensiero Scientifico e Pensiero Filosofico. Conflitto, Alleanza o Reciproco Sospetto?, Franco Muzzio Editore, Padova, pp. 135-62. Turnbull, P.W. and Valla, J.P. (Eds) (1985), Strategies for International Industrial Marketing, Croom Helm, London. Uusitalo, L. (1998), “Consumption on postmodernity. Social structuration and the construction of the self”, in Bianchi, M. (Ed.), The Active Consumer. Novelty and Surprise in Consumer Choice, Routledge, London, pp. 215-35. Vattimo, G. (1991), La Fine della Modernita`, Garzanti, Milan. Vavra, T.G. (1995), Aftermarketing. How to Keep Customers for Life through Relationship Marketing, Irwin Professional, Chicago, IL. Walter, G.A. and Marks, S.E. (1981), Experiential Learning and Change. Theory Design and Practice, John Wiley, New York, NY. Walters, C.G. and Paul, G.W. (1970), Consumer Behavior. An Integrated Framework, Richard D. Irwin, Homewood, IL.

Long life to marketing research 411

EJM 39,3/4

412

Wensley, R. (2002), “A bridge over troubled water?”, European Journal of Marketing, Vol. 36 No. 3, pp. 391-400. Williams, J. (1998), Lyotard. Towards a Postmodern Philosophy, Polity Press, Cambridge. Zagzebski, L. (1999), “What is knowledge?”, in Greco, J. and Sosa, E. (Eds), The Blackwell Guide to Epistemology, Blackwell, Malden, MA, pp. 92-116. Zaltman, G. and Wallendorf, M. (1979), Consumer Behavior: Basic Findings and Management Implications, John Wiley & Sons, New York, NY. Further reading Bocchi, G. and Ceruti, M. (Eds) (1985), La Sfida della Complessita`, Feltrinelli, Milan. Firat, A.F., Sherry, J.F. Jr and Venkatesh, A. (1994), “Postmodernism, marketing and the consumer”, International Journal of Research in Marketing, Vol. 11 No. 4, pp. 311-16. Venkatesh, A., Sherry, J. Jr and Firat, A.F. (1993), “Editorial. Postmodernism and the marketing imagery”, International Journal of Research in Marketing, Vol. 10 No. 3, pp. 215-23.

About the authors Michela Addis Michela Addis, PhD, is Assistant Professor of Management at the Institute of Management, Bocconi University, Milan. She also teaches at the Marketing Department, SDA Bocconi Business School.

About the authors

413

Andrea Davies Andrea Davies is a Senior Lecturer in Marketing at the University of Leicester. She received her PhD from the Open University. Her research interests focus on the consumer experience of tourism, branding and on methodological issues in quantitative research. Her work has been published by the Association of Consumer Research, Journal of Tourism Management, and the International Journal of Museum Management and Curatorship. James A. Fitchett James A. Fitchett is a Senior Lecturer in Marketing at Nottingham University Business Schoool, UK. His research interests focus on critical cultural readings of consumption-related phenomena, and has looked at issues such as the environmental implications of consumption, the characteristics of buyer behaviour in illicit markets, and the consumer experience of tourism. His work has been published in the Journal of Macromarketing, Marketing Theory, and Consumption, Markets and Culture. Christina Goulding Christina Goulding is Professor of Consumer Research at the Business School, Wolverhampton University. Her research interests include subcultures of consumption, the body, identity and consumption and qualitative research methodologies. She has published her work in a number of journals including Psychology and Marketing, European Journal of Marketing, Consumption, Markets and Culture, Journal of Marketing Management, Qualitative Marketing Research and Advances in Consumer Research. Evert Gummesson Evert Gummesson is Professor of Management and Marketing at the Stockholm University School of Business, Sweden. After graduation he worked for 25 years as marketing manager and management consultant. His research is directed to relationship marketing, customer relationship management, and quality management with a particular emphasis on services. In 1977 he published the first book on services marketing in Scandinavia; in 2000 he received the American Marketing Association’s Award for Leadership in Services Marketing; and in 2003 he was listed by the Chartered Institute of Marketing as one of the 50 most important contributors to the development of marketing. His book Total Relationship Marketing (revised edition, 2002) has won two prizes for excellence, and his book Qualitative Methods in Management Research (revised 2000) has been reprinted ten times. He has authored more than 20 books and numerous articles in refereed journals, such as European Journal of Marketing, Journal of the Academy of Marketing Science, Journal of Marketing Management, Management Decision, Long Range Planning and Service Industry Management. He is a member of

European Journal of Marketing Vol. 39 No. 3/4, 2005 pp. 413-415 q Emerald Group Publishing Limited 0309-0566

EJM 39,3/4

editorial and review boards of, among others, Journal of the Academy of Marketing Science, Journal of Marketing Management and Journal of Service Research.

414

Robert P. Hamlin Robert Hamlin is a Senior Lecturer in the Faculty of Marketing at the University of Otago. He was trained in Agriculture at the University of Oxford, and held a number of posts in the British Agriculture Industry before taking an MBA at Indiana University as a Harkness Fellow. He obtained his PhD at Otago University, where he specialises in food marketing and point of sales communication. Graham Hooley Graham Hooley is a Professor of Marketing and Senior Pro-Vice Chancellor of Aston University. His research interests include marketing strategy, competitive positioning, marketing in developing economies, and marketing research methods. His work has appeared in journals such as the Journal of Business Research, Journal of International Business Studies, International Journal of Research in Marketing as well as the European Journal of Marketing. He also co-edited the book Quantitative Methods in Marketing, the second edition of which was published in 1999. Nick Lee Nick Lee is a Lecturer in Marketing at Aston Business School. His research interests include sales management, electronic commerce, and marketing research philosophy and methodology. He has presented his work at international conferences around the world, and was the recipient of the 2002 EMAC Conference award for best paper based on doctoral work, as well as a 2002 UK Chartered Institute of Marketing Research Initiative Award. He lectures in marketing research at undergraduate and postgraduate level. Dale Littler Dale Littler is Professor Marketing at the Manchester School of Management, UMIST where he has held a number of senior positions including the Head of the Manchester School of Management and Dean of Management. He established the Customer Research Academy and the Centre for Applied Management Research, which was part of a successful £5.7 million competitive bid to a UK Government initiative. He has been a visiting professor at many major European business and management schools and departments. He has had several grants from the ESRC and was a principal investigator for around eight years on the ESRC-funded Programme of Information and Communication Technologies. He has also had research grants from the European Commission, the Chartered Institute of Management Accountants, the Department of Trade and Industry, the Department of Education and Science, the Teaching Company Secretariat and from the private sector. In 2002, the TCS for which he was responsible was given the ESRC award for the TCS with the best application of Management Science or Social Science. He has published extensively. He is on the editorial board of many marketing and technology management journals. He has undertaken consultancies with a wide range of companies. He has been a member of the Research Grants Board of the ESRC; and is currently on the Academic Senate of the Chartered Institute of Marketing and Chair of the national research committee of the Academy of Marketing.

Stefano Podesta` Stefano Podesta`, full professor, is the Dean of the Institute of Management at Bocconi University, Milan, Italy. He also teaches at the Marketing Department, SDA Bocconi Business School. Ype H. Poortinga Ype H. Poortinga is a former Secretary-General, President, and Honorary Member of the International Association for Cross-cultural Psychology. Currently, he is a part-time Professor of Cross-cultural Psychology at Tilburg University in The Netherlands and at the Catholic University of Leuven in Belgium. His empirical research has been on a range of topics, with consistent attention to theoretical and methodological issues in the interpretation of cross-cultural differences. John Saunders John Saunders is Professor of Marketing, Head of Aston Business School and Pro-Vice Chancellor of Aston University. Besides being Dean of the Senate of the CIM, he has senior responsibilities within the EFMD and AACSB International. In addition, he has served on the research selectivity exercises of the UK, The Netherlands and New Zealand and on accreditation panels for EQUIS, AACSB International and the QAA. His research has covered market modelling and marketing strategy but focuses most recently on managerial incompetence. His articles have appeared in Journal of Marketing Research, Journal of Marketing, Marketing Science, International Journal of Research in Marketing, Journal of International Business Studies, Journal of Business Research, European Journal of Marketing, and many others. Caroline Tynan Caroline Tynan is a Professor of Marketing and Head of the Marketing Division at Nottingham University Business School, Chair of the Academy of Marketing and visiting Professor of Marketing at The University of Ljubljana in Slovenia. Her research interests include relationship marketing, particularly within business-to-consumer and cross-cultural contexts, consumption rituals and services marketing. She has published in a number of journals including the Journal of Business Research, The European Journal of Marketing, The Journal of Marketing Management and The Journal of Strategic Marketing. Hester van Herk Hester van Herk received her PhD from Tilburg University, The Netherlands. Currently she is Assistant Professor of Marketing and Marketing Research at Vrije Universiteit, Amsterdam. Her research interests include methodological issues in culture-comparative research, and cross-cultural factors that influence consumer attitudes and behaviours. Theo M.M. Verhallen Theo M.M. Verhallen received his PhD in Economic Psychology at Tilburg University, The Netherlands. He is Professor in Marketing and Marketing Research and Dean of the Faculty of Economics and Business Administration at Tilburg University. His research interests are in the area of strategic marketing research, the market orientation of firms and international chain and network studies.

About the authors

415

Call for papers European Journal of Marketing Special Issue on

Trust in marketing Guest Editors: David C. Arnott and David Wilson, Warwick Business School Trust is a pivotal concept in both social and business life. It is fundamental to almost every successful relationship (personal or business) and yet research into this most basic of concepts is limited. What research exists emanates mostly from the fields of psychology and organisation theory, leaving a research gap in the role of trust in a marketing context. Five key arguments support a need for this special issue. First, the heavily researched areas of relationship marketing, customer relationship management, branding, etc. support the need to understand, to gain and to maintain long-term relationships which, at least from the customers’ perspectives, are based almost entirely on trust and confidence in the organization. Second, there is a growth in trust-related research activity evident in a number of recent articles in the major marketing journals and conferences. Third, there is a growth of arm’s-length dealings via direct sales and via the internet with growing consumer uncertainty about security issues, about globalised, anonymous, ‘‘faceless’’ companies, and about corporate ethics. Fourth, and as a consequence of point three, various research forums (e.g. World Economic Forum, 2003) adopt Trust as their conference theme and the research presented suggests that the two groups least trusted (and equally mistrusted) are politicians and large companies. The latter of these spend enormous amounts of scarce resource (money and time) on understanding customer relationships but very little on managing the basic elements of trust building – integrity, benevolence and ability. And finally, there is the ease with which reputations built on trust can be destroyed and a consequent need to avoid disasters. Marketing is at the post-introduction/early-growth phase of research into the trust concept, the ideal point for a special issue (especially for a journal at the forefront of marketing thought). This special issue of EJM is aimed at filling some of the theoretical and empirical gaps in research into the trust concept in a marketing context. Consequently, we are seeking original, marketing-focused, conceptual and/or empirical, quantitative and/or qualitative papers on the topic and invite submissions which investigate or explore issues such as: . Conceptualising trust in a marketing context . Dimensions (determinants, antecedents) of trust/distrust in a

marketing context . Modelling of trust/distrust in a marketing context

. Measurement of trust/distrust in a marketing context . Trust in the buyer-seller relationship . Trust in business-to-business marketing relationships . Trust in business-to-consumer marketing relationships . Building trust in online environments . Building trust in entrepreneurial or new businesses . The interaction between trust and branding . The interaction between trust and product involvement . The impact of competitors’ activities on trust-based relationships . The interaction between trust and time and contact frequency . Managing trust-based relationships . Trust recovery/rebuilding after marketing disasters . Measuring the consequences of trust (i.e. cost of losing or

benefits of retaining) . Intra-organisational trust between marketing and other

departments or functions . Marketing as the initiator of trust-based relationships . Role of and/or impact of marketing intermediaries in the

development of trust.

Prior to submission please consult the author guidelines available on the journal home page at www.emeraldinsight.com/ejm.htm or on the inside back cover of any hard copy of the journal. Papers, with a maximum length of 4,000 words, should be sent in standard EJM submission format but to the attention of the guest editors: Dr David C. Arnott, Lecturer in Marketing, Warwick Business School, University of Warwick, Coventry CV4 7AL, UK Prof. David Wilson, Professor in Strategic Management, Warwick Business School, University of Warwick, Coventry CV4 7AL, UK Please address telephone or e-mail queries to David Arnott at: +44 (0)2476 524487 (voicemail); [email protected] The deadline for submission of papers is: 1 December 2005