Advertising Research

ADVERTISING
The Quarterly Review of Marketing Communications
International Journal of
Volume 26 Number 4 2007
Contents
Editorial The academicianpractitioner gap in advertising Gergely Nyilasy and Leonard Reid The web motivation inventory: replication, extension and application to internet advertising Shelly Rodgers, Ye Wang, Ruth Rettie and Frank Alpert More than meets the eye: investigating the hidden impact of brand placements in television magazines Jrg Matthes, Christian Schemer and Werner Wirth The influence of message format on the effectiveness of print advertisements for tourism destinations Alain Decrop Internet advertising effectiveness: the effect of design on click-through rates for banner ads Helen Robinson, Anna Wysocka and Chris Hand Directions in marketing communications research: an analysis of the International Journal of Advertising Douglas West Comments International advertising issues and challenges From Charles R. Taylor and Barbara Mueller Book reviews W. Timothy Coombs and Sherry J. Holladay Its not just PR: public relations in society Andrew Purcell Paul Gillin The New Influencers: A Marketers Guide to the New Social Media Susan Dobscha Ex Libris Malcolm White Global economy and adspend prospects
The Advertising Association is not responsible for the opinions and data contained in the editorial or articles in this journal Advertising Association 2007
419 425 447
477
505
527
543
557 565
571
EDITORIAL ADVISORY BOARD

T Ambler, Senior Fellow London Business School Professor M J Baker University of Strathclyde Dr L Bergkvist Professor P Berthon Bentley College A L Biel Alexander L Biel & Associates Professor S Brown University of Ulster Professor S Burgess University of Cape Town, South Africa Professor L Carlson Clemson University Professor A Caruana University of Malta Professor C S Craig New York University Professor G Davies Manchester Business School Professor L de Chernatony University of Birmingham Professor A S C Ehrenberg South Bank University Professor R Elliott Warwick Business School Professor M Ewing Monash University Professor P W Farris University of Virginia Professor J Ford Old Dominion University A T Green Ipsos MORI Professor L Ha Bowling Green State University Professor C Hackley Royal Holloway University of London Professor F Hansen Copenhagen Business School Professor H Henry IMCA Professor M Holbrook Columbia University Professor J P Jones RGC Consulting Corporation Professor M A Kamins University of Southern California Professor K Karande Old Dominion University Professor S Kates Simon Fraser University Professor E Kaynak Pennsylvania State University at Harrisburg Professor P Kitchen Hull Business School Professor S Koslow University of Waikato Professor A Kover Fordham and Yale Universities Professor D Krugman University of Georgia Professor J J Lambin University Catholique de Louvain J Lannon Market Leader Dr H Li Michigan State University C McDonald McDonald Research Professor P Michell Leeds University Business School Dr A Nairn E M Lyon Professor P Neijens University of Amsterdam Dr S ODonohoe University of Edinburgh Professor S Okazaki Autonomous University of Madrid Professor D Olsen University of Alberta Professor S Paliwoda University of Strathclyde Professor C Patti University of Denver Professor N Piercy Warwick Business School A E Pitcher, CBE International Advertising Association Professor L Pitt Simon Fraser University Professor G Prendergast Hong Kong Baptist University Professor L Reid University of Georgia Professor J Rossiter University of Wollongong Professor C Rungie University of South Australia Professor S Sasser Eastern Michigan University Professor D Schultz Northwestern University J Scriven South Bank University Professor B Stern Rutgers University Professor D Stewart University of Southern California Professor M Sutherland Bond University Professor R Taylor Villanova University A C Tempest Director General, FEDMA, Brussels Professor R Terlutter University of Klagenfurt M J Waterson The Advertising Association R White Admap Dr B Young University of Exeter Professor G Zinkhan University of Georgia
418
Editorial
Douglas West
University of Birmingham
We have some thought-provoking comments to accompany the main articles in this last issue of IJA of 2007, with Overcoming barriers to publishing international advertising research in top journals from Charles R. Taylor (Villanova University) and Just where does corporate responsibility end and consumer responsibility begin? The case of marketing food to kids around the globe from Barbara Mueller (San Diego State University). Along with the book reviews is a fascinating Ex Libris from our invited practitioner, Malcolm White, and our regular global adspend trends from Colin Macleod at WARC. There are a couple of web-related developments worthy of note for this end-of-the-year report. One is that the journal launched its new website in spring www.internationaljournalof advertising.com. The site features full text of the previous 12 months issues, all IJA article abstracts since 1990, as well as abstracts from the entire warc.com database. Full text of leading IJA papers on key industry debates of the past ten years are also accessible, as are book reviews and comment pieces. Subscribers to the journal (here comes a small plug) also benefit from access to global adspend data by country, and
full text of the adspend trends article published in each issue. The other development is that we will soon be changing our submissions process to an online paper-tracking system. Authors will be able to submit their papers over the internet and check on their progress, and reviewers will be able to download papers and complete reviews through the system. It will bring greater clarity and visibility to the review process for authors, and will make online reviews quicker and easier to complete. Turning to this years submission statistics, there has been an increase in submissions on 2006 figures of just over 10% to 102 submissions. It is notoriously difficult to be precise about the acceptance rate in any calendar year given the overhang of papers under review from year to year, but our estimate is an 18% acceptance rate for 2007. Turning to author domicile, the United States has increased its share of submissions from 27% last year to 31%; submissions from Europe fell slightly over the same period, from 24% to 17%, and those from the UK fell significantly, from 25% to 9%. I suspect the fall in submissions from the UK may be linked to the UKs research assessment exercise for 2008
International Journal of Advertising, 26(4), pp. 419420 2007 Advertising Association Published by the World Advertising Research Center, www.warc.com
419
INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)
(see http://www.rae.ac.uk/), which sees this year as the closing period. However, there was an enormous increase in submissions from the rest of the world, from 23% to 43% of the total. A new statistic to pass on is that 95% of submissions were from academics and 5% made up of MARCOMS practitioners. Regarding topics, advertising practice (32%) and advertising effects (31%) continued to dominate the submissions, followed by advertising content (19%). Compared to last year, submissions on cultural topics declined sharply, from 23% to 14%, as did public policy, which fell from 18% to 4%. Perhaps the shift in topic areas reflects the changes in the domicile of submissions. Thanks to the members of our Editorial Advisory Board for their continued support of the IJA and their valuable advice to people submitting papers. A particular thanks to John Ford (Old Dominion) and Stephanie ODonohoe (Edinburgh) for their work and development of the Comments and Book review sections respectively. We will be announcing our two top reviewers of 2007 in the next issue, which is always an extremely difficult decision, made jointly by the Editor and production team.
Lastly, I would like to thank (in alphabetical order) the following ad hoc reviewers: Jennifer Argo, University of Alberta George Balabanis, Cass Business School, City University Bradley Barnes, University of Kent Helen Borland, University of Birmingham Louise Canning, University of Birmingham Marylyn Carrigan, University of Birmingham George Christodoulides, University of Birmingham Geoffrey Crouch, La Trobe University Janine Dermody, University of Gloucestershire Lynne Eagle, University of Middlesex Jaafar El Murad, University of Westminster Michael La Tour, University of Nevada, Las Vegas Larry Percy, Copenhagen Business School Ian Phau, University of Curtin Mike Read, Monash University Helen Robinson, Kingston University Herbert Rotfeld, Auburn University Isabelle Szmigin, University of Birmingham
420
ABSTRACTS
The academicianpractitioner gap in advertising

Gergely Nyilasy and Leonard Reid
The existence of the academicianpractitioner gap is readily acknowledged and widely discussed in the marketing/advertising literature. This paper analyses key writings on the nature of the academicianpractitioner rift and proposes a new approach complementing the literature. The review identifies five prevailing explanations why miscommunication between academicians and practitioners exists: (1) the failure of academic knowledge dissemination systems; (2) problems with the knowledge content and knowledge form academicians produce; (3) counterproductive academic organisational systems; (4) questions of philosophy of science; and (5) practitioners inability and unwillingness to process academic information. The study concludes that one potential explanation is entirely missed in these accounts: the possibility that practitioners knowledge about how advertising works is an autonomous construct, which has its own rules and deep structure, and resists simple assimilation attempted by academicians. The study also complements the existing literature by basing the review on firm theoretical grounds: the authors apply the influential sociological theory of professionalisation. Finally, future directions for research investigation are suggested, which moves the predominantly normative discourse into the empirical world.
425
The web motivation inventory: replication, extension and application to internet advertising
Shelly Rodgers, Ye Wang, Ruth Rettie and Frank Alpert
The Web Motivation Inventory (WMI) is used in internet advertising research, and is frequently used and cited in advertising, marketing and communication literature. Investigations of the robustness of the WMI have been somewhat limited. Additionally, new uses of the internet are not accounted for by the WMI since its publication in 2002. This paper replicates and extends the original WMI using participants in the US, UK and Australia and includes internet motives not previously examined. The results show that the four-factor WMI remained reliable and valid for all three samples. Findings suggest the extended WMI may be broken into 12 sub-scales that represent the original four-factor measure.
447
More than meets the eye: investigating the hidden impact of brand placements in television magazines
Jrg Matthes, Christian Schemer and Werner Wirth
A growing research literature suggests that visual brand placements need not be recalled in order to have an impact on brand attitudes. In an experimental study working with an authentic television broadcast, the authors investigated the
477
421
influence of frequency of placement exposure, viewers involvement and viewers persuasion knowledge on the attitudes towards the brand and brand recall. Results clearly indicate a mere exposure effect. A frequently presented brand placement can have a positive effect on brand evaluations although viewers do not recall the brand. However, this effect can only be found when there is a high involvement in the programme and low persuasion knowledge. In contrast, when persuasion knowledge is high and involvement is low, frequently presented placements lead to a deterioration of brand attitudes.
The influence of message format on the effectiveness of print advertisements for tourism destinations
Alain Decrop
In a context of overabundant ads and saturated consumers, message format is crucial in developing effective advertising campaigns. In this study, four major format components of print advertisements are considered: picture, logo, text and headline. The goal is to investigate the effectiveness of each of these components in triggering a response by the target audience. Three types of response are taken into account: knowledge (information), liking (attraction) and behavioural intention. Four series of hypotheses related to the influence of message format on the effectiveness of print ads are tested for an urban tourism destination through an experimental research design. Findings show that picture and text are the prevailing elements, while logo and headline are of marginal importance. Pictures are especially effective in attracting the consumer and arousing a behavioural intention, whereas text is most powerful in conveying information.
505
Internet advertising effectiveness: the effect of design on click-through rates for banner ads
Helen Robinson, Anna Wysocka and Chris Hand
Online advertising has experienced phenomenal growth since its inception in 1994. This empirical study investigates the impact of seven creative characteristics of banner ads on the effectiveness of online advertising using a multiple regression model. A random sample of 209 banner ads was drawn from a sampling frame of advertisers, provided by an advertising agency specialising in internet advertising for the gaming industry. The findings of this study are broadly consistent with past research into online advertising efficiency, indicating that the creative characteristics of effective banner ads in the online gaming arena include: a larger size, absence of promotional incentives and the presence of information about casino games. In contrast, banner features such as animation, action phrase and presence of company brand or logo were ineffective in generating click-throughs. Contrary to expectations, long messages on banners were associated with higher click-through rates.
527
422
ABSTRACTS
Directions in marketing communications research: an analysis of the International Journal of Advertising

Douglas West
This paper provides a content analysis of the inputs and outputs of the International Journal of Advertising (IJA) during the period 19922006 and was sparked by the 25th anniversary of the journal (19822006). A total of 348 papers were surveyed using a content analysis in order to provide researchers and readers with a better sense of the contribution of the IJA over the past 15 years. The analysis reveals a journal largely focused upon topics involving practice and effects with increasingly sophisticated statistical techniques employed. Single-authored papers appear to be in decline in favour of two- to three-authored papers and author institutions widely domiciled across North America, Asia, the UK, Europe and Australasia.
543
423
The academicianpractitioner gap in advertising

Gergely Nyilasy
Hall and Partners, New York
Leonard N. Reid
Grady College, University of Georgia
The existence of the academicianpractitioner gap is readily acknowledged and widely discussed in the marketing/advertising literature. This paper analyses key writings on the nature of the academicianpractitioner rift and proposes a new approach complementing the literature. The review identifies five prevailing explanations why miscommunication between academicians and practitioners exists: (1) the failure of academic knowledge dissemination systems; (2) problems with the knowledge content and knowledge form academicians produce; (3) counterproductive academic organisational systems; (4) questions of philosophy of science; and (5) practitioners inability and unwillingness to process academic information. The study concludes that one potential explanation is entirely missed in these accounts: the possibility that practitioners knowledge about how advertising works is an autonomous construct, which has its own rules and deep structure, and resists simple assimilation attempted by academicians. The study also complements the existing literature by basing the review on firm theoretical grounds: the authors apply the influential sociological theory of professionalisation. Finally, future directions for research investigation are suggested, which moves the predominantly normative discourse into the empirical world.
Introduction
Advertising academicians and advertising practitioners seem to live in different worlds. The separation between these two groups, commonly referred to as the academicianpractitioner gap (Hunt 2002a), would not be too alarming if it only denoted the fact that there are always discrepancies between theoretical modelling in the field and practical applications. The gap in the case of advertising, however, is much wider and is manifested on deeper levels than would be expected in the case of occupations such as medicine, engineering or law.
425
The importance of the subject is reflected in a recent action of the American Academy of Advertising. At the 2006 AAA Conference, a special topic session focused specifically on the problem of the academicianpractitioner gap; the central mission of the session was to find ways to create more connections between advertising practitioners and those in academia (Katz 2007, p. 1). The outcome was the creation of an AAA Newsletter series on applied research summaries to build and strengthen the relationship between the academic producers and the applied users of advertising knowledge. This paper offers another means of continuing the dialogue and promotes the discussion of the gap problem between the academic and professional communities of advertising. The objective of this paper is twofold: (1) to give a structural review of the academicianpractitioner divide literature in advertising and (2) to propose a new approach that moves the discussion forward. Specifically, the paper analyses key writings from the marketing and advertising literature that discuss the nature of the academicianpractitioner rift, why it exists and if it is possible to narrow it. The paper also suggests that the reviewed literature suffers from three shortcomings: (1) the explanations offered are incomplete; (2) they lack a clear theoretical framework they tend to be ad hoc assessments; and (3) they have little empirical support. In contrast, the authors recommend a new approach that addresses each of these points: (1) the approach offers a new explanation that complements the existing literature; (2) it utilises a firm theoretical base; and (3) it maps out directions for future empirical investigations. The review has some inherent limitations. Although the authors have made a significant effort to search and compile an exhaustive list of articles for reviewing the subject (through searching databases such as EBSCO Business Source Premier, following reference links, consulting colleagues, etc.), some sources might have been missed. Further, the review claims about the nature of academicianpractitioner relations are limited geographically: they are assumed to be valid only in the US and the UK the origins of the referenced sources. It is quite possible that some dynamics of the above relations have strong local flavours in other countries.
426
THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING
The first part of the paper surveys the existing literature. Next it introduces a new approach, which places the debate in a relevant theoretical context and maps out future directions for empirical inquiries.
The review
The existence of the academicianpractitioner gap is readily acknowledged and widely discussed in marketing, one of advertisings umbrella disciplines. As Hunt (2002a, p. 305) argues: Throughout its 100-plus year history, one of the most recurring themes has been that there is a gap or divide between marketing academe and marketing practice. Indeed, over the years, a sizeable volume of literature has developed focusing on this issue (e.g. Ehrenberg 1969; Kover 1976; Maiken et al. 1979; Peters 1980; Parasuraman 1982; Holbrook 1985, 1987; Preston 1985; AMA Task Force 1988; Wells 1993; McQuarrie 1998; Rossiter 2001; Hunt 2002b; McKenzie et al. 2002; Baker & Holt 2004; Baldridge et al. 2004; Brennan & Ankers 2004; November 2004; Tapp 2004a, 2004b; Gabriel et al. 2006; Southgate 2006). Most scholars define the gap as a communication problem between academia and advertising/marketing practice. According to this view, while academicians continually add to a body of abstract knowledge about advertising and marketing phenomena, practitioners do not seem connected to this information. Practitioners do not read journals, and they do not even consider academic knowledge very relevant (Hunt 2002a). While few commentators question the existence of the gap, they offer radically different explanations, and consequently suggest different potential solutions. Most commonly, the cause of the problem is attributed to academicians themselves, and the academicianpractitioner rift linked to five problems: (1) information dissemination; (2) knowledge content and form; (3) academic organisational structures; (4) philosophy of science; and (5) practitioner users. Knowledge dissemination To many, the divide is a dissemination problem and exists because academicians are not successful in disseminating the knowledge they generate. According to the AMA Task Force on the Development of Marketing
427
Table 1: Understandings of the marketing academicianpractitioner gap based on a review of the academic literature
Definition: miscommunication between marketing academia and practice
Who to blame Causes Knowledge dissemination Academicians Failure to create adequate distribution systems for academic advertising knowledge Too much focus on knowledge production that has no relevance or presented in an incomprehensible format for practice Certain organisational characteristics of academia prohibit a practical focus Philosophical presuppositions about the nature of academic advertising knowledge that prohibit a practical focus Unwillingness and inability of practitioners to process advertising research results Practitioners
Knowledge content/form
Academic organisational structures
Philosophy of science
Practitioner knowledge utilisation
Thought, academic researchers do not communicate well with their constituencies, most importantly with practitioners (AMA Task Force 1988, p. 4). The Task Force specifically brought together for assessing and potentially reducing the gap described the classic ideal of a direct communication flow from academia to praxis: Primarily, marketing knowledge is developed somewhat formally by academic researchers and commercial marketing researchers or consultants and more experientially by practitioners or users. Knowledge developed by academic researchers tends to be disseminated to the discipline through research journals or academically oriented conferences (AMA Task Force 1988, p. 17). It is the breakdown in academias knowledge distribution system that causes the divide. The assumption is that if only these systems improved, then the academicianpractitioner problem would cease to exist. Similarly, Baker and Holt (2004) fault marketing education for the fact that despite long decades of research, marketing is perceived to be the least accountable function in business: One of the key findings is that marketers are perceived to be unaccountable by the rest of the
428
organisation; they are seen as unable to demonstrate a return on investment in the activities they have control over (p. 560). Baker and Holt argue that this is especially embarrassing as there are useful tools out there that simply do not get to the practitioners. Some commentators, however, highlight that it is unrealistic to assume a direct flow from academe to praxis, and it is through the facilitation of indirect flows that the gap can be narrowed. Brennan (2004), for example, argues: There is evidence that marketing practitioners do not read academic marketing journals. Perhaps the surprise here is not that practitioners eschew these journals, but that anyone would expect them to read such material at all. Brennan argues that immediate and obvious applications of academic research are neither possible nor desirable. Ehrenberg, almost four decades ago (1969), very similarly stated that it took a considerable amount of time and energy to apply theory to practice; technological application is not an automatic or a direct process in any field, rather it is a long and painful one. The gap in this sense is to some extent natural, and careful nurturing of indirect channels that can effectively translate academic developments into technological applications (such as textbooks, association work, and consultant and research services) would ultimately resolve the issue. In his rejoinder to the AMA Task Force report, Garda (1988, p. 35) agrees with this assessment: The Task Force report implies that marketing knowledge is solely original research at the concept/theory level. Original research is surely needed in each of the knowledge levels, but also needed are resynthesis, repackaging, and repetition of old knowledge for the new generation of managers. Academic researchers (or others such as consulting or research firms) need to develop this secondary form of knowledge to make academic research palatable for practitioners. McKenzie et al. (2002) also acknowledge the importance of channels for indirect communication flows such as trade journals, textbooks, conferences, training and development courses. Others, however, disagree with the assessment that no direct communication is necessary and express concern over the efficiency of indirect dissemination routes. Tapp (2004b, pp. 497498), for example, suggests: The argument that we dont have to concern ourselves with how our work might be used, on the grounds that there is often a time lag between the development of underlying theory and its use in practice, is wearing
429
increasingly thin. He argues that if academicians do not concern themselves with the lack of direct flow, there will be no flow whatsoever, and academic research will quickly become obsolete. Knowledge content and form It is not only knowledge dissemination that is problematic but also the content of advertising/marketing knowledge itself. November (2004), for example, in his satirical article, enumerates seven reasons why practitioners should continue to ignore academic research: (1) academic research does not contain knowledge that is relevant or actionable for practitioners; (2) academic knowledge is inadequately structured: The reality is that, while we do seem to have an agreed standard as to what a brick is, there is no agreement as to which bricks need to be made first, no foundations, no architect of the final wall, and no idea as to what the wall is expected to do when, if ever, it is built (November 2004, p. 41); (3) academics sometimes make false, misleading claims about the existence of causality where, in reality, it is not warranted; (4) academic research is often reductionistic: While a narrowly focused study is manageable and likely to lead to a definitive result, the results, assuming they have statistical validity, cannot be applied outside the scope of the study. This means that we can never generate any generalisations from a single reductionist study (November 2004, p. 43); (5) measurement in marketing is imprecise: Because our measurement systems lack precision in comparison with those used in classical sciences (November 2004, p. 44); (6) knowledge is too general and therefore does not help; (7) there is little replication in market research. In essence, Novembers caustic satire implies there is no useful knowledge in marketing academia and practitioners should not expect there to be. In a similarly self-critical manner, the AMA Task Force (1988, p. 6) pummels academic researchers for producing research that is not good enough (and not good enough for practitioners). The Task Force suggests that there are no real innovative ideas in academic research, only shortpayoff studies; only knowledge creep and not knowledge spurt. At another point in the report, they formulate this criticism much more strongly; they suggest that there is, in fact, very little knowledge available at all in marketing: Further, there is little generalizable, accumulated
430
marketing knowledge to be disseminated to marketings constituents (AMA Task Force 1988, p. 17). Further criticisms include the suggestion that academic research is very difficult to read and uses a lot of jargon (Brennan 2004; Ottesen & Gronhaug 2004) one study even provides empirical evidence for this claim (Crosier 2004). Another potential reason is that academic researchers are not familiar with the problems practitioners face and therefore are unable to develop research programmes that are useful for this constituency (Easton 2000, cited in Brennan & Ankers 2004). Parasuraman (1982, p. 78) suggests that this detachment is aggravated by the fact that little practitioner input is sought or allowed in academic projects: Lack of managerial involvement or at least some managerial emphasis at the theory development stage can greatly reduce the chances of the theory ever being applied in practice. Finally, Katsikeas et al. (2004, p. 568) argue that the problem may simply be topicality: if academic researchers are able to identify the relevant hot topics for research, academia automatically ceases to be irrelevant: Emphasis is placed on identifying a number of hot topics worthy of future investigation. It is hoped that the identification and discussion of these topics will spark greater research on fundamental marketing issues, and that the allied explication of research rigour will likewise enhance the efficacy of research in marketing. Academic organisational structures Others point to the organisational context of the academic world and argue that the gap between practitioners and academicians can partly be explained by the fact that academic incentive and reward systems are not conducive to research that is of direct use for practitioners. One of the key findings of the AMA Task Force study was that the understanding of the sociological context of academia has predictive power when it comes to explaining the gap. The Task Force (1988, p. 6) concluded: The [incentive and reward] system, however, does truly deserve its appellation publish or perish. It produces some very strong and undesirable incentives toward knowledge development on the part of young academicians: it is extremely short-term in orientation, is almost entirely peer-reviewed, and
431
is strongly directed toward achieving a maximum number of publications as a means to the end of promotion. Similarly, November (2004, p. 41) suggests that the sociological context, in which academic knowledge is conceived, has a tremendous impact on the knowledge produced: The relevance of this published material to practitioners has nothing to do with your promotional prospects or its chance of being published. At most universities, the critical factor is the number of publications and the type of journal in which they are published not their relevance. The absence of relevance can readily be seen in the published products. Brennan and Ankers (2004, p. 511) also claim that it is the organisational structure of research at universities, and not individual researchers, that is at fault: It seems clear that although academics would like to get closer to practitioners, they are inhibited by institutional factors, such as academic reward systems and the publish or perish culture. Philosophy of science Some commentators dig deeper and examine the fundamental philosophical presuppositions of academic research. The discussion has crystallised around the dichotomy of whether academic advertising/marketing research is a basic vs applied or academic vs professional discipline. Those who claim the field is an academic discipline argue that scholars are under no obligation to produce knowledge that is directly relevant for practitioners. For this group, the gap between practitioners and academicians is ultimately not a very serious issue. The other group, on the contrary, suggests that since advertising/marketing is by definition an applied field, if academicians are not relevant for practitioners, they are not producing any useful knowledge. This latter approach can be summed up by the proverbial question Webster (1988, p. 49) posed: What kind of knowledge in marketing is there that is not relevant for practice? Although the debate was very heated in the 1980s, it seems the professional discipline side has won the dispute. Hunts (2002b) monumental summary of marketing theory concludes that problem-oriented research is what marketing academia should strive for. In his assessment, problemoriented research represents a healthy equilibrium between purely basic research orientation and purely pragmatic applied research. Problem-
432
oriented research is applied research, which produces theoretically based generalisations that also have practical value. Similarly, Myers argued in a famous roundtable discussion on the issue: Marketing academicians should recognize that the overall purpose of research and knowledge development in this field, over the short-run or the long-run, is to improve marketing practice, and decision-making, and in general, to advance the state of knowledge useful for the profession (cited in Maiken et al. 1979, p. 62). There is still a group of thinkers that insist, quite forcefully, that advertising/marketing academia should have nothing to do with practice. Kassarjian suggests, for example: I see no reason why just because we are in marketing we want to force a kind of practitioner view, or a real-world view into other places. I see absolutely no reason why I should have a value system imposed on me that says do something useful. I dont want to do anything useful; I want to do what I want to do. Those of us in academia want to push our value system onto the other side of the world and the other side of the world is trying to push their value system on us and maybe we just ought to part company (cited in Maiken et al. 1979, p. 71). Holbrook (1985, p. 145; 1987), another famous advocate of the academic discipline viewpoint, argues: I believe that business does to consumer research approximately what the comedian Gallaghers Sledge-O-Matic does to watermelon. It smashes, crushes, and pulverizes. If you want to sit in the front row at a Gallagher concert, you had better wear a raincoat. While the above described standpoint seems to be in the minority in academia, it represents a very powerful dynamic: the need for autonomy in advertising/marketing (or any type of academic) research. Those criticising the academic discipline orientation blame this on what they believe to be a misconstrued notion of scholarly autonomy and suggest that this deep-seated belief is responsible for academias inability to produce useful knowledge for the advertising practitioner. Preston (1985, p. 14) suggests, for example, that academic freedom may be better understood within the boundaries of the mission of the academic unit under which the scholar is operating. If freedom is understood this way, marketing/advertising academicians may not be free to be irrelevant: There is, however, another relevant academic concept that of the stated mission of an academic unit. Many university departments have written
433
mission statements, and for advertising or marketing units they undoubtedly make reference to the study of problems faced by practitioners. If the faculty is studying the consumer in such ways as to be not explicitly studying practitioners problems, then perhaps the mission is not being carried out. Yet another way of characterising the above dichotomy between academic and professional discipline orientation is the rigour vs relevance debate: Marketing, perhaps more than any other management functions, has had to struggle with a presumed conflict between rigor and relevance (Webster 1988, p. 48). According to this formulation of the dichotomy, knowledge cannot be both rigorous and relevant in advertising/marketing research; you have to pick sides. Another way of stating this conflict is that researchers have to find a compromise between reliability and external validity. According to the critics of the academicianpractitioner divide, however, the dichotomy is false and should not be an excuse for the production of professionally irrelevant research. Some suggest that perhaps the focus has been too much on reliability and not enough on external validity in academic research (McQuarrie 1998). The consensus seems to be that responsible advertising/marketing scholarship should strive for both rigour and relevance: the rigor and relevance dichotomy is not only false but counterproductive and misleading. Research quality suffers if the only concern is analytical rigor because marketing problems are so easily misspecified, leading to results that are neither valid nor credible. However, managers and practitioners certainly do not come to the academic world of marketing thought looking only for relevance. Nothing is more useful to a professional marketing manager than a good theory that can help to bring order out of chaos, insights out of data, meaning out of patterns (Webster 1988, p. 61). A similar assessment is provided by Katsikeas et al. (2004, p. 568): Enhancing the relevance and rigour of our research in order to arrive at better explanations of contemporary and prospective marketing problems and issues is central to the continued development of the discipline. Even if, theoretically, the professional discipline view represents the majority opinion in academic practice, the actual research often ends up less than useful for practitioners: It seems to be the case that the type of research output that is viewed by academics as being of the highest
434
quality, is the type of research that is viewed by managers as being of the least interest (Brennan 2004). Some critics suggest that even stating the dichotomy of basic vs applied is dangerous, because it allows for work that will never be useful for anyone. The danger to suggest that there is a basic orientation is that the academicianpractitioner gap becomes trivialised and ignored. As Parasuraman (1982, p. 78) argues: Surely, marketing theory building as an end in itself is not worth pursuing. Nevertheless, there is a very real danger of this happening due to the insistence on labeling research projects as either scientist-oriented or technologist-oriented, and further claiming that only scientist-oriented research can contribute to theory construction. Practitioner knowledge utilisation So far we have only discussed arguments placing the blame for the academicianpractitioner divide on academicians. Some reviewers, however, clearly charge practitioners as well for the existence of the gap. First, some critics point to the fact that practitioners often do not use academic information even if it is useful for them. According to this view, even if academia sometimes does have problems with communicating relevant information to practitioners, many times there is relevant information available practitioners simply do not use it. Brennan and Ankers (2004), for example, provide in-depth interview data with academicians pondering on this issue: [The objective is to] provide leading edge knowledge to society but if that society chooses not to use it I dont think it is our job to beat up on them and say youre idiots. You can put the water in the trough and bring the horse to the trough, but if they dont want to drink then thats not an academics problem. McKenzie et al. (2002) report survey data suggesting that, out of their sample population of practitioners (n = 47), not a single marketing manager reads academic journals: It is clear from this survey that academic journals devoted to marketing are largely unknown and unread by marketing managers (p. 12). Related to the previous commentary, the gap may also be explained by negative attitudes held by practitioners (irrespective of whether or not there is any justification for it). The AMA Task Force (1988, p 8), for instance, points to the possibility that the whole issue of irrelevance may
435
be more of a practitioner perception than reality: The work of a marketing academician may be dismissed as ivory tower and having little relevance to the real world when, in fact, some marketing academicians do focus on translating theory into practice. Holbrook (1985, 1987) suggests that there is a generalised negative attitude among practitioners (business people) against academia: anti-intellectualism. In this view, the problem does not have to do with opinions about advertising research in particular. Rather, the problem has to do with a general negative opinion among American business people about the utility and value of academia. Holbrook argues as cited above that this is the main reason academia should not be concerned about practical relevance: business anti-intellectualism can only ruin academic marketing research. Finally, some critics focus on individual cognitive capabilities rather than structural features. Ottesen and Gronhaug (2004) argue that part of the problem may be that professionals simply lack the necessary knowledge to be able to comprehend complex presentations of academic data. They might also have a limited attention capacity to process academic information. They suggest: Also, the research information may not be understood, because the potential users lack the required knowledge (Ottesen & Gronhaug 2004 1990, p. 521). It is also possible that the relevant users are unaware of the information, because potential users like other human beings have limited attention capacity (Cohen & Levinthal 1990, p. 521). Myers, similarly, argues that part of the explanation for the gap lies in managers lack of receptivity to academic information (cited in Maiken et al. 1979, p. 64).
Summary and new directions

To summarise, the growing literature on the academicianpractitioner gap has expressed serious concern about the status quo. Commentators emphasise that the current situation is unfortunate and detrimental to the future interests of both advertising/marketing academia and practice. Most define the problem as a communication issue and attribute the cause of the gap to academic research itself: the inability of academia to produce and disseminate relevant research knowledge to practitioners. The consensus is that the discipline is by definition applied and not basic;
436
therefore academicians should conduct problem-oriented research (Hunt 2002b), or research that addresses general problems advertising/marketing practitioners may face. Some commentators also point to the possibility that at least partly practitioners should be blamed for their unwillingness or inability to process information that is practically useful and readily available. The authors contention is that this classic understanding of the academicianpractitioner gap has shortcomings on three fronts: (1) the explanations offered are incomplete; (2) they lack theoretical foundations they tend to be ad hoc assessments; and (3) they do not have much empirical support. Let us consider each of these limitations and offer a new approach that strives to overcome them. A new explanation: practitioner knowledge autonomy What is striking from the reviewed literature is the extent to which the discussion is centred on academia. Most of the literature focuses on academicians themselves and identifies academic knowledge distribution, content/form, organisational and philosophical factors as the root causes of the problem. When practitioners are mentioned, they are conceptualised as empty vessels (and very leaky ones at that) to be filled with academic wisdom. Despite bona fide attempts by academicians, they are deemed unwilling and incapable of accepting a knowledge transfer. This classic view does not allow for a very possible alternative scenario: practitioners may have their own autonomous knowledge forms about how advertising works. It is very conceivable that as a distinct social group, they have developed their own sets of beliefs about advertising, which in both content and form are independent of academic knowledge. Practitioner knowledge may carry either of the following markers of knowledge autonomy: own set of theories own sets of postulated boundary conditions and domains of applicability unique underlying philosophical assumptions (meta-theories) unique forms of validity/reliability testing independent social systems to negotiate, distribute and consume knowledge-systems that ignore or may actively resist academic influence
437
own style of thinking and normative ideas about what is acceptable vs unacceptable to believe in, irrespective of a strict sense of validity. This is not to say that practitioners are entirely isolated from academia. Indeed, the history of advertising research offers great examples of academicianpractitioner interaction. Many authors have actually been on both sides of the fence, both practising and publishing in academic journals literally embodying a knowledge crossover (consider such classic theorists as Ernest Dichter, Claude C. Hopkins, Herbert Krugman, Richard Vaughn; or contemporaries such as Tim Ambler, Paul Feldwick, Robert Heath, John Philip Jones, Erik du Plessis, John R. Rossiter, to name a few). Further, practitioners may in fact use, unadmittedly, some forms of academic knowledge. Even though direct flows seem to be obstructed, in indirect ways some ideas may still percolate through. Media planners, for instance, may use the mental models of wear-in/wear-out theory when working with media response models even if they would not be able to reflect on the fact that they have borrowed anything from academia. It seems that intermediaries (consultancies, research services and products, journalists, general educators) have a very instrumental role in these implicit transfers. Nevertheless, the existence of interactions, dialogues and crossovers (which still represent relatively isolated occurrences, rather than the norm) does not challenge the contention that there may be structural differences between academic and practitioner knowledge forms. Even though they may influence each other, the two forms of knowledge seem to represent different centres of gravity. The existence of a conversation does not imply that the conversing partners have identical norms, underlying assumptions, styles of expression or ways of knowing. In fact, that is very rarely the case. Finally, as an analytical strategy, it makes sense to fully understand the individual components of an intertwined system, before interactions can be described. The classic view clearly ignores the possibility of practitioner knowledge autonomy and implicitly or explicitly presupposes a unidirectional flow from academia to praxis. Some innovative thinkers, however, have already pointed to the possibility of a reverse flow. The most emblematic example is Zaltman et al.s (1982) theory-in-use approach. Others have also emphasised the legitimacy of a praxis-to-academe transfer. Ottesen
438
and Gronhaug (2004, p. 526), for instance, argue: In order to enhance the usefulness of academic marketing knowledge to practitioners, we need to understand what types of information they perceive as useful as well as factors that might impair the transfer of research information from academia to practice. Similarly, Parasuraman (1982, p. 79) suggests: To the extent possible, attempting to incorporate some managerial focus in the process of marketing theory development is a useful goal to strive toward. However, this may be easier said than done. For, it would not only require that the theory builders be aware of the ultimate theory users (i.e. practitioners) perspectives, but also require that the ultimate theory users be appreciative of the potential benefits of developing marketing theories. Rossiter (2001, p. 21) proposes that one potential way of overcoming the divide between academia and practice is back-engineering practitioner knowledge into academia: What is circulating as practitioner marketing knowledge must be codified and translated into the form of strategic principles, and this work will doubtless have to be done by academics. What we argue for in this paper, however, goes beyond the idea of a simple reverse flow. It is possible that, precisely because of the postulated autonomy of practitioner knowledge, certain aspects of this knowledge do not lend themselves to easy back-engineering into academia. It is also possible that fundamental philosophical, metatheoretical and methodological rifts between the two domains (as well as sociological aspects of their production and negotiation) make the reverse flow just as problematic as the forward flow depicted by the classic literature. When thinking about knowledge flows, we can no longer assume, without critically assessing the validity of this assumption, that the only difference between academic and practitioner knowledge is their topical content. It is quite possible that they are autonomous knowledge systems that differ in complex, multifaceted and interacting structural ways. Although at the developmental stage, the authors have collected data that have found some initial empirical support for the above outlined new approach. The authors encourage others to also investigate the hypothesis. Broader theoretical context The classic literature on the academicianpractitioner divide also fails to put its claims into a larger theoretical context. However, there is a theory
439
in the sociology of occupations that can be of great help. The theory of professions focuses precisely on the phenomenon we want to understand: the nature of the link between an occupation and its theoretical core (if it at all exists). Professionalisation explains why occupations need to develop an academic knowledge base, thereby maintaining and elevating their standing in society (see for example MacDonald 1992; Abbott 1998). All occupations, even ones that are not considered professions in a sociology of occupations sense, strive for a professional status. If there seems to be a breakdown in the otherwise universal professionalisation process (discrepancies and detachments between the knowledge base and its application), the analysis needs to focus on the characteristics of the knowledge base and how key constituencies think about it, in order to understand the discrepancy better. In other words, the idea of professionalisation puts the academicianpractitioner gap in a broader-scale theoretical context that may provide better explanations than an enumeration of ad hoc causes based on ad hoc opinions, an approach, characteristic of the above reviewed literature. Among other things, professionalisation sheds light on the fact that what is really at stake is not just a problem of communication disturbances, personally held attitudes or reading habits, but the societal status of an occupation. If the divide between practitioners and academicians is truly rooted in knowledge, and if practitioners and academicians disagree what knowledge is or can be in advertising, this can seriously threaten, if not altogether undermine, the (aspired) professional status of the advertising industry. The need for empirical studies on practitioner knowledge Finally, as suggested above, the current literature discussing the academicianpractitioner discrepancies is largely normative: a collection of ad hoc observations and recommendations without positive support (some exceptions using empirical data are McKenzie et al. (2002) and Brennan & Ankers (2004)). With a Copernican turn, we need to start using our own social scientific methods and observe the gap in an empirical-positive manner. If we truly want to understand what seems to be a key component of the issue, namely the autonomy of practitioner knowledge, we need to launch research projects to investigate. Such empirical studies can give
440
support (or can refute) the new conceptualisation of the academician practitioner gap described above. Sadly, with very few exceptions (Kover 1995; Gabriel et al. 2006) there are currently no empirical studies available to answer the question of whether there are knowledge autonomy-based discrepancies between academicians and practitioners. We do not know if the gap exists partly because of any of the following: 1. ad practitioners believe advertising works differently from what academics claim; 2. if practitioners have the same presuppositions whether this knowledge is even possible; 3. whether it is even relevant for them to have such a theoretical knowledge base when dealing with clients. Such an investigation is long overdue. As mentioned earlier, the authors have recently collected and analysed data to be able to answer these questions, focusing on: 1. practitioner theories; 2. practitioner metatheories; and 3. pseudo-professionalisation techniques. Their approach in this research was qualitative interviewing, using the grounded theory method. Another potential empirical direction may be the analysis of written practitioner texts. Advertising professionals seem to love to write about their trade in the form of business books, trade articles and even crossover papers published in academic publications such as the Journal of Advertising Research, International Journal of Advertising or International Journal of Market Research; such documents can be subjected to systematic analysis. A third potential source is advertising effectiveness case studies collected and published by professional organisations such as the IPA and MRS in the UK. Finally, conference proceedings can be analysed to content/textual analyses. When executing such empirical investigations, it is important to note that practitioner is a heterogeneous construct. At the macro-level argument that has been presented in this paper, the practitioner clearly differs from the academic; on a more micro-level analysis, important differences are expected. It is very reasonable to assume that there will be important
441
variations in the knowledge autonomies of advertising agencies (and even functional units within) as well as marketing personnel or higher management on the client side. Further, intermediaries (research services, consultancies, associations, etc.) represent yet another layer of complexity as these groups often take an intervening position between academia and practice. The differences between these groups should be uncovered by empirical investigations.
Conclusion
For advertising academicians to advance advertising knowledge, while fulfilling their professional responsibilities to the academic and professional communities, investigations from the practitioner perspective are sorely needed in order to accomplish two fundamental objectives: 1. to narrow the divide between academicians and practitioners, which endangers both academic research as a discipline and advertising as a professionalising occupation; communication and exchange which are impossible without the understanding of each others language are necessary preconditions for long-term survival for both parties 2. to aid advertising education by uncovering the types of knowledge advertising practitioners possess, use and expect from novices entering the academy; advertising educators may benefit from this research by using these insights for the development of improved educational programmes ones that do a better job in anticipating the realities of advertising work and the needs of the industry. To the benefit of all, research on what practitioners think about the workings of advertising will allow us to compare and contrast practitioner perspectives with academic ones, thus allowing us to understand the academicianpractitioner gap on an even deeper level. If discrepancies are found, this can be indicative of the knowledge-based nature of the academicianpractitioner gap and relevant implications can be drawn to minimise it. Additionally, if investigated in the social context in which practitioner knowledge is used, research can provide a rich description of a knowledge-based occupations professionalisation dynamics, including advertisings scholarly responsibilities to the academic community where
442
advertising is taught and studied, and to the professional community where advertising is practised.
Acknowledgements
The authors wish to thank the editor and the manuscripts three reviewers for their insightful and constructive comments and suggestions.
References
Abbott, A. (1998) The System of Professions: An Essay on the Division of Expert Labor. Chicago: University of Chicago Press. AMA Task Force on the Development of Marketing Thought (1988) Developing, disseminating and utilizing marketing knowledge. Journal of Marketing, 52 (October), pp. 115. Baker, S. & Holt, S. (2004) Making marketers accountable: a failure of marketing education. Marketing Intelligence and Planning, 22(5), pp. 557567. Baldridge, D.C., Floyd, S.W. & Markoczy, L. (2004) Are managers from Mars and academicians from Venus? Toward an understanding of the relationship between academic quality and practical relevance. Strategic Management Journal, 25 (November), pp. 10631074. Brennan, R. (2004) Should we worry about an academicpractitioner divide in marketing? Marketing Intelligence and Planning, 22(5), pp. 492500. Brennan, R. & Ankers, P. (2004) In search of relevance: is there an academicpractitioner divide in business-to-business marketing? Marketing and Intelligence Planning, 22(5), pp. 511519. Cohen, W.M. & Levinthal, D.A. (1990) Absorptive capacity: a new perspective on learning and innovation. Administrative Science Quarterly, 35 (March), pp. 128152. Crosier, K. (2004) How effectively do marketing journals transfer useful learning from scholars to practitioners? Marketing Intelligence and Planning, 22(5), pp. 540556. Easton, G. (2000) Is relevance relevant? Paper presented at the Annual Conference of the IMP Group, Bath, UK. Ehrenberg, A.S.C. (1969) The discovery and use of laws of marketing. Journal of Advertising Research, 9 (June), pp. 1117. Gabriel, H., Kottasz, K. & Bennett, R. (2006) Advertising planning, ad-agency use of advertising models, and the academic practitioner divide. Marketing Intelligence and Planning, 24(5), pp. 505527. Garda, R.A. (1988) Developing, disseminating, and utilizing marketing knowledge: comment. Journal of Marketing, 52 (October), pp. 3241. Holbrook, M.B. (1985) Why business is bad for consumer research: the three bears revisited. Advances in Consumer Research, 12, pp. 145156. Holbrook, M.B. (1987) What is consumer research? Journal of Consumer Research, 14 (June), pp. 128132.
443
Hunt, S.D. (2002a) Marketing as a profession: on closing stakeholder gaps. European Journal of Marketing, 36 (March), pp. 305312. Hunt, S.D. (2002b) Foundations of Marketing Theory: Toward a General Theory of Marketing. Armonk, NY: M.E. Sharpe. Katsikeas, C.S., Robson, M.J. & Hulbert, J.M. (2004) In search of relevance and rigour for research in marketing. Marketing Intelligence and Planning, 22(5), pp. 568578. Katz, H. (2007) Building bridges with industry to advance advertising knowledge. AAA Newsletter, 3(1) (March), pp. 1, 8. Kover, A.J. (1976) Careers and noncommunication: the case of academic and applied marketing research. Journal of Marketing Research, 13 (November), pp. 339344. Kover, A.J. (1995) Copywriters implicit theories of communication: an exploration. Journal of Advertising Research, 21 (March), pp. 596611. MacDonald, C. (1992) How Advertising Works: A Review of Current Thinking. Henley-onThames, UK: Advertising Association. Maiken, J.M., Myers, J.G., Peters, W.H., Schwartz, G. & Westing, J.H. (1979) What is the appropriate orientation for the marketing academician? in O.C. Ferrell, S.W. Brown & C.W. Lamb, Jr (eds) Conceptual and Theoretical Developments in Marketing. Chicago, IL: American Marketing Association, pp. 4975. McKenzie, C.J., Wright, S., Ball, D.F. & Baron, P.J. (2002) The publications of marketing faculty who are we really talking to? European Journal of Marketing, 36 (November), pp. 11961208. McQuarrie, E.F. (1998) Have laboratory experiments become detached from advertiser goals? A meta-analysis. Journal of Advertising Research, 38 (NovemberDecember), pp. 1525. November, P. (2004) Seven reasons why marketing practitioners should ignore marketing academic research. Australian Marketing Journal, 12(2), pp. 3950. Ottesen, G.G. & Gronhaug, K. (2004), Barriers to practical use of academic marketing knowledge. Marketing Intelligence and Planning, 22(5), pp. 520530. Parasuraman, A. (1982) Is a scientist versus technologist research orientation conducive to marketing theory development? in R.F. Bush & S.D. Hunt (eds) Marketing Theory: Philosophy of Science Perspectives. Chicago, IL: American Marketing Association. Peters, W.H. (1980) The marketing professorpractitioner gap: a possible solution. Journal of Marketing Education, 2 (Fall), pp. 411. Preston, I.L. (1985) The developing detachment of advertising research from the study of advertisers goals. Current Issues and Research in Advertising, 8(2), pp. 115. Rossiter, J.R. (2001) What is marketing knowledge? Stage 1: forms of marketing knowledge. Marketing Theory, 1(1), pp. 926. Southgate, N. (2006) The academicpractitioner divide: finding time to make a difference. Marketing Intelligence and Planning, 24(6), pp. 547551. Tapp, A. (2004a) A call to arms for applied marketing academics. Marketing Intelligence and Planning, 22(5), pp. 579590.
444
Tapp, A. (2004b) The changing face of marketing academia: what can we learn from commercial market research and practitioners? European Journal of Marketing, 38(56), pp. 492499. Webster, F.L., Jr (1988) Developing, disseminating, and utilizing marketing knowledge: comment. Journal of Marketing, 52 (October), pp. 4851. Wells, W.D. (1993) Discovery-oriented consumer research. Journal of Consumer Research, 19 (March), pp. 489504. Zaltman, G., LeMasters, K. & Heffring, M. (1982) Theory Construction in Marketing: Some Thoughts on Thinking. New York: Wiley.
About the authors

Gergely Nyilasy is a senior account executive in the New York City office of Hall & Partners, a market research firm specialising in advertising and brand research. Beyond regular quantitative research projects for various fmcg, B2B/technology and service industry clients, he plays a major role in the firms R&D and new product development initiatives. He graduated from the University of Georgia with a PhD in mass communication, where he also taught classes in copywriting (a former career). His research interests include digital communication, word of mouth, sociology of occupations/professionalisation and practitioneracademician relations. Leonard N. Reid is professor of advertising at the University of Georgia. He received his PhD from the University of Illinois at UrbanaChampaign. He is a Fellow of the American Academy of Advertising and former editor of the Journal of Advertising. His research on advertising and related topics has been published in many of the leading advertising, marketing, mass communication and health communication journals. His most recent research focuses on advertising knowledge and professional practice, trust in advertising, and direct-to-consumer prescription drug advertising. Address correspondence to: Greg Nyilasy, Hall & Partners New York, 72 Spring Street, New York, NY 10012, USA. Email: g.nyilasy@hall-and-partners.com
445
Public policy issues Global trends in advertising Media development Case histories Advertising effectiveness measurement World advertising statistics
This leading journal is a unique source of high quality research and practice-based articles, state-of-the-art studies, statistical analysis and polemical contributions from an international authorship of academics and practitioners. The Journals aim is to present articles that have originality, depth and clarity of insight into significant issues and developments of interest to advertising and related industries. Its editorial policy is to reflect a variety of perspectives and to invite the different viewpoints of advertising and marketing professionals, academics and those shaping and administering public policy.
Subscription Order Form

I wish to subscribe to the International Journal of Advertising (ISSN 0265 0487), UK/Europe Institution: Individual: 215/364 107/182 Rest of the World 215/$348/364 107/$174/182
I enclose a cheque for _____ made payable to WARC

Cheques in sterling drawn on a UK bank.
Please charge /$/_____ to my Visa/American Express/MasterCard Card No. Name Company Main Activity (please tick):
Individual Retailer/Distributor Research Agency Academic Institution Advertising or Media Agency Other (please specify): Media Owner Manufacturer
Expiry date: Job Title
Address
Tel. No.
(in case of order query)
Fax No. Date
Signed
If you do not wish to receive promotional mailings from other companies, please tick here
Please return this form with payment to: World Advertising Research Center, Farm Road, Henley-on-Thames, Oxfordshire, RG9 1EJ, UK Tel: +44 1491 411000 Fax: +44 1491 418600
IJOA_997
The Web Motivation Inventory

Replication, extension and application to internet advertising
Shelly Rodgers
University of Missouri-Columbia
Ye Wang
Missouri School of Journalism
Ruth Rettie
Kingston University
Frank Alpert The University of Queensland
The Web Motivation Inventory (WMI) is used in internet advertising research, and is frequently used and cited in advertising, marketing and communication literature. Investigations of the robustness of the WMI have been somewhat limited. Additionally, new uses of the internet are not accounted for by the WMI since its publication in 2002. This paper replicates and extends the original WMI using participants in the US, UK and Australia and includes internet motives not previously examined. The results show that the four-factor WMI remained reliable and valid for all three samples. Findings suggest the extended WMI may be broken into 12 sub-scales that represent the original fourfactor measure.
Introduction
Internet advertisers in the US spent $7.9 billion during the first six months of 2006 a 37% increase over the first half of 2005 (IAB 2006). Increase in online advertising reaffirms the internets growing importance for advertisers and marketers who want to engage consumers and leverage branding opportunities. From search engines, behavioural targeting, consumer-generated content, broadband and new emerging platforms such as mobile and iPTV, the internet is projected to continue to increase its share of total ad spending (IAB 2006).
447
Despite advances in interactive options available to internet advertisers and marketers, and a shift in spending towards these interactive venues, measuring the impact of online promotional efforts continues to be a challenge. As with traditional media, decisions made about online marketing strategies and ad expenditures must be justified to demonstrate some level of ad effectiveness. In this push for greater accountability, consumer motives for internet use have been identified as a key to understanding the effectiveness of interactive advertising strategies (Rodgers & Thorson 2000). Studies have shown that internet motives influence website effectiveness (Ko et al. 2005) as well as attitudes and behaviours towards internet advertising (Rodgers 2002). Numerous scales have been developed to measure consumer motives for internet use (e.g. Korgaonkar & Wolin 1999; Papacharissi & Rubin 2000) but the Web Motivation Inventory has received particular attention by internet advertising, communication and marketing scholars in the US and internationally (e.g. Sheehan 2002; Cai & Jun 2003; Faber et al. 2004; Francis & White 2004; Weib 2005; LaFerle & Kim 2006). The WMI was developed out of a comprehensive review of the literature, and a number of data collections and factor analyses have been conducted on US (Rodgers & Sheldon 2002) and international samples (Rodgers et al. 2005). However, the scale has not always produced consistent results and new uses of the internet are not accounted for by the WMI since its inception in 1997. For the WMI to be useful, it must be updated to reflect current internet motives. The scales unidimensionality must also be tested and validated with internet ad variables to provide a robust measure of internet motives that is useful to internet advertisers and marketers. This research replicates and extends the WMI, where replication broadly refers to a duplication of a target study and an extension is duplication with the alteration of one or more key parameters (Berthon et al. 2003). This was accomplished with two studies. Study 1 replicates the WMI with a survey of internet users from the US, UK and Australia. Study 2 replicates and extends the original WMI to include new internet motives; the extended version of the WMI is validated with confirmatory factor analysis and internet ad-specific items. The paper begins with a brief review of the WMI and then compares it to other similar measures to assess the scales properties. An overview is then provided about analyses to be conducted to examine the WMIs reliability and validity, and to
448
THE WEB MOTIVATION INVENTORY
evaluate the scales psychometric properties. The paper concludes with future research possibilities.
Internet motives: instrument evaluation

Web Motivation Inventory The Web Motivation Inventory (WMI) quantifies four primary internet motives on 12 five-point Likert scales (Rodgers & Sheldon 2002). The four motives are: research (information acquisition), communicate (socialisation), surf (entertainment) and shop. The instrument was developed out of a series of data collections, factor analyses, and reliability and validation techniques. Study 1 synthesised more than 100 internet motives from the literature, and conducted qualitative interviews with internet users and categorised these into four motives (see Rodgers & Sheldon 2002). Study 2 purified the instrument and narrowed the number of scale items from 15 to 12 using five-point Likert scales. Study 3 cross-validated the WMI using a student sample and Study 4 cross-validated the WMI using a non-student sample. The four-factor structure remained stable across both samples. Study 5 tested the temporal stability of the WMI with a testretest procedure that administered the scale at two different times to the same participants over a six-week period. All but the communication factor remained stable over the six weeks. Study 6 validated the WMI using a student sample. Individuals filled out the WMI and were exposed to 12 banner ads, three for each of the four motives, and rated the banners on ten five-point semantic differential scales that measured liking, persuasion and intent to click. Three of the four motives successfully predicted at least one of the dependent variables except for communication. Study 7 validated the WMI with the same procedures as in Study 6 using a non-student sample. Three of the four motives successfully predicted at least one of the dependent variables except for surfing (Rodgers & Sheldon 2002). A replication of the WMI was published in 2005. The 12-item WMI was administered to four groups of students in the US, UK, Australia and South Korea. The scale was translated into Korean and half of the Korean students were given the English and half the Korean version. Factor analysis was conducted with Varimax rotation, as were coefficient alphas.
449
The four-factor structure emerged consistently across all but the Korean sample; two of the motives, namely research and surf, loaded on the same factor for this sample. Notwithstanding its use and scale development procedures, the WMI has not yielded consistent results for every study in which it was examined. Investigations of the robustness of the WMI also have been somewhat limited. In all iterations noted above, principal component factor analyses with Varimax rotations were conducted, as were reliability tests using alpha coefficients. Factor analysis and coefficient alphas are needed but are not a sufficient test for the validation of an instrument; confirmatory factor analysis is also needed to assess the unidimensionality of a measure to determine the scales acceptability (Gerbing & Anderson 1988). To further assess the WMIs properties and identify possible new motives not currently measured by the scale, we compared the WMI to similar internet motives scales. A broad search of the literature was conducted across disciplines such as advertising, marketing, computer science and mass communication to identify scales for possible inclusion. Dozens of studies were located that identified and/or examined internet motives, also called uses and gratifications of the internet, defined as general dispositions that influence peoples inner desires to actively fulfil a need or want (Deci & Ryan 1985; Papacharissi & Rubin 2000). We limited our discussion to nine instruments based on three criteria: (1) instrument development was a major focus of the research; (2) individual scale items were identified and were predominantly focused on internet motives; and (3) internet motives were measured in a general (e.g. internet as a whole) vs specific context (e.g. email, eWOM, travel). To avoid repetition, measures that used or adapted existing internet motive scales were also excluded (e.g. Parker & Plank 2000; Sheehan 2002; Amiel & Sargent 2004). This helped to narrow the universe of scales and provided measures that were sufficiently similar to the WMI to make comparisons. Table 1 summarises the nine scales that met our criteria. The scales are listed in chronological order starting with Eighmeys (1997) user perceptions of commercial websites measure and ending with Song et al.s (2004) internet gratification factors. The number of items per scale ranged from 12 (Rodgers & Sheldon 2002) to 41 (Korgaonkar & Wolin 1999). Likert and semantic differential scales were employed by the
450
Table 1: Internet motives instrument evaluation

Number of items Author(s) Eighmey 1997 Journal of Advertising Research Year Journal 14 SD No No Factor analysis Scale Validity type* Reliability checks Method of analysis Replication(s) Not found
Instrument name Dimensions
User perceptions of commercial websites
Six factors: 1. Marketing perceptions 2. Entertainment value 3. Informational value 4. Ease of use 5. Credibility 6. Interactivity 15 SD No No Factor analysis Eighmey & McCord 1998
Uses and gratifications perspective of websites
Nine factor themes: 1. Entertainment value 2. Personal involvement 3. Personal relevance 4. Information involvement 5. Clarity of purpose 6. Controversy 7. Credibility 8. Interest in continuing communication 9. Purchase interest LIK 0.650.91 No Factor analysis Korgaonkar & Wolin
Journal of Business Research
Not found
Gratifications and concerns
Seven factors: 41 1. Social escapism 2. Transaction-based security and privacy concerns 3. Information 4. Interactive control 5. Socialisation 6. Non-transactional privacy concerns 7. Economic 27 LIK 0.780.93 No
1999
Journal of Advertising Research
Not found
Internet motives & electronic media
Five factors: 1. Interpersonal utility 2. Pass time 3. Information seeking 4. Convenience 5. Entertainment
Factor analysis
Papacharissi & Rubin
2000
Journal of Broadcasting
Not found
(continued)
451
452
Number of items Author(s) Flanagin & Metzger 2001 Human Communication Research Year Journal 21 SD No No Cluster analysis Scale Validity type* Reliability checks Method of analysis Replication(s) Not found 25 SD 0.730.82 No Factor analysis Stafford & Stafford 2001 Information Resources Management Journal 2002 Journal of Advertising Research LaRose & Eastin 2004 Journal of Broadcasting & Electronic Media Not found 12 LIK 0.700.93 Testretest Yes Factor analysis Rodgers & Sheldon Rodgers et al. 2005 Not found 29 SD 0.730.93 Yes Factor analysis and confirmatory factor analysis 36 SD 0.630.92 No Factor analysis Song et al. 2004 Cyber Psychology and Behavior Not found
Table 1: Internet motives instrument evaluation (continued)
Instrument name Dimensions
Motivations for communication technology use
Ten clusters: 1. Information 2. Learn 3. Play 4. Leisure 5. Persuasion 6. Social bonding 7. Relationship management 8. Problem solving 9. Status 10. Insight General consumer Five factors: motives for 1. Search accessing 2. Cognitive commercial areas 3. New and unique of the web 4. Social 5. Entertainment Web Motivation Four factors: Inventory (WMI) 1. Shop 2. Surf 3. Research 4. Communicate Expected outcomes Six factors 1. Activity outcomes 2. Monetary outcomes 3. Novel outcomes 4. Social outcomes 5. Self-reactive outcomes 6. Status outcomes Internet Seven factors gratification 1. Virtual community factors 2. Information seeking 3. Aesthetic experience 4. Monetary compensation 5. Diversion 6. Personal status 7. Relationship maintenance
*SD = semantic differential scale, LIK = Likert scale
instruments. Six of the scales reported alpha coefficients as evidence of the scales reliability and three did not (Eighmey 1997; Eighmey & McCord 1998; Flanagin & Metzger 2001). One conducted a validation check with confirmatory factor analysis (LaRose & Eastin 2004) and one, that we could find, had been replicated (Rodgers & Sheldon 2002). Three of the nine scales had been applied to internet advertising, including the WMI. A total of 59 internet motive factors are identified by the nine scales. Three primary factors emerged across the scales, including information, entertainment and socialisation. Information acquisition was identified by all nine scales with the following factors: informational value (Eighmey 1997), information involvement (Eighmey & McCord 1998), information (Korgaonkar & Wolin 1999), information seeking (Papacharissi & Rubin 2000), information (Flanagin & Metzger 2001), cognitive (Stafford & Stafford 2001), research (Rodgers & Sheldon 2002), novel outcomes (LaRose & Eastin 2004) and information seeking (Song et al. 2004). The second internet motive that emerged across all nine scales is socialisation, which broadly refers to the communication, interactivity and/or relationship-building function of the internet. Factors include: interactivity (Eighmey 1997), interest in continuing communication (Eighmey & McCord 1998), socialisation (Korgaonkar & Wolin 1999), interpersonal utility (Papacharissi & Rubin 2000), social bonding and relationship building (Flanagin & Metzger 2001), social (Stafford & Stafford 2001), communicate (Rodgers & Sheldon 2002), social outcomes (LaRose & Eastin 2004) and virtual community (Song et al. 2004). The third internet motive, namely entertainment, is captured by eight of the nine scales and is represented by the following factors: entertainment value (Eighmey 1997), entertainment value (Eighmey & McCord 1998), entertainment (Papacharissi & Rubin 2000), play (Flanagin & Metzger 2001), entertainment (Stafford & Stafford 2001), surf (Rodgers & Sheldon 2002), activity outcomes (LaRose & Eastin 2004) and aesthetic experience (Song et al. 2004). Shopping also emerged as a factor in five out of the nine scales, as represented by the following factors: purchase interest (Eighmey & McCord 1998), economic and transaction-based security/privacy concerns (Korgaonkar & Wolin 1999), shop (Rodgers & Sheldon 2002), monetary outcomes (LaRose & Eastin 2004) and monetary compensation (Song et al. 2004).
453
From this review, it appears that there are advantages and disadvantages of the WMI in its current state. The WMI has been tested and retested on a number of student, non-student, US and non-US samples, which increases its ability to generalise to different populations (Peter 1979). In comparison to existing measures, the validity and reliability of the WMI has been assessed at different points in time and across different samples, providing additional evidence of the scales construct validity and reliability. The WMI has been applied to internet advertising, suggesting its potential use in this area. The four predominant motives that emerged from our review information, socialisation, entertainment and shopping were present in four out of the nine scales including the WMI, suggesting that the scale captures perhaps the most predominant internet motives. However, the identification of internet motives beyond these four, such as pass time (Papacharissi & Rubin 2000), persuasion (Flanagin & Metzger 2001) and diversion (Song et al. 2004), suggest that additional internet motives may exist that are not captured by the WMI. The usefulness of any scale is dependent on the robustness of its psychometric properties (Churchill 1979). Given the somewhat limited assessment of the scales robustness and the inconsistencies produced by the scale, a reassessment of the WMI seems appropriate.
Overview
The research reported in the remainder of this paper describes a replication of the WMI and the development and validation of the extended WMI. First, we describe Study 1, which replicates the WMI on three different student samples of internet users in the US, UK and Australia. We provide evidence from Study 1 that the original four-factor structure of the WMI is reliable across samples. Then, we describe the procedures used to replicate and validate the WMI with confirmatory factor analysis (CFA) using three different student samples from the US, UK and Australia in Study 2. We provide evidence of the scales unidimensionality in measuring the construct, namely internet motives. We then describe the procedures used to generate and purify new uses of the internet. An exploratory factor analysis is undertaken to examine the original WMI and new motives collected in Study 2. Subsequent CFAs are conducted to examine the extended version of the WMI for each sample. We report
454
further evidence from a series of analyses in Study 2 of the WMIs discriminant, convergent and nomological validity.
Method
Study 1 Participants A pen-and-paper survey of the 12 items comprising the WMI was administered to undergraduate students at three universities: US (N = 185, male = 40% and female = 60%), UK (N = 337, male = 49% and female = 52%), and Australia (N = 129, male = 31% and female = 69%). Students received extra credit for their participation. Students are an appropriate sample since they are heavy users of the internet and are adequate when research addresses methodological issues such as scale development and validation (Chen et al. 2002). Survey items Survey items included the 12 items comprising the WMI, rated on fivepoint Likert scales ranging from (1) strongly disagree to (5) strongly agree. The stem, I use the internet to: was followed by the items: email other people, connect with my friends, make a purchase, do research, explore new sites, buy things, communicate with others, get information I need, surf for fun, find interesting web pages, purchase a product Ive heard about and find out things I need to know. Participants were also asked to list any additional uses of the internet, to be examined later on in Study 2. Results Factor analysis followed by Varimax rotation was used to reduce the dimensionality of the WMI (Bollen & Lennox 1991). Items with eigenvalues of 1.0 or higher and item loadings of 0.40 were retained (Nunnally & Bernstein 1993). The same four-factor structure resulted for the three samples (see Table 2). For the US sample, the resulting motives and their alphas were: research ( = 0.81), communicate ( = 0.78), shop ( = 0.93) and surf ( = 0.90), which explained 79% of the total item variance. For the UK sample, the resulting motives and their alphas were: research
455
Table 2: Rotated factor pattern from principal component analysis of the WMI (Study 1)
Motives US SAMPLE I use the internet to: Shop 0.94 0.93 0.93 0.83 0.88 0.92 0.85 0.83 0.80 0.80 0.80 0.90 1. Make a purchase 2. Buy things 3. Purchase a product Ive heard about 4. Explore new sites 5. Surf for fun 6. Find interesting web pages 7. Do research 8. Get information I need 9. Find out things I need to know 10. Email other people 11. Connect with my friends 12. Communicate with others Shop Surf Research Communicate Items
Surf
Research
Communicate
UK SAMPLE I use the internet to: Shop 0.95 0.96 0.92 0.73 0.85 0.86 0.82 0.82 0.81 0.82 0.85 0.76 1. Make a purchase 2. Buy things 3. Purchase a product Ive heard about 4. Explore new sites 5. Surf for fun 6. Find interesting web pages 7. Do research 8. Get information I need 9. Find out things I need to know 10. Email other people 11. Connect with my friends 12. Communicate with others
Surf
Research
Communicate
AUSTRALIAN SAMPLE I use the internet to: Shop 0.94 0.95 0.86 0.66 0.91 0.91 0.74 0.83 0.80 0.74 0.87 0.79 1. Make a purchase 2. Buy things 3. Purchase a product Ive heard about 4. Explore new sites 5. Surf for fun 6. Find interesting web pages 7. Do research 8. Get information I need 9. Find out things I need to know 10. Email other people 11. Connect with my friends 12. Communicate with others
Surf
Research
Communicate
456
( = 0.79), communicate ( = 0.75), shop ( = 0.95) and surf ( = 0.80), which explained 75% of the total item variance. For the Australian sample, the resulting internet motives and their alphas were: research ( = 0.74), communicate ( = 0.78), shop ( = 0.90) and surf ( = 0.84). The combined items explained 75% of the total item variance. The high-reliability alpha for each motive across the three countries demonstrates good internal consistency of the WMI scale. Study 2 Participants The participants in Study 2 were students at the same three universities examined in Study 1, including the US (N = 117; females: 79%, males: 21%), UK (N = 136; females: 45%, males: 55%) and Australia (N = 111; females: 67%, males: 33%). Students received extra credit for their participation. Survey items Survey items included the original WMI plus an additional 42 items collected from qualitative data in Study 1 and identified in the academic and trade literature. As with Study 1, the items were measured on five-point Likert scales ranging from (1) strongly disagree to (5) strongly agree. The stem, I use the internet to: was followed by the new items listed in the appendix. Analysis and results With the goal of replicating and extending the WMI, three sets of analyses were performed. First, a confirmatory factor analysis was conducted to assess the validity and reliability of the WMI instrument (the original 12 items) for the three countries. Second, a multi-group analysis of factorial invariance was conducted to assess the measurement equivalence of the WMI instrument within the three samples. Last, the relationship between the extended WMI and internet advertising variables was examined to establish the criterion-related validity and to further investigate the scales application to internet advertising.
457
CFA analysis of original WMI instrument A confirmatory factor analysis (CFA) was conducted to assess the validity and reliability of the WMI. CFA is a more rigorous test of equality than exploratory factor analysis (Alwin & Jackson 1981). First, the WMI structure was validated using overall responses from the three samples. Then, the WMI was validated with a multi-group analysis of factorial invariance, which rigorously assesses cross-national equivalence of measures (Calantone & Zhao 2001). A CFA using AMOS assessed the WMI factorial structures, convergent validity, discriminant validity and the internal consistency of the scale. Measurement items were force-loaded on the original four-factor WMI. The overall goodness of fit was evaluated according to the similarity of the predicted and actual correlation (Gerbing & Anderson 1988). Results show that the model fit indices are a good fit between the model and data (2 = 51.892 df = 41; p = 0.119; GFI = 0.978, AGFI = 0.958, NFI = 0.977, TLI = 0.992, RMSEA = 0.027) (see Figure 1). Table 3 documents the detailed estimates from the CFA analysis.
Table 3: CFA results Study 2 (three samples combined)

Standardised estimate SE 0.209 0.617 0.951 0.487 0.756 0.909 0.660 0.145 0.616 0.783 0.859 0.904 0.959 0.741 0.979 1.550 0.068
Items Labels C1 C2 C3 C1 R1 R2 R3 Su1 Su1 Su2 Su3 Sh1 Sh2 Sh3 Email other people Connect with my friends Communicate with others Email other people Do research Get information I need Find out things I need to know Explore new sites Explore new sites Surf for fun Find interesting web pages Make a purchase Buy things
Constructs Communicate motive Communicate motive Communicate motive Research motive Research motive Research motive Research motive Research motive Surf motive Surf motive Surf motive Shop motive Shop motive
Estimate 1.000* 4.092 5.608 0.623 1.000* 1.149 0.871 0.264 1.000* 1.248 1.475 1.000* 1.092 0.796
CR 4.182 3.619 9.195
p *** *** *** *** *** *** *** *** ***
0.075 15.296 0.068 12.730 0.086
3.056 0.002
0.113 11.068 0.133 11.079 0.041 26.385 0.043 18.607
Purchase a product Ive heard about Shop motive
* We constrained the parameter to fix the scale of the latent construct; *** p < 0.001
458
Figure 1: CFA for original WMI Study 2 (three samples)

C1 C2 C3 0.49 R1 R2 R3 0.76 0.91 0.66 0.15 0.35
Su1 Su2 Su3
0.21 0.62 0.95 0.50 Communicate motive
Research motive
0.28
0.62 0.78 0.86 Surf motive 0.13
0.19
Sh1 Sh2 Sh3
0.90 0.96 0.74 Shop motive
Convergent and discriminant validity Convergent validity was assessed with structural equation modelling by determining whether indicator variables loaded significantly on the specified construct (Gerbing & Anderson 1988). The significance of the associated t-value for the respective value of lambda was assessed (Byrne 1989). This information is presented in Table 3 along with each indicators loading coefficient. Results show that each variable yielded a highly significant estimate, which demonstrates high convergent validity (Gerbing & Anderson 1988).
459
An accepted test of discriminant validity is to constrain the estimated correlation parameter between pairs of constructs to 1.0 and conduct a chisquared difference test on the values obtained from the models containing the constrained pairs and the unconstrained model in which the correlation parameter varies freely (Joreskog 1971). The results yielded a significantly lower 2 value on the unconstrained model (as compared to the constrained model), which provides evidence that the comparison dimensions are not perfectly correlated (i.e. the motives are distinct) an indication that discriminant validity is achieved (Bagozzi & Phillips 1982). Calibration of original WMI instrument The overall CFA findings confirm the original WMI factorial structure and indicate two changes in the original scale. The first change was the communication factor. The original WMI item email other people shifted from the communication motive in Study 1 to the research motive in Study 2. This shift was evidenced by a higher loading for research (0.49) vs communication (0.21). The second change in the original WMI pertained to the item explore new sites. This item loaded exclusively on the surf motive in Study 1 but loaded on both the research motive (0.15) and the surf motive (0.62) in Study 2. Nomological validity Nomological validity assesses whether the construct of interest behaves as it should (Bagozzi 1980; Cronbach & Meehl 1955) with regard to its external relationships with other constructs (Reise et al. 1993). To confirm the nomological validity of the WMI, we investigated the relationship between the four motives and two criterion variables, specific to internet advertising. The first captured approach behaviour, measured by the item I tend to click on internet ads and the second measured avoidance behaviour with the item I typically try to avoid internet ads (Table 4). Table 5 provides the bivariate correlations between each WMI motive and criterion variable. The findings show that the surf (r = 0.209, p < 0.01) and shop motive (r = 0.139, p < 0.01) significantly correlated with approach behaviour, whereas the communication (r = 0.104, p < 0.05), research (r =
460
Table 4: Correlation matrix between WMI and criterion variables (Study 2)

Motive Communication Research Surf Shop I tend to click on internet ads 0.046 0.067 0.209** 0.139** I typically try to avoid internet ads 0.104* 0.132* 0.077 0.108*
** Correlation is significant at the 0.01 level (two-tailed) * Correlation is significant at the 0.05 level (two-tailed)
Table 5: Regression analysis (Study 2)

Dependent variables Predictors WMI: Communication motive Research motive Surf motive Shop motive 0.09 0.30 0.25 0.10 0.084** I tend to click on internet ads P 0.250 0.001 0.000 0.022 (0.074**) I typically try to avoid internet ads P 0.508 0.138 0.776 0.105 (0.016*)
0.08 0.22 0.03 0.12 0.027*
R2 (Adj. R2)
** Correlation is significant at the 0.01 level (two-tailed) * Correlation is significant at the 0.05 level (two-tailed)
0.132, p < 0.05) and shop motives (r = 0.108, p < 0.05) were associated with avoidance behaviour pertaining to internet ads. To further explore the predictive power of the WMI, we regressed each of the two criterion variables on the four motives (see Table 6). The results show that the four-factor model predicted approach behaviour (R2 = 0.084; adjusted R2 = 0.074; p < 0.01); the research ( = 0.30, p = 0.001), surf ( = 0.25, p = 0.000) and shop motives ( = 0.10, p = 0.022) were significant predictors of approach behaviour. The research motive decreased, and the surf and shop motives increased, consumers intention to click on internet ads. None of the four WMI motives significantly predicted avoidance behaviour, although the overall models predictive power was acceptable (R2 = 0.027; adjusted R2 = 0.016; p < 0.05).
461
Multi-group analysis of factorial invariance To assess the cross-national equivalence of WMI measures, CFA in the multi-group level was performed. An unconstrained CFA was conducted that allowed the factor structure to vary across samples in the US, UK and Australia. After a slight modification, the results revealed a high level of consistency in model form and measurement across the three groups. Factor loadings for each indicator for its respective construct were significant (p < 0.01) except for the items: connect with my friends and communicate with others for the US and Australian samples; explore new sites for the UK sample; and connect with my friends for the UK and Australian samples. Factor loadings yielded similar patterns for the three samples, shown in Tables 68, and the multi-group model yielded a good fit (2 = 118.272, df = 102; p = 0.129; GFI = 0.954, AGFI = 0.894, NFI = 0.952, TLI = 0.986, RMSEA = 0.021) (see Figure 2). Invariance testing (i.e. test of equivalence) is a particularly demanding test of an instruments robustness that helps diagnose measurement equivalence; minor changes in item loadings may not be critical to the interpretation of research results
Table 6: CFA results Study 2 (United States)
Standardised estimate SE 0.124 0.531 1.089 0.787 0.918 0.744 0.648 0.155 0.710 0.851 0.910 0.934 0.966 0.785 0.327 4.593 9.428 0.121 0.095 0.168 0.131 0.146
Items Labels C1 C2 C3 C1 R1 R2 R3 Su1 Su1 Su2 Su3 Sh1 Sh2 Sh3 C2 Email other people Connect with my friends Communicate with others Email other people Do research Get information I need Find out things I need to know Explore new sites Explore new sites Surf for fun Find interesting web pages Make a purchase Buy things Connect with my friends
Constructs Communicate motive Communicate motive Research motive Research motive Research motive Research motive Research motive Surf motive Surf motive Surf motive Shop motive Shop motive Shop motive
Estimate 1.000* 8.327 1.000* 1.222 1.040 0.688 0.361 1.000* 1.137 1.296 1.000* 1.072 0.897 0.313
CR
1.813 0.070 1.564 0.118 8.624 7.211 8.665 8.858 *** *** *** *** *** *** *** ***
Communicate motive 14.748
0.119 10.275
2.149 0.032
0.057 18.899 0.074 12.141 0.076 4.123
* We constrained the parameter to fix the scale of the latent construct *** p <0 .001
462
Table 7: CFA results Study 2 (United Kingdom)

Standardised estimate SE 0.346 0.683 0.794 0.716 0.920 0.595 0.437 0.120 0.638 0.762 0.824 0.913 0.943 0.633 0.053 0.654 0.867 0.134 0.108 0.113 0.124 0.168 0.184 0.071
Constructs Communicate motive Communicate motive Communicate motive Research motive Research motive Research motive Research motive Research motive Surf motive Surf motive Surf motive Shop motive Shop motive Shop motive
Estimate 1.000* 2.457 2.455 1.000* 1.120 0.776 0.603 0.203 1.000* 1.166 1.332 1.000* 1.033 0.662 0.055
CR 3.756 2.832 8.370 7.158 5.340 1.644 6.956 7.223 9.313
p *** 0.005 *** *** *** 0.100 *** *** *** *** 0.539
0.071 14.453 0.090 -0.614
* We constrained the parameter to fix the scale of the latent construct *** p < 0.001
Table 8: CFA results Study 2 (Australia)

Standardised estimate SE 0.140 0.563 1.048 0.784 0.901 0.666 0.459 0.195 0.456 0.723 0.898 0.845 0.916 0.747 0.003 3.293 6.461 0.123 0.119 0.129 0.170 0.376 0.527 0.095
Constructs Communicate motive Communicate motive Communicate motive Research motive Research motive Research motive Research motive Research motive Surf motive Surf motive Surf motive Shop motive Shop motive Shop motive
Estimate 1.000* 5.252 9.673 1.000* 1.186 0.864 0.575 0.336 1.000* 1.650 2.206 1.000* 1.115 0.879 0.003
CR
1.595 0.111 1.497 0.134 9.668 7.261 4.466 4.385 4.184 9.229 *** *** *** *** *** *** ***
1.983 0.047
0.099 11.235
0.079 0.036 0.971
463
Figure 2: CFA for original WMI Study 2 (multi-group comparison: US sample shown in CFA factorial structure)
C1 C2 C3 0.12 0.53 1.09 0.65 0.43 R1 R2 R3 0.79 0.92 0.74 0.15 0.33
Su1 Su2 Su3
Communicate motive
Research motive
0.26
0.36
0.22
0.71 0.85 0.91 Surf motive 0.26
0.24
Sh1 Sh2 Sh3
0.93 0.97 0.78 Shop motive
(Byrne & Shavelson 1987). Thus, despite minor differences in item-factor loadings, it was concluded that the three samples have virtually the same pattern of variables associated with the four latent motive constructs (McGowan & Sternquist 1998). To further test the equivalence of the measurement model across the three samples, a constrained CFA was conducted. If the measurement properties are the same for the three samples, factor patterns and factor loadings should be equal. Therefore, we set the factor structure to be invariant in the constrained model. The results indicated identical factor
464
patterns. Compared with the unconstrained model, the difference in 2 value of 36.72 with df of 22 suggests that factor structure was indifferent across samples. The constrained model also exhibited a good fit (2 = 155.024, df = 124; p = 0.031; GFI = 0.940, AGFI = 0.886, NFI = 0.937, TLI = 0.978, RMSEA = 0.026). Extension of web motivation inventory Data analysis To examine the validity and reliability of the extended version of the WMI, exploratory factor analysis with the original 12 items and 42 new items was used to explore the emergence of possible new internet motive factors. An eigenvalue of 1.00 was used as the stopping criterion rather than force a specified factor solution. Principal component analysis with Varimax rotation yielded 12 factors, explaining 67% of the matrix variance (see Table 9). The new factors were labelled: Community, Entertainment, Product trial, Information, Transaction, Game, Survey, Downloads, Interaction, Search, Exploration and News. The highest three loadings above 0.40 were used to represent each factor, as shown in Table 9. The 12 factors were mapped conceptually to form the sub-factors of the original WMI scale where: Community, Survey and Interaction are submotives of the original Communication motive; Information, Search and Exploration are sub-motives of the original Research motive; Entertainment, Game and Downloads are sub-motives of the original Surf motive; and Transaction and Product trial are sub-motives of the original Shopping motive. News emerged as a sub-motive for both of the original Research and Surf motives (see Figure 3). A confirmatory factor analysis (CFA) was then conducted to assess whether the conceptual structure of the WMI fit the data. Table 10 shows that although the 2 statistic (2 = 1051.272, df = 500, p < 0.0001) did not indicate a perfect fit, it is not always the best indication of model fit (e.g. Bagozzi & Yi 1988; Mulaik et al. 1989); therefore a range of additional fit indices is reported. The additional indices taken together (GFI = 0.859, AGFI = 0.833, NFI = 0.851, TLI = 0.905, CFI = 0.915, RMSEA = 0.054) indicate a reasonably good fit between the data and extended version of the WMI.
465
Table 9: Exploratory factor analysis with original WMI and new motives (Study 2)
Sub-motives 1. Community Items Get to know other people Participate in an online chat Join a group Amuse myself Entertain myself Find information to entertain myself Try on the latest fashions Experience a product Try out a product Do research Get information I need Search for information I need Make a purchase Buy things Purchase a product Ive heard about Play online games Entertain myself with internet games Play online games with individuals from other countries Take a survey on a topic I care about Fill out an online survey Give my opinion on a survey Download music Listen to music Watch online videos Connect with my friends Communicate with others Instant message others I know Get answers to specific questions Find information I can trust Find interesting web pages Explore new sites Surf for fun Read about current events and news Read entertainment news Loading 0.767 0.754 0.733 0.761 0.758 0.714 0.716 0.659 0.652 0.827 0.804 0.693 0.907 0.901 0.828 0.874 0.858 0.706 0.799 0.771 0.674 0.694 0.532 0.450 0.784 0.712 0.511 0.727 0.536 0.703 0.635 0.578 0.595 0.580 Variance explained 11.30
2. Entertainment
8.07
3. Product trial
7.83
4. Information
7.00
5. Transaction
5.89
6. Game
5.14
7. Survey
4.93
8. Downloads
3.91
9. Interaction
3.75
10. Search
3.48
11. Exploration
3.42
12. News
2.71
466
Figure 3: Extended WMI conceptual model Study 2

WMI sub-motives
Community Communication motive Survey
WMI motives
Interaction
Information
Search
Research motive
Exploration
News
Entertainment
Game
Surf motive
Downloads
Transaction
Shopping motive
Product trial
467
Table 10: CFA results for extended WMI (Study 2)

Labels Sub-motives: Community Survey Interaction Exploration Product trial Information Search Exploration News Interaction Entertainment Entertainment Game Downloads News Exploration Product trial Transaction Product trial Survey Search Items: Play online games Entertain myself with internet games Play online games with individuals from other countries Download music Listen to music Watch online videos Try on the latest fashions Experience a product Try out a product Get to know other people Participate in an online chat Join a group Take a survey on a topic I care about Fill out an online survey Give my opinion on a survey Connect with my friends Communicate with others Instant message others I know Do research Get information I need. Search for information I need Get answers to specific questions Find information I can trust Find information to entertain myself Entertain myself Amuse myself Make a purchase Buy things Purchase a product Ive heard about Find interesting web pages Surf for fun Explore new sites Read about current events and news Read entertainment news Constructs Estimate Standardised estimate SE CR p
Communication motive Communication motive Communication motive Communication motive Communication motive Research motive Research motive Research motive Research motive Research motive Research motive Surf motive Surf motive Surf motive Surf motive Surf motive Surf motive Shopping motive Shopping motive Shopping motive Shopping motive Game Game Game Downloads Downloads Downloads Product trial Product trial Product trial Community Community Community Survey Survey Survey Interaction Interaction Interaction Information Information Information Search Search Entertainment Entertainment Entertainment Transaction Transaction Transaction Exploration Exploration Exploration News News
1.000* 0.482 0.266 0.026 0.482 1.000* 0.634 0.767 0.690 0.897 0.841 1.000* 1.114 1.279 0.554 0.766 0.015 1.000* 1.066 0.976 0.674 1.000* 1.042 0.747 1.000* 0.891 0.873 1.000* 1.269 1.045 1.000* 0.943 0.822 1.000* 1.122 1.069 1.000* 0.845 0.980 1.000* 1.019 1.109 1.000* 1.256 1.000* 0.895 0.991 1.000* 1.089 0.811 1.000* 0.963 0.756 1.000* 1.375
0.931 0.517 0.293 0.025 0.576 0.824 0.528 0.418 0.539 0.558 0.434 0.604 0.605 0.747 0.506 0.487 0.012 0.406 0.639 0.525 0.497 0.889 0.932 0.729 0.747 0.675 0.708 0.635 0.847 0.746 0.809 0.746 0.768 0.762 0.814 0.834 0.778 0.741 0.561 0.785 0.841 0.683 0.539 0.696 0.864 0.758 0.818 0.904 0.958 0.748 0.817 0.842 0.654 0.624 0.766
0.075 0.061 0.110 0.124 0.128 0.148 0.119 0.132 0.141 0.145 0.164 0.086 0.190 0.143 0.247 0.219 0.180 0.044 0.044 0.081 0.077 0.106 0.092 0.067 0.057 0.074 0.070 0.077 0.106 0.068 0.087 0.214 0.054 0.054 0.041 0.044 0.059 0.059 0.147
6.463 4.381 0.236 3.889 4.944 5.198 5.813 6.775 5.977 7.684 7.810 6.470 4.041 0.108 4.318 4.458 3.741 23.600 17.071 10.953 11.356 11.985 11.372 14.160 14.537 15.116 15.355 10.961 9.244 14.923 12.757 5.865 16.662 18.406 26.471 18.544 16.450 12.758 9.331
*** *** 0.813 *** *** *** *** *** *** *** *** *** *** 0.914 *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
468
Discussion and conclusion

Overview The purpose of this research was to replicate and extend the original Web Motivation Inventory (WMI), which categorises internet motives into four primary factors: research, communication, surfing and shopping. The scales psychometric properties were examined with two studies: Study 1 replicated the WMI with samples from the US, UK and Australia, and Study 2 replicated and extended the WMI with three new samples from the same three countries. The results presented here provide support for the psychometric soundness of the original WMI and the extended WMI. The original WMI was shown to be valid and reliable with US and non-US samples with somewhat different market and consumer settings, suggesting that the WMI is a culturally neutral scale that may be useful in a global environment. The extended WMI yielded 12 new sub-motives that proved to be adequate extensions of the original four-factor scale, providing additional evidence of the scales unidimensionality, validity and reliability. The original WMI yielded meaningful responses among internet users on the approach and avoidance behavioural items regarding internet advertising, thereby suggesting that the scale possesses useful advertising and managerial applications. Limitations and future research There are several limitations and directions for future research. First, although the respondents were from three different countries, the three countries were English-speaking and shared perhaps other cultural characteristics. It may be informative to consider how the three countries differ, and also how cultures within the same country differ in their responses to the WMI. Ideally, countries beyond those examined here could also be included in future studies that examine the WMI. Second, follow-up studies could consider possibly improving the original and extended versions of the WMI by rewording items that comprise, for example, the communication motive. The items communicate with others and connect with my friends appear to be somewhat confusing to
469
some respondents. A re-examination of these and other items using different wording would help to further establish the cross-national equivalence of the scale (see Ewing et al. 2002). Third, a limitation of this research is that the samples were undergraduate students. We argued earlier that this is an appropriate sample given the task of validating a scale, particularly across different countries where individuals likely differ with regard to internet needs and motives. A future direction for research is to collect more data from different samples using an improved version of the WMI, and then examine whether the items that shifted from one motive to another in the current research could be applied to other samples. The items that shifted here could be sample and/or country specific (Strasheim et al. 2007). More research is needed to determine which items to retain. It may also be advantageous to include more items that have emerged in the trade and academic literature since the writing of this paper and then reassess the stability and psychometric robustness of the WMI instrument. Additionally, it may be useful to consider other forms of new technology that either supplement and/or replace current applications of the internet to account for possible shifts in the items that comprise the WMI, particularly with regard to communication (e.g. Svennevig 2000). Fourth, internet advertising practitioners and scholars can consider additional external factors with the goal of building a more comprehensive model of how internet motives influence individuals psychological and behavioural responses to internet ads. In addition to internet ad approach and avoidance behaviour, other variables may include: recall and recognition of internet ads, attitude towards the site, and purchase intentions, to name a few. Last, the WMI may be applied to scholarly research to better understand and develop theories about how internet users perceive and process advertising stimuli. Future research could consider pairing the WMI with existing internet advertising scales and constructs such as interactivity (McMillan & Hwang 2002), virtuality (Griffith & Chen 2004), and website personality (Chen & Rodgers 2006) to examine applications of the WMI in the field of internet advertising and marketing research.
470
Conclusion
If the state of internet marketing and advertising is to develop beyond its current condition, it is useful to assess the scales that have been developed for use in this area. Findings presented here suggest that the original and extended versions of the WMI are valid and reliable scales. Internet advertisers may therefore use the WMI with confidence when gauging consumer motives for internet use.
Appendix: Web Motivation Inventory with new motives

Please circle the number that best represents your agreement or disagreement
I use the internet to Strongly disagree Strongly agree
Original 12 items: Email other people Connect with my friends Make a purchase Do research Explore new sites Buy things Communicate with others Get information I need Surf for fun Find interesting web pages Purchase a product Ive heard about Find out things I need to know Additional 42 items: Download music Watch online videos Play online games Read entertainment news Entertain myself with internet games Instant message others I know Take a survey on a topic I care about Play online games with individuals from other countries Search for information I need Find information to entertain myself
1 1 1 1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3 3 3
4 4 4 4 4 4 4 4 4 4 4
5 5 5 5 5 5 5 5 5 5 5
1 1 1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3 3
4 4 4 4 4 4 4 4 4 4
5 5 5 5 5 5 5 5 5 5
(continued)
471
I use the internet to Amuse myself Use a search engine to find the latest trends Read about current events and news Write a product review Give my opinion on a survey Create a wish list on a website Buy a gift card for a family member or friend Talk to a knowledgeable individual Listen to other peoples problems Respond to a posting on a website Join a group Participate in an online chat Try on the latest fashions Experience a product Download coupons Shop for the best price on a product Comparison shop Get to know other people Chat with others Find information for someone else Find a website someone recommended to me Fill out an online survey Try out a product Entertain myself Visit entertaining websites Talk to a live person Build a relationship with others Talk to a person in real time Get answers to specific questions Find information I can trust Cast my vote Listen to music
Strongly disagree 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4
Strongly agree 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
References
Alwin, D.F. & Jackson, D.J. (1981) Applications of simultaneous factor analysis to issues of factorial invariance, in Jackson, D.D. & Borgotta, E.P. (eds) Factor Analysis and Measurement in Sociological Research: A Multidimensional Perspective. Beverly Hills, CA: Sage. Amiel, T. & Sargent, S.L. (2004) Individual differences in internet usage motives. Computers in Human Behavior, 20(6), pp. 711726.
472
Bagozzi, R.P. (1980) Causal Models in Marketing. New York: Wiley. Bagozzi, R.P. & Phillips, L.W. (1982) Representing and testing organizational theories: a holistic construal. Administrative Science Quarterly, 27(3), pp. 459489. Bagozzi, R.P. & Yi, Y. (1988) On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), pp. 7494. Berthon, P., Ewing, M., Pitt, L. & Berthon, J.P. (2003) Reframing replicative research in advertising. International Journal of Advertising, 22(9), pp. 511530. Bollen, K. & Lennox, R. (1991) Conventional wisdom on measurement: a structural equation perspective. Psychological Bulletin, 110(2), pp. 305314. Byrne, B.M. (1989) A Primer of LISREL. New York: Springer-Verlag. Byrne, B.M. & Shavelson, R.J. (1987) Adolescent self-concept: testing the assumption of equivalent structure across gender. American Educational Research Journal, 24(3), pp. 365385. Cai, S. & Jun, M. (2003) Internet users perceptions of online service quality: a comparison of online buyers and information searchers. Managing Service Quality, 13(6), 504519. Calantone, R.J. & Zhao, Y.S. (2001) Joint ventures in China: a comparative study of Japanese, Korean, and US partners. Journal of International Marketing, 9(1), pp. 123. Chen, Q., Clifford, S.J. & Wells, W.D. (2002) Attitude toward the site II: new information. Journal of Advertising Research, 42(2), pp. 3345. Chen, Q. & Rodgers, S. (2006) Development of an instrument to measure web site personality. Journal of Interactive Advertising, 7(1), online at http://jiad.org/vol7/no1/chen/index.htm. Churchill, G.A., Jr (1979) A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16(1), pp. 6473. Cronbach, L.J. & Meehl, P.E. (1955) Construct validity in psychological tests. Psychological Bulletin, 52(4), pp. 281302. Deci, E.L. & Ryan, R.M. (1985) The general causality orientations scale: selfdetermination in personality. Journal of Research in Personality, 19(2), pp. 109134. Eighmey, J. (1997) Profiling user responses to commercial web sites. Journal of Advertising Research, 37(3), pp. 5966. Eighmey, J. & McCord, L. (1998) Adding value in the information age: uses and gratifications of sites on the World Wide Web. Journal of Business Research, 41(3), pp. 187194. Ewing, M.T., Caruana, A. & Zinkhan, G.M. (2002) On the cross-national generalisability and equivalence of advertising response scales developed in the USA. International Journal of Advertising, 21(3), pp. 323343. Faber, R.J., Lee, M. & Nan, X. (2004) Advertising and the consumer information environment online. American Behavioral Scientist, 48(4), pp. 447466. Flanagin, A.J. & Metzger, M.J. (2001) Internet use in the contemporary media environment. Human Communication Research, 27(1), pp. 153181. Francis, J.E. & White, L. (2004) Value across fulfillment-product categories of internet shopping. Managing Service Quality, 14(2/3), pp. 226234.
473
Gerbing, D.W. & Anderson, J.C. (1988) An updated paradigm for scale development incorporating unidimensionality and its assessment. Journal of Marketing Research, 25(2), pp. 186192. Griffith, D. & Chen, Q. (2004) The influence of virtual direct experience on online ads message effectiveness. Journal of Advertising, 33(1), pp. 5569. Internet Advertising Bureau (2006) Internet advertising revenues continue to accelerate at an unprecedented rate with a 36% increase for first half of 06. Press release online at http://iab.net (accessed 11 December 2006). Joreskog, K.G. (1971) Statistical analysis of sets of congeneric tests. Psychometrika, 36(2), pp. 109133. Ko, H., Cho, C.H. & Roberts, M.S. (2005) Internet uses and gratifications. Journal of Advertising, 34(2), pp. 5770. Korgaonkar, P.K. & Wolin, L.D. (1999) A multivariate analysis of web usage. Journal of Advertising Research, 39(2), pp. 5368. LaFerle, C. & Kim, H.J. (2006) Cultural influences on internet motivations and communication styles: a comparison of Korean and US consumers. International Journal of Internet Marketing and Advertising, 3(2), pp. 142157. LaRose, R. & Eastin, M.S. (2004) A social cognitive theory of internet uses and gratifications: toward a new model of media attendance. Journal of Broadcasting & Electronic Media, 48(3), pp. 358377. McGowan, K.M. & Sternquist, B.J. (1998) Dimensions of price as a marketing universal: a comparison of Japanese and US consumers. Journal of International Marketing, 6(4), pp. 4966. McMillan, S. & Hwang, J.S. (2002) Measures of perceived interactivity: an exploration of communication, user control, and time in shaping perceptions of interactivity. Journal of Advertising, 31(3), pp. 4154. Mulaik, S.A., James, L.R., Van Alstine, J., Bennett, N., Lind, S. & Stilwell, C.D. (1989) Evaluation of goodness-of-fit indices for structural equation models. Psychological Bulletin, 105(3), pp. 430445. Nunnally, J.C. & Bernstein, I.H. (1993) Psychometric Theory (3rd edn). Englewood Cliffs, NJ: McGraw-Hill. Papacharissi, A. & Rubin, A.M. (2000) Predictors of internet use. Journal of Broadcasting & Electronic Media, 44(2), pp. 175196. Parker, B.J. & Plank, R.E. (2000) A uses and gratifications perspective on the internet as a new information source. American Business Review, 18(2), pp. 4349. Peter, J.P. (1979) Reliability: a review of psychometric basics and recent marketing practices. Journal of Marketing Research, 16(1), pp. 617. Reise, S.P., Widaman, K.F. & Pugh, R.H. (1993) Confirmatory factor analysis and item response theory: two approaches for exploring measurement invariance. Psychological Bulletin, 114(3), pp. 552566. Rodgers, S. (2002) The interactive advertising model tested: the role of motives in ad processing. Journal of Interactive Advertising, 2(2), available online at http://jiad.org/vol2/no2/rodgers. Rodgers, S. & Sheldon, K.M. (2002) An improved way to characterize internet users. Journal of Advertising Research, 42(5), pp. 8594.
474
Rodgers, S. & Thorson, E. (2000) The interactive advertising model: how users perceive and process online ads. Journal of Interactive Advertising, 1(1), available online at http://jiad.org/vol1/no1/rodgers. Rodgers, S., Jin, Y., Rettie, R., Alpert, F. & Yoon, D. (2005) Internet motives of users in the United States, United Kingdom, Australia, and Korea: a cross-cultural replication of the WMI. Journal of Interactive Advertising, 6(1), online at http://jiad.org/vol6/no1/rodgers/index.htm (accessed on 11 May 2006). Sheehan, K.B. (2002) Of surfing, searching, and newshounds: a typology of Internet users online sessions. Journal of Advertising Research, 42(5), pp. 6271. Song, I., LaRose, R., Eastin, M.S. & Lin, C. (2004) Internet gratifications and Internet addiction: on the uses and abuses of new media. Cyber Psychology and Behavior, 7(4), pp. 385395. Stafford, T.F. & Stafford, M.R. (2001) Identifying motivations for the use of commercial web sites. Information Resources Management Journal, 14(1), pp. 2230. Strasheim, A., Pitt, L. & Caruana, A. (2007) Psychometric properties of the Schlinger Viewer Response Profile: evidence from a large sample. Journal of Advertising, 36 (forthcoming). Svennevig, M. (2000) Needs, not nerds: researching technological change. International Journal of Advertising, 19(5), pp. 645663. Weiss, D. (2005). Internetnutzung im studium Erklarung and vorhersage der internetnutzung von studierenden in Deutschland und Osterreich. PsychologieMedienpsychologie, 135(1, 3), p. 135.
About the authors

Shelly Rodgers, Ph.D., is an associate professor of strategic communication at the Missouri School of Journalism. Her research areas are interactive advertising, marketing and communication with emphasis on information processing of health messages in new media contexts. Qimei Chen, Ph.D., is chair of the Department of Marketing and associate professor of marketing in the Shidler College of Business, University of Hawaii at Manoa. Her research interests include advertising effectiveness, online/offline consumer behavior and innovation knowledge. Ye Wang is an MA student of strategic communication at the Missouri School of Journalism. Her research interests include international and cross-cultural advertising and new technology. Ruth Rettie, Ph.D., is an assistant professor and director of the eCommerce Consumer Research Unit in the School of Marketing, Kingston University, London. Her research interests include eCommerce, internet marketing and communication theory.
475
Frank Alpert, PhD, is an associate professor at the UQ Business School, University of Queensland, Australia. His research areas include marketing strategy, marketing management, brand management, product management, marketing education and entertainment software industry. Address correspondence to: Shelly Rodgers, Associate Professor, Strategic Communication, University of Missouri-Columbia, 364 McReynolds Hall, Columbia, MO 65211 Email: srodgers@missouri.edu
476
More than meets the eye

Investigating the hidden impact of brand placements in television magazines
Jrg Matthes, Christian Schemer and Werner Wirth
University of Zurich
A growing research literature suggests that visual brand placements need not be recalled in order to have an impact on brand attitudes. In an experimental study working with an authentic television broadcast, the authors investigated the influence of frequency of placement exposure, viewers involvement and viewers persuasion knowledge on the attitudes towards the brand and brand recall. Results clearly indicate a mere exposure effect. A frequently presented brand placement can have a positive effect on brand evaluations although viewers do not recall the brand. However, this effect can only be found when there is a high involvement in the programme and low persuasion knowledge. In contrast, when persuasion knowledge is high and involvement is low, frequently presented placements lead to a deterioration of brand attitudes.
Introduction
In todays oversaturated media landscape, product placements are seen as promising tools for the advertising of brands because they are less likely to be recognised as persuasive attempts than more obtrusive forms of advertising. However, throughout almost the entire research literature it is assumed that product placements must be recognised or recalled in order to influence brand evaluations. Therefore, in studying the effects of product placements, scholars have focused primarily on measuring brand recall. Accordingly, many studies seem to indicate that the hidden nature of product placements makes them hard to be recognised, and, therefore, there might often be no effect on brand recall (Russell 1998; Karrh et al. 2003). In fact, when brands are not recalled, placements are thought to be unsuccessful (Karrh et al. 2003).
477
More recently, however, there have been some hints in the research literature that brand placements need not be recalled in order to have an impact on brand attitudes. As Law and Braun-LaTour (2004, p. 64) argue, recall and recognition measures are not capable of detecting the more subtle effects of product placements. For instance, research by Law and Braun (2000) and Russell (2002) suggests that placement recall might be independent of placement evaluation. This view is supported by research in cognitive psychology on the so-called mere exposure effect (Zajonc 1968; Kunst-Wilson & Zajonc 1980). According to the mere exposure effect, a repeated exposure to a stimulus can lead to favourable feelings towards that stimulus, even though people do not remember that they were exposed to it. In recent years, the consideration of mere exposure effects has become an important and lively research area in advertising, consumer research and marketing research (Baker 1999; Lee and Labroo 2004; Fang et al. 2007; Grimes & Kitchen 2007; Tom et al. 2007). As Grimes and Kitchen (2007, p. 192) argue, the development of methodologies to study mere exposure effects represents one of the major contemporary challenges for marketing research. Despite the relevance of this research area, however, little is known about mere exposure effects when it comes to product placements. In fact, although the mere exposure effect is regarded as a possible and important mechanism for the impact of product placements (Russell 2002; McCarty 2004), there are to date no studies in advertising research that have put this notion to a direct test. Moreover, there is a dearth of studies investigating the specific conditions under which mere exposure effects are likely to occur in the context of product or brand placements. This is of utmost importance because product placements are embedded in a rather complex visual field. Unlike traditional mere exposure stimuli in cognitive psychology, placements are not presented directly to viewers; in contrast, they are popping up in moving pictures and compete against a rich array of other stimuli that are a part of the plot. Therefore, the crucial question is, are there boundary conditions for the mere exposure effects of product placements? The aim of this paper is to fill these pressing research gaps. The paper will first discuss previous research that has dealt with the impact of product placements on viewers recall and brand evaluations. It will then briefly summarise research on the mere exposure effect in the context of cognitive psychology, and will discuss several explanations of the mere
478
BRAND PLACEMENTS IN TELEVISION MAGAZINES
exposure effect. Following this there will be a discussion of possible boundary conditions or moderators of mere exposure effects in the context of product placements. Based on these insights, an experiment will be presented that explicitly tests the notion of mere exposure effects for product placements.
The impact of product placements on recall and attitudes

A plethora of studies call attention to the ways in which product placements influence viewers brand recall (e.g. Karrh et al. 2003; Brennan and Babin 2004; Law and Braun-LaTour 2004). The idea is that placements can activate already existing images. Or, if such images do not exist, placements can enhance brand awareness. Therefore, in order for the placements to have an impact, they must be recognised in the first place. According to this line of reasoning, placements that are not attentively processed and that cannot be recalled later on are rather useless. However, previous studies have shown that viewers generally show a relatively low brand recall (Babin & Carder 1996; Law & Braun 2000; Tiwsakul et al. 2005). Although recall can be increased when placements are more obtrusive (Brennan et al. 1999; dAstous & Chartier 2000), it can be concluded that simple, unobtrusive screen placements are rather marginally efficient in terms of brand recall (Law & Braun-LaTour 2004). These results point to a very serious dilemma. On the one hand, the subtlety of placements makes them attractive for practitioners because viewers do not recognise the persuasive attempt, and, therefore, strategies of counter-arguing become less likely (dAstous & Chartier 2000). On the other hand, however, the non-aggressive, non-persuasive method of placing a brand runs the risk of being overlooked by the viewers. One explanation is that placements compete against a rich array of other stimuli that are a part of the plot. In contrast to research gauging the impact of placements on brand recall, there are comparatively few studies that examine effects on brand attitudes. In most instances, it is assumed that recall is the precondition for any attitudinal effects (Law & Braun-LaTour 2004). Other researchers argue that placements must have an impact on attitudes simply because they are more accepted and more tolerated by the audience (Karrh 1998;
479
Tiwsakul et al. 2005). However, tolerating placements does not necessarily mean that placements influence attitudes and evaluations. Recently, several researchers advanced the notion that mere exposure effects may be at work when viewers are confronted with placements (e.g. Russell 2002; McCarty 2004). According to this reasoning, a placement can have an impact on attitudes, even though viewers cannot recall the brand or the product. This mere exposure effect will be outlined in more detail in the next section.
The mere exposure effect: an unmediated phenomenon

The mere exposure effect is a learning effect that apparently can take place entirely outside of conscious awareness, involving implicit rather than explicit knowledge (Bornstein & DAgostino 1992, p. 545). A plethora of studies in cognitive psychology have shown that the repeated mere exposure to a previously unknown stimulus leads to an increase in liking for that stimulus. As Zajonc (2004, p. 198) puts it, a stimulus object is made accessible to the individuals sensory receptors without offering any positive or negative reinforcement. Mere exposure effects have been shown for a wide array of stimuli, ranging from simple syllables to sounds or pictures of persons, and from supraliminal to subliminal exposure times (Zajonc 1968, 2001; Bornstein 1989). Typically, this line of research is based on focal exposure to target stimuli, and the frequencies of exposure are experimentally varied. The mere exposure effect works best under low-attention conditions and for very short presentation times (Bornstein 1989). In a seminal study by Moreland and Zajonc (1977), Japanese ideographs were presented to subjects at five different frequencies (0, 1, 3, 9, 27 times). After that, the subjects were asked to evaluate the stimuli as positive or negative. The results showed an enhancement of liking with increasing stimulus exposure. Even more importantly, significant correlations between frequency and liking were found, but not between liking and recognition. These results suggest that, in contrast to exposure frequency, recognition is not necessary for the enhancement of liking. This idea has become the fundamental backdrop of mere exposure research over the past 40 years (Zajonc 1968, 2001, 2004), and it has been replicated in a vast number of experiments (see Bornstein 1989). Summing up the comprehensive evidence for mere exposure effects, Zajonc (2004)
480
concludes that they are unmediated that is, recognition is not a necessary mediator of the exposure effect (p. 200). Several explanations for the occurrence of mere exposure effects have been proposed (for a review see Fang et al. 2007). Zajonc (2001, 2004) draws on a classical conditioning explanation, arguing that the absence of aversive consequences following the presentation of the conditioned stimulus is a sufficient condition for the development of an approach tendency to that stimulus (Zajonc 2004, p. 202). However, the most prominent and widely accepted mechanism is a perceptual fluency-based account of mere exposure effects (Winkielman et al. 2003; Lee & Labroo 2004; Reber et al. 2004; Fang et al. 2007; Grimes & Kitchen 2007). The idea is that prior exposure enhances the ease with which individuals can process the stimuli in subsequent encounters. This perceptual fluency is an important source of information for individuals. Winkielman et al. (2003) argue that individuals monitor the fluency with which they can extract information from a stimulus. The fluency signal itself is hedonically marked: high perceptual fluency will elicit a positive affective reaction. As Winkielman et al. (2003, p. 195) further explain, individuals by default assume that their fluency-based affective reactions reflect their disposition toward the stimulus. In other words, the same stimulus is evaluated more positively when it can be processed with high as opposed to low fluency. According to this line of reasoning, then, the more frequently a stimulus is presented, the faster individuals can process the stimulus at a later point in time. The reason for the increase in processing fluency lies in implicit memory traces (Schacter 1987; Shapiro 1999). The intriguing point is that those implicit memory traces cannot be consciously retrieved from memory, so that there is no explicit memory for the stimulus. However, the retrieval of those memory traces can take place non-consciously, thus influencing subsequent choice without conscious recollection of the encoding episode. [] As such, it is widely thought that mere exposure effects are primarily driven by automatic, preconscious processes (Grimes & Kitchen 2007, p. 195). In recent years, there has been a flourish of interest in mere exposure effects in advertising research (e.g. Baker 1999; Lee & Labroo 2004; Fang et al. 2007; Grimes & Kitchen 2007; Tom et al. 2007). Even Zajonc (2004) himself has argued that the mere exposure effect must be well known in the advertising industry (p. 202). However, in research on the effects of
481
product placements, there are only some first hints for mere exposure effects. Janiszewski (1993) demonstrated that the mere presentation of a brand in a print ad had a positive impact on the attitude towards the brand, even though viewers did not recall the exposure to it. In an interesting study, Russell (2002) investigated the impact of modality (visual or auditory) and degree of connection between plot and placement on viewers recall and attitudes. The results showed that placements which could not be remembered did indeed exert an influence on viewers attitudes towards the brand. In contrast, placements that were better remembered had no impact on attitudes. As Russell (2002) explains, when placements are the focal point of attention, they become obtrusive and viewers might respond with counter-arguing: The fact that subtly placed visual brands appeared more natural and were therefore not counter-argued advances our knowledge of visual persuasion (Russel 2002, p. 314). The results of Janiszewski (1993) and Russell (2002) suggest that placement recall might be independent of placement evaluation. Taken together, insights from cognitive psychology and advertising research support the notion that mere exposure effects are not mediated by stimulus recognition and recall. However, little is known about mere exposure effects when it comes to product placements. Although Russell (2002) mentions the mere exposure effect as one possible explanation of her results, there are no studies that have put this notion to a direct test. Even more importantly, the placement-specific boundary conditions or moderators for the mere exposure effects are an unresolved issue.
The mere exposure effect for product placements: a moderated phenomenon

While the mere exposure effect is unmediated, this does not mean that there are no moderators at work (Bornstein 1989). As Muller et al. (2005) explain, mediational analyses aim at identifying the intermediary process that leads to an effect on a dependent variable, but moderational analyses attempt to identify [] contextual variables that strengthen and/or change the direction of the relationship between the treatment variable and the dependent variable (p. 852). In his meta-analysis, Bornstein (1989) identifies several moderators for mere exposure effects, such as placement frequency or exposure time. However, research has yielded
482
little knowledge about the moderators that are likely to occur when it comes to product placements. First of all, the studies by Janiszewski (1993) and Russell (2002) investigated singular placements. The most important moderator, the frequency of exposure (Bornstein 1989), has not been manipulated in product placement research so far. Based on the results in psychological mere exposure research, it can be assumed that mere exposure effects are more pronounced when a brand is presented frequently, i.e. more than once. Therefore, a first moderator of product placement mere exposure effects should be placement frequency. Second, there is a crucial difference between the rather simple stimuli presented in psychological mere exposure research and product placements in audiovisual media. In movies or television shows, product placements are embedded in a rather complex visual stimulus and are watched with varying degrees of attention. In contrast, psychological mere exposure research works with unembedded stimuli. Therefore, the effectiveness of product placements highly depends on the context in which placements are shown and on the involvement viewers devote to a media programme. Surprisingly, this aspect has rarely been taken into account in previous research: studies testing the effectiveness of placements mostly work with simple placement contexts, such as slideshows of placements or newspaper articles (Janiszewski 1993; see also Baker 1999). Bearing in mind that fluency-based mere exposure effects occur only when viewers do not consciously elaborate on the occurrence of placements, it can be assumed that placements should be processed with a rather limited amount of attention (Law & Braun 2000; Fang et al. 2007). Also, Grimes and Kitchen (2007) call for low-attention conditions in order to observe mere exposure effects. As stated by Winkielman et al. (2003), it can be expected that fluency effects increase under resource limitations, because these limitations interfere with the integration of additional information about the stimulus. In other words, the less cognitive resources are devoted to the placement itself, the stronger the attitudinal effects (McCarty 2004; Fang et al. 2007; Grimes & Kitchen 2007). More specifically, when cognitive resources are devoted to the movie or television programme itself which means there is a high programme involvement cognitive resources for the placements are rather low. Put simply, if viewers are cognitively engaged with the story, they lack the cognitive capacity to watch out for placements. A study by Schneider and Cornwell (2005)
483
gives clear support for this argument: highly involved players of a racing game showed a poorer placement recall compared to players with low involvement. This line of reasoning is also supported by research on traditional forms of advertising, which demonstrates that higher involvement with the story reduces the cognitive capacity for the commercial (Norris & Coleman 1992). Or, as Grigorivici and Constantin (2004, p. 34) have put it, placing brands in highly arousing, immersive programme context leaves less processing capacities to encoding placements as commercial messages, but rather wrapping users/viewers in the program context itself. Therefore, it can be assumed that a mere exposure effect is more likely to occur when programme involvement is high, because high programme involvement limits processing capacity for the embedded placements, thus impeding deliberative processing of the placement (Winkielman et al. 2003, p. 211; Fang et al. 2007, p. 102; Grimes & Kitchen 2007, p. 210). This should be the second moderator for a mere exposure effect of a product placement. Along with the frequency of placements and programme involvement, another crucial variable can be derived from the research literature (Russell 2002): viewers persuasion knowledge. Friestad and Wright (1995) argue that people can detect the persuasive attempt, they indulge in counter-arguing and, consequently, develop more negative attitudes towards the brand (see also Brehm 1966). Persuasion knowledge is best defined as an activated attitude towards a persuasive attempt. More specifically, it comprises knowledge about actors, themes and outcomes of the persuasion process (Friestad & Wright 1995). In a seminal study, Boush et al. (1994) operationalised persuasion knowledge as a stable attitude that can be activated during persuasion. As Russell (2002) has convincingly argued, persuasion knowledge can impede mere exposure effects: audiences may realise that the placements were placed there to affect their judgements, and they may counter-argue them just as they do traditional advertising messages (p. 307). That means persuasion knowledge is a moderator that impedes a mere exposure effect. When there is persuasion knowledge, implicit memory effects are no longer possible. Persuasion knowledge evolves from experiences with persuasion that may differ among individuals (Boush et al. 1994; Friestad & Wright 1995). Hence, it can be assumed that individuals differ in their degree of having already developed persuasion knowledge about product placements.
484
In fact, people who have developed a high degree of persuasion knowledge during their media socialisation are less likely to be influenced by product placements via the mere exposure effect. In contrast, individuals who have no accessible and applicable persuasion knowledge for product placements are unlikely to employ strategies of counter-arguing. Thus, persuasion knowledge is the third moderator that can be identified based on the product placement literature. Taken together, consistent with the literature previously discussed, it is proposed that there are three moderators of mere exposure effects to occur for product placements: first, the frequency of exposure (Bornstein 1989); second, the involvement in the plot (Winkielman et al. 2003); and, third, viewers persuasion knowledge (Russell 2002). Of course, involvement and persuasion knowledge were not investigated in psychological mere exposure research outside the advertising context. The reason is, simply, that they were not relevant. For the simple stimuli in Zajoncs (1968) study, there is no context in which viewers can get involved. Likewise, syllables or ideographs cannot invoke persuasion knowledge, because they are not meant to persuade. In contrast, placements are meant to persuade, and they are embedded in a programme context that can be watched with varying degrees of involvement.
Hypotheses
Throughout this paper, it has been argued that a frequently presented brand placement can have a positive effect on brand evaluations although viewers do not recall the brand. This is a mere exposure effect. Therefore, the following general hypothesis can be derived: H1: The repeated exposure to a brand placement can lead to an increase in liking for the brand, when people do not remember that they were exposed to a placement (mere exposure effect). However, as should be apparent from the discussion above, this general hypothesis is bound to three conditions, or moderators. First of all, based on the insights of mere exposure research in cognitive psychology, it is suggested that a placement must be presented with a high frequency. Second, frequently presented product placements can gain an attitudinal
485
impact only when viewers have limited cognitive capacity to process the placements (Winkielman et al. 2003; Fang et al. 2007; Grimes & Kitchen 2007). This means that the media programme must be processed with high involvement, leaving little capacity for processing the placement information itself. In this case, placements are not attentively processed. While they nevertheless leave memory traces (Schacter 1987; Shapiro 1999), these traces are not accessible to recall: viewers have in fact seen the placements but will not remember them explicitly. Nevertheless, the appearance of the brand leaves implicit memory traces, which, in turn, lead to a higher fluency. Due to the positive marking of perceptual fluency, viewers should evaluate the brand positively. Third, a crucial variable in any persuasion process is the viewers persuasion knowledge. Persuasion knowledge is not generated while viewing the movie or television programme, it is rather brought to the process based on previous persuasion experiences. People who have developed high degrees of persuasion knowledge during previous contact with placements are more susceptible to persuasion cues. Therefore, they are more likely to counterargue when they view a movie or a television broadcast. Taken together, for a mere exposure effect to occur, placements must be presented with high frequency, viewers must watch the movie or programme with high involvement, and they should have a low amount of persuasion knowledge. Hence, it is argued that a mere exposure effect is possible only when these three moderators are present. Therefore, we can formulate three sub-hypotheses: H1aH1c: For mere exposure effects to occur, brand placements must be presented with high frequency (H1a), viewers must watch the media programme with high involvement (H1b) and have a low degree of persuasion knowledge (H1c). In contrast to a mere exposure effect, the presentation of placements can also lead to a decrease in liking for a brand. As Russell (2002) has argued, when placements become the focal point of attention and when they can be remembered well by viewers, a deterioration of brand attitudes can be the result. This is a reactance effect. Hence, we can formulate the second general hypothesis:
486
H2: The repeated exposure to a brand placement can lead to a decrease in liking for the brand, when people do remember the exposure to a placement (reactance effect). Similar to the mere exposure effect, we argue that there are conditions for a reactance effect. In general, when persuasion knowledge is high, no positive attitude change can be expected at all (Friestad & Wright 1995). However, when there is high persuasion knowledge and low programme involvement, individuals have the maximum amount of resources to elaborate on the persuasion attempt. We assume that persuasion knowledge has the most negative impact on brand evaluations when there is a high cognitive capacity to activate and elaborate persuasion knowledge (for a similar argument, see Russell 2002; Lee & Labroo 2004). Based on that, we assume that the frequent presentation of a placement can even have a negative impact on brand evaluations when persuasion knowledge is high and involvement in the programme is low. This negative effect can be interpreted as a reactance effect (Brehm 1966) which is completely different to a mere exposure effect. This leads us to three sub-hypotheses: H2aH2c: For a reactance effect to occur, brand placements must be presented with high frequency (H2a), viewers must watch the media programme with low involvement (H2b) and have a high degree of persuasion knowledge (H2c). For the other combinations of the independent variables, we can only state some assumptions. Therefore, no hypotheses are formulated. First, we suppose that high programme involvement and high persuasion knowledge, as well as low programme involvement and low persuasion knowledge, have no impact on brand attitudes. Individuals with low involvement and low persuasion knowledge might process placement information; however, as long as they have only a low degree of persuasion knowledge, they will not engage in counter-arguing. Similarly, individuals with high programme involvement and high persuasion knowledge might start counter-arguing, but they are still captured by the narration of the story. In this case, persuasion knowledge is without any consequences. Taken together, we assume that there should be no attitude change in these cells as a function of placement frequency.
487
Method
Design We tested our hypotheses in a 3 2 2 experimental design. Subjects were randomly assigned to the experimental groups. First, the frequency of placement appearance was manipulated: we took a broadcasted audiovisual report from a TV magazine focusing on science topics. In this report, a screen placement for an agricultural company (the brand logo DeLaval, see Figure 1) occurred several times. Similar to psychological mere exposure research, the target brand DeLaval is quite unknown to the general public. The occurrence of the brand placement could easily be varied. We created one version without any placement, one version in which the brand appeared seven times, and one version with 13 appearances of the brand. These frequencies were chosen for three reasons: (1) in order to prevent wear-out effects, there should not be too many placement repetitions (see Laroche et al. 2006); (2) the frequencies were deemed realistic from a practical standpoint creating a report with a higher amount of placements could adversely affect the credibility of the report; (3) the whole design was tested in an extensive pilot experiment (n = 74), which gave clear hints for the expected results. The duration of a placement appearance was 0.5 to 3 seconds (see Figure 1 for screenshots). The content of the report was exactly the same in the three versions. Second, programme involvement was varied by telling one half of the subjects that the story they were going to see was quite important, that they should concentrate on the content of the story, and that a knowledge test would follow (Wirth 2006). The other half of the subjects were told that the story they were about to see was rather uninspiring and not very important to the general public. Persuasion knowledge was not manipulated but measured after the experiment and treated as a quasi-experimental factor. Subjects A total of 115 students participated in the experiment (55% female; age: M = 24.4, SD = 3.35). Subjects were enrolled in graduate and undergraduate courses; most subjects majored in mass communication. No relationship
488
Figure 1: Screenshots (target brand is ringed in white)
489
between students area of study and the results could be observed. They were paid US$10 for their participation. Of course, student samples can be problematic in terms of population validity. However, as Oakes (1972) explained, when the dependent variables do not interact with subject characteristics, student samples can be appropriate. In this study, we examine mere exposure effects. In the research literature, there is no indication that mere exposure effects depend on age, gender or education (Bornstein 1989). Stimulus material The topic of the story was new methods of cattle-breeding. Subjects saw two reports: one warm-up report at the beginning and our target report. Overall, the whole stimulus lasted about 10 minutes. In our target report dependent on experimental condition a brand placement was shown varying in frequency. For instance, there was a placement on perimeter advertising at a cattle exhibition or on the coat of a farmer. These short scenes were cut out for the versions without placements. A first pre-test (n = 27) showed that all versions of the stimulus were perceived as natural, not arousing any suspicion of manipulation. A second pre-test (survey data, n = 70) showed that our subjects did not know that particular brand, and that they evaluated it equally neutral compared to other unknown brands (e.g. Flaco) that were invented especially for that pre-test. In a third pre-test a pilot experiment (n = 74) we tested all measures and the whole procedure. Procedure Subjects were run in groups of six to eight individuals in our advertising lab. They watched the stimulus on a personal computer (equipped with headsets), and worked independently. No disturbance among subjects was observed. As a cover story, they were told that the investigation was about attitudes towards audiovisual science reports. After the stimulus was shown, subjects filled out an online questionnaire. In that questionnaire, we asked for the purpose of the study. As the results indicate, no one guessed the true purpose. Because of that and because of the strict randomisation, we believe that there were no other ways that the respondents
490
could have been influenced during the fieldwork. After filling out the questionnaire, subjects were debriefed, thanked and paid US$10. Measures All items were extensively tested in several pre-tests prior to the fieldwork. Involvement was measured with five items ( = 0.78) assessing the degree to which subjects were captured by the story and how much they concentrated on it (for all measures see the appendix). Persuasion knowledge was measured with five items ( = 0.85). In order to capture the negative preoccupation, we operationalised persuasion knowledge as an activated attitude of how disturbing product placements were perceived (see Friestad & Wright 1995). As persuasion knowledge served as an independent variable, we divided our subjects into two groups by means of median split (Md = 0.26). As the main dependent variable, we measured the attitude towards the brand with a semantic differential scale consisting of five items ( = 0.84). Subjects were shown several brands and were asked how much they intuitively liked those brands. This measure followed the suggestions by Bornsteins (1989) extensive meta-analysis on mere exposure effects. He found that liking-ratings, goodness-ratings and appealing-ratings showed best results in terms of mere exposure effects. It was quite important for us to build on these measures because the mere exposure effect we were looking for demands a spontaneous, unelaborated assessment of brand preference. Brand choice measures were deemed too complicated to implement in our laboratory. Furthermore, similar scales were used in the study by Russell (2002) that were crucial for the derivation of our hypotheses. We also checked the Marketing Scales Handbook by Bruner et al. (2005), which documents several attitudes towards the brand measures that are similar to ours (e.g. Zhang & Schmitt 2001). In a pilot experiment (n = 74), we found a social desirability bias (King & Bruner 2000) in the condition with zero placements. In fact, subjects evaluated the placements quite positively, although no placements were included in that version of the stimulus. Such a bias can emerge when subjects have to evaluate a brand they could not have seen. We controlled for this social desirability bias by partialling out the influence of such false evaluation effects. We computed a difference score between the evaluation of our target brand (DeLaval) and another unknown brand (Flaco),
491
which we call ghost brand. As shown in our pre-test survey (n = 70), the ghost brand is completely equal in familiarity and evaluation compared to DeLaval. The logic is as follows: the ghost brand was not shown in our stimulus, but we asked for the evaluation of that brand in our questionnaire. If a social desirability bias occurs, it should also affect the evaluation of the ghost brand. Hence, we can partial out that social desirability bias by computing a difference score between our target brand and the ghost brand. This difference score is, therefore, an adjusted evaluation of the target brand controlled for a social desirability bias. To capture viewers brand recognition, we measured aided recall: subjects reported to what extent they could remember the brand. Both for the attitude and the recall measure, brands were shown as an icon in the questionnaire. In order to control for possible confounds, we asked participants for their evaluation of the report with a seven-item scale ( = 0.81). Moreover, we tried to avoid measurement effects by asking about our most important dependent variable, brand evaluation, at the beginning of the questionnaire. Only some filler questions and the programme involvement were asked prior to that. Then followed persuasion knowledge and aided recall of several brands.
Results
Manipulation check First of all, there were no differences between the three versions of the stimulus in terms of perceived professionalism, credibility and overall impression of the programme. Second, the involvement manipulation worked quite well: respondents in the high involvement group reported a significant higher programme involvement (M = 0.38, SD = 0.88) than respondents in the low involvement group (M = 0.39, SD = 0.96, t (112) = 4.51, p < 0.01). However, the involvement manipulation had no impact on the attitude towards the brand (t < 1). Third, we checked whether placement frequency or involvement had an impact on viewers persuasion knowledge. This was not the case: neither placement frequency nor viewers involvement influenced persuasion knowledge (Fs < 1). Hence, there is no indication of a confounding among our independent variables.
492
Experimental hypotheses testing We assumed that the mere exposure to a frequently presented brand will have a positive impact on the attitude towards the brand when programme involvement is high, and persuasion knowledge is low. We call this experimental group the mere exposure group. In contrast, when a brand placement is presented with high frequency, and viewers devote low involvement to the report, and have a high degree of persuasion knowledge, the exposure to product placements will have a negative impact on the attitude towards the brand (i.e. reactance group). Therefore, we expected a three-way interaction of placement frequency, programme involvement and persuasion knowledge. As the results indicate, this is the case: there is a significant three-way interaction of the independent variables (F (11, 97) = 3.54, p < 0.01, 2 = 0.29). No main effects of placement frequency, programme involvement and persuasion knowledge on the attitude towards the brand emerged (Fs < 1). Put differently, we have a disordinal interaction. There is no main effect of placement frequency on attitude towards the brand because the attitude towards the brand can deteriorate or increase depending on the other conditions. Likewise, there are no main effects of persuasion knowledge and involvement. Persuasion knowledge per se does not lead to a worsening of brand attitudes. And high programme involvement per se does not lead to better brand attitudes. In other words, all three independent variables interact in their impact, i.e. they can only exert an effect in their combination. As can be seen in Table 1, subjects who were exposed to the broadcast with 13 placement appearances, watched the programme with high involvement, and had no persuasion knowledge (i.e. the mere exposure group), have the highest scores on the attitude towards the brand index (M = 1.23, SD = 0.31). To further elaborate on this finding, we tested whether the mean of the mere exposure group is significantly different from the mean of all other cells. In fact, all 11 tests are significant (p < 0.05) in the expected direction. However, this positive attitude towards the brand can be interpreted as a mere exposure effect only when subjects fail to remember the brand explicitly. Indeed, subjects in the mere exposure group (13 placements, high programme involvement, and low persuasion knowledge), have extremely low scores on the memory towards the brand index (M = 1.25,
493
Table 1: Evaluation of the brand and brand recognition as a function of placement frequency, involvement and persuasion knowledge (standard deviations in brackets)
Involvement High (n = 58) Persuasion knowledge high (n = 29) Attitude towards the brand: 0 placements (n = 33) 7 placements (n = 41) 13 placements (n = 41) Recognition: 0 placements (n = 33) 7 placements (n = 41) 13 placements (n = 41) Persuasion knowledge low (n = 29) Low (n = 57) Persuasion knowledge high (n = 30) Persuasion knowledge low (n = 27)
0.39 (0.38) 0.26 (0.30) 0.13 (0.35)
0.12 (0.35) 0.13 (0.37) 1.23 (0.31)
0.22 (0.40) 0.80 (0.35) 0.67 (0.29)
0.34 (0.35) 0.25 (0.31) 0.05 (0.44)
1.14 (0.38) 3.00 (1.62) 3.23 (1.42)
1.40 (0.52) 1.33 (0.71) 1.25 (0.46)
1.50 (0.58) 1.71 (0.76) 3.31 (1.82)
1.17 (3.89) 1.14 (0.54) 1.75 (0.96)
Note: Values for brand evaluation are z-standardised means; brand recognition is measured on a 5-point scale ranging from 1 do not remember the brand to 5 do remember the brand.
SD = 0.46). In other words, viewers with the most positive evaluation of the brand show a very low brand recall. Because explicit brand recall is not responsible for the bettering of brand attitudes, an implicit mere exposure effect must be at work here. In order to prove this, we have calculated a correlation between brand recall and brand attitudes for this mere exposure group. As expected, we can find no significant correlation between brand recall and attitude towards the brand (r = 0.15, p = 0.72). Thus, explicit brand recall has no impact on the attitude towards the brand for this group. To put it in the words of Zajonc (2004), the mere exposure effect is unmediated by recall. Taken together, our results lend strong support for our first hypothesis, and the three sub-hypotheses 1a, 1b and 1c. Furthermore, we hypothesised that the attitude towards the brand will deteriorate when there is a high frequency of placements, low involvement and high persuasion knowledge (reactance group). This effect is no mere exposure effect, but a reactance effect, as the memory for the brand
494
should be very high. As expected, subjects who were exposed to 13 placements, and had a low involvement but high persuasion knowledge, showed a very negative attitude towards the brand (M = 0.67, SD = 0.29). This effect can even be found for the low involvement/high persuasion knowledge group that was exposed to seven placements (M = 0.80, SD = 0.35). In fact, there is a significant decrease regarding the attitude towards the brand compared to the conditions with zero placements. Even more importantly, subjects in the reactance group had very high scores on the memory towards the brand index (M = 3.31, SD = 1.82). Obviously, the high persuasion knowledge and the low programme involvement lead to a high placement recall. This impedes a mere exposure effect but leads to reactance. In other words, in contrast to the mere exposure effect described above, placement recall has an impact on brand attitudes for the reactance group. In order to prove this, we have again calculated the correlation between recall and attitude towards the brand, just as above. As expected, this time we can find a significant correlation between both variables (r = 0.58, p < 0.05). Overall, this result supports the second hypothesis, and the three sub-hypotheses 2a, 2b and 2c. To reiterate, it is the specific combination of the independent variables that can either lead to more positive brand attitudes (mere exposure effect) or to more negative brand attitudes (reactance effect). In the first case, explicit brand recall is not responsible for the effect the mere exposure effect is not mediated by recall. In the second case, however, a high brand recall can support the worsening of brand attitudes. Regarding the other experimental groups, as expected, subjects with low programme involvement and low persuasion knowledge show no differences in their attitudes at all. For those participants, the frequency of product placements has no impact on the attitude towards the brand (zero placements: M = 0.34, SD = 0.35; seven placements, M = 0.25, SD = 0.31; 13 placements, M = 0.05, SD = 0.44). Put differently, when there is low involvement and low persuasion knowledge, no mere exposure effect can be expected, because there is no high involvement. However, no reactance effect can be expected either, because these viewers are lacking the necessary persuasion knowledge. Likewise, we find no effect of placement frequency for subjects with high programme involvement and high persuasion knowledge (zero placements: M = 0.39, SD = 0.38; seven placements, M = 0.26, SD = 0.30; 13 placements, M = 0.13, SD = 0.35). For this
495
group, a possible mere exposure effect is suppressed by persuasion knowledge. Furthermore, a reactance effect is unlikely, because reactance requires cognitive resources that are bound by high involvement. In other words, persuasion knowledge and involvement neutralise each other in their impact on viewers attitudes.
Discussion
The findings of this study indicate that the frequent and unreinforced presentation of a previously unknown brand can have an effect on viewers attitudes, even if memory for the placement is low. This lends support to the idea that the impact of product placements on viewers attitudes can be interpreted as a mere exposure effect. More precisely, this effect can be explained in terms of processing fluency (Reber et al. 2004). Although people cannot remember the stimulus, repeated exposure leads to an increase in processing fluency. This fluency signal itself is hedonically marked: high fluency will elicit a positive affective reaction. It is important to stress that the mere exposure effect is not mediated by brand recall, because there is no significant relationship between brand recall and the bettering of the attitude towards the brand. In this study, we have identified three important moderators for the mere exposure effect of product placements: a mere exposure effect is likely to occur when a previously unknown brand is presented with high frequency, viewers watch the programme with high involvement, and have a rather low persuasion knowledge. Besides the mere exposure effect, we can also formulate the conditions for a reactance effect: when a brand placement occurs with high frequency, the audience programme involvement is low but persuasion knowledge is high, the attitude towards the brand can be negatively affected compared to a condition where no placements are shown. One possible explanation for this reactance effect is that viewers possess sufficient cognitive resources to elaborate on their persuasion knowledge. In other words, when watching a media programme with low involvement and feeling disturbed by product placements, people have the cognitive capacity to critically scrutinise the placements. Therefore, when subjects figure out that the placements are intended to persuade, they try to resist this persuasion attempt and evaluate the brand more negatively. Although other explanations for this effect
496
cannot be fully ruled out by the present data, this reasoning is in line with the persuasion knowledge model of Friestad and Wright (1995) which explicitly points to reactance effects. One could object that, in terms of priming, we have studied a word rather than a brand. However, it is important to stress that the placement we have used is a real brand placement. Our stimulus was not an artificial slideshow with invented placements or just words, but an authentic placement broadcasted in a TV magazine. In short, the placements were real. Moreover, subjects were asked about their evaluation and aided recall with respect to a brand, not just a word. Even in theoretical terms, the placement we have used is a combination of an icon with a brand name, which perfectly matches the definition of a screen placement (Gupta & Lord 1998; Russell 1998). Bearing in mind that we have used a real product placement presented in an authentic broadcast, there are of course differences to classical mere exposure studies in cognitive psychology. Compared to these studies, the time of placement exposure is quite high, lasting for 0.5 to 3 seconds. However, we have to keep in mind that the placements are embedded in a complex audiovisual field. Thus, individuals are not really directly exposed to those placements. In contrast, the placements are popping up at some place in moving pictures. And again, these are realistic exposure times in real-world contexts, therefore enhancing the external validity of our results. We have studied a real, but comparatively unknown brand. It is known that mere exposure effects are stronger for unknown stimuli (Bornstein 1989). For mere exposure effects to occur, attitude strength might be the most important predictor. When individuals hold strong attitudes towards a brand, mere exposure will certainly exert no influence. However, in the global advertising world, consumers are literally surrounded by thousands of brands without having a strong attitude for them. For such brands, mere exposure effects are a realistic scenario. In considering the broader implications of our findings, our results point to the importance of examining how product placements are processed by viewers. Among practitioners and marketing researchers, there is a widespread agreement that product placements must catch as much attention as possible. However, our findings contradict the idea that placements must be remembered in order to have an attitudinal impact: in contrast, our experiment demonstrated that well-remembered product placements
497
can even lead to reactance effects. Moreover, our findings suggest that effects on memory can be independent of effects on attitudes: placements can have an effect on attitudes although there is no effect on explicit memory. This hidden impact of product placements is quite important because it contributes to our understanding of some otherwise paradoxical phenomena. Previous research on the impact of product placements has consistently pointed to a rather serious problem. On the one hand, product placements are attractive for practitioners because viewers do not recognise the persuasive attempt. Therefore, compared to classical forms of advertising (such as commercials), placements are accepted among viewers and reactance towards advertising is less likely (dAstous & Chartier 2000). On the other hand, however, the subtle and non-persuasive way of a placing a brand runs the risk of its being overlooked by viewers. In fact, our results seem to indicate that this risk is no drawback; in contrast, it can even be an advantage: when placements are not recognised, they can nevertheless have an impact on viewers attitudes. In other words, effective placements are more than meets the eye. Moreover, the results of this study are also quite important both for practitioners and marketers. For instance, although we do not yet know how strong and enduring such mere exposure effects are, it can be assumed that mere exposure effects can be quite important when new brands are introduced to the market. By subtly placing brands in a highly involving media programme, positive attitudes towards the brand can be created in a first step. In a second step, then, this positive attitudinal base can be strengthened by more obtrusive forms of advertising. So far, we have reached some preliminary conclusions about when mere exposure effects and reactance effects are likely to occur; however, we are far from having a comprehensive analysis of all antecedents and consequences of such effects. First of all, a replication of this study with a larger sample could further test the extent to which our findings can be generalised. Second, further research should also explore the relationship between programme involvement and persuasion knowledge. In this study, we argued that persuasion knowledge has the most negative impact when viewers can elaborate consciously on the placement information (i.e. low programme involvement). This assumption is confirmed by the high brand recognition scores of low-involvement viewers. However, it may be equally possible that, once persuasion knowledge is activated, viewers
498
disengage from the media programme. In order to examine the relationship between involvement and persuasion knowledge, future studies will have to manipulate the two variables independently. A further shortcoming of our study refers to the magic number of placement appearances. In psychological mere exposure research, the optimal frequency (i.e. to produce the most positive attitudes) varies from stimulus to stimulus. Therefore, we may expect that an absolute magic number does not exist. Nevertheless, future research should test the mere exposure effect with different placement frequencies and different media contexts. There is probably an exposure level (i.e. extremely high placement frequency) at which no mere exposure effects can be found. As Sawyer (2006) states, the effectiveness of the frequency of placements might have an inverted U-curve effect. The examination of this issue might be a fertile topic for upcoming research on product placements.
Appendix: Test measures

All scales were administered with a five-point scale ranging from 1 not agree at all to 5 fully agree. Involvement: I thought about the topic of this story extensively. I concentrated on the story. I was captured by the story. I thought about things that had to do with the story. I was absorbed by the things that were shown in the story. Persuasion knowledge: I think some persons will feel disturbed by the placements that were shown in the programme. The brand icons in the story should have been cut out. There was too much advertising in the TV programme. I found the brand appearances in the story annoying. Obviously, companies were given the opportunity to place their brands in the programme.
499
Attitudes towards the brand (five-point semantic differential): In my opinion, this brand seems to be friendly unfriendly positive negative appealing not appealing interesting uninteresting attractive unattractive Attitude towards the report: In my opinion the report was good bad interesting uninteresting credible not credible professional unprofessional informative uninformative exciting boring realistic unrealistic
References
Babin, L.A. & Carder, S.T. (1996) Viewers recognition of brands placed within a film. International Journal of Advertising, 15(2), pp. 140151. Baker, W.E. (1999) When can affective conditioning and mere exposure directly influence brand choice? Journal of Advertising, 28 (Winter), pp. 3146. Bornstein, R.F. (1989) Exposure and affect: overview and meta-analysis of research, 19681987. Psychological Bulletin, 106(2), pp. 265289. Bornstein, R.F. & DAgostino, P.R. (1992) Stimulus recognition and the mere exposure effect. Journal of Personality and Social Psychology, 63(4), pp. 545552. Boush, D.M., Friestad, M. & Rose, G.M. (1994) Adolescent scepticism toward TV advertising and knowledge of advertiser tactics. Journal of Consumer Research, 21 (June), pp. 165175. Brehm, J.W. (1966) A Theory of Psychological Reactance. New York: Academic Press. Brennan, I. & Babin, L.A. (2004) Brand placement and recognition: the influence of presentation mode and brand familiarity. Journal of Promotion Management, 10(1/2), pp. 185202. Brennan, I., Dubas, K.M. & Babin, L.A. (1999) The effects of placement type and exposure time on product placement recognition. International Journal of Advertising, 18(3), pp. 323338. Bruner, G.C., Hensel, P.J. & James, K.E. (2005) Marketing Scales Handbook. Mason, OH: Thomson.
500
dAstous, A. & Chartier, F. (2000) A study of factors affecting consumer evaluations and memory of product placements in movies. Journal of Current Issues and Research in Advertising, 22 (Fall), pp. 3140. Fang, X., Singh, S. & Ahluwalia, R. (2007) An examination of different explanations for the mere exposure effect. Journal of Consumer Research, 34 (June), pp. 97103. Friestad, M. & Wright, P. (1995) Persuasion knowledge: lay peoples and researchers beliefs about the psychology of advertising. Journal of Consumer Research, 22 (June), pp. 6274. Grigorovoci, D. & Constantin, D. (2004) ?Experiencing interactive advertising beyond rich media: impacts of ad type and presence on brand effectiveness in 3D gaming immersive virtual environments. Journal of Interactive Advertising, 5(1), available online at http://www.jiad.org/vol5/no1/index.htm. Grimes, A. & Kitchen, P.J. (2007) Researching mere exposure effects to advertising. Theoretical foundations and methodological implications. International Journal of Market Research, 49(2), pp. 191219. Gupta, P.B. & Lord, K.R. (1998) Product placement in movies: the effect of prominence and mode on audience recall. Journal of Current Issues and Research in Advertising, 20 (Spring), pp. 4759. Janiszewski, C. (1993) Pre-attentive mere exposure effects. Journal of Consumer Research, 20 (December), pp. 376392. Karrh, J.A. (1998) Brand placement: a review. Journal of Current Issues and Research in Advertising, 20 (Fall), pp. 3149. Karrh, J.A., McKee, K.B. & Pardun, C.J. (2003) Practitioners evolving views on product placement effectiveness. Journal of Advertising Research, 43 (June), pp. 138149. King, M.F. & Bruner, G.C. (2000) Social desirability bias: a neglected aspect of validity testing. Psychology & Marketing, 17 (February), pp. 79103. Kunst-Wilson, W.R. & Zajonc, R.B. (1980) Affective discrimination of stimuli that cannot be recognized. Science, 207 (February), pp. 557558. Laroche, M., Cleveland, M. & Maravelakis, I. (2006) Competitive advertising interference and ad repetition effects: comparing high-share and low-share brands. International Journal of Advertising, 25(3), pp. 271307. Law, S. & Braun, K.A. (2000) Ill have what shes having: gauging the impact of product placements on viewers. Psychology & Marketing, 17 (December), pp. 10591075. Law, S. & Braun-LaTour, K.A. (2004) Product placements: how to measure their impact, in L.J. Shrum (ed.) The Psychology of Entertainment Media: Blurring the Lines between Entertainment and Persuasion. Mahwah, NJ: Lawrence Erlbaum Associates, pp. 6378. Lee, A.Y. & Labroo, A.A. (2004) The effect of conceptual and perceptual fluency on brand evaluation. Journal of Marketing Research, XLI (May), pp. 151165. McCarty, J.A. (2004) Product placement: the nature of the practice and potential avenues of inquiry, in L.J. Shrum (ed.) The Psychology of Entertainment Media: Blurring the Lines between Entertainment and Persuasion. Mahwah, NJ: Lawrence Erlbaum, pp. 4563.
501
Moreland, R.L. & Zajonc, R.B. (1977) Is stimulus recognition a necessary condition for the occurrence of exposure effects? Journal of Personality and Social Psychology, 35 (April), pp. 191199. Muller, D., Judd, C.M. & Yzerbyt, V.Y. (2005) When moderation is mediated and mediation is moderated. Journal of Personality and Social Psychology, 89(6), pp. 852863. Norris, C.E. & Colman, A.M. (1992) Context effects on recall and recognition of magazine advertisements. Journal of Advertising, 21 (September), pp. 3746. Oakes, W. (1972) External validity and the use of real people as subjects. American Psychologist, 27 (October), pp. 959962. Reber, R., Schwarz, N. & Winkielman, P. (2004) Processing fluency and aesthetic pleasure: is beauty in the perceivers processing experience? Personality and Social Psychology Review, 8(4), pp. 364382. Russell, C.A. (1998) Toward a framework of product placement: theoretical propositions. Advances in Consumer Research, 25(1), pp. 357362. Russell, C.A. (2002) Investigating the effectiveness of product placements in television shows: the role of modality and plot connection congruence on brand memory and attitude. Journal of Consumer Research, 29 (December), pp. 306318. Sawyer, A.G. (2006) Possible psychological processes underlying the effectiveness of brand placements. International Journal of Advertising, 25(1), pp. 110112. Schacter, D.L. (1987) Implicit memory: history and current status. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13(3), pp. 501518. Schneider, L.-P. & Cornwell, T.B. (2005) Cashing in on crashes via brand placement in computer games. International Journal of Advertising, 24(3), pp. 321343. Shapiro, S. (1999) When an ads influence is beyond our conscious control: perceptual and conceptual fluency effects caused by incidental ad exposure. Journal of Consumer Research, 26 (June), pp. 1636. Tiwsakul, R., Hackley, C. & Szmigin, I. (2005) Explicit, non-integrated product placement in British television programmes. International Journal of Advertising, 24(1), pp. 95111. Tom, G., Nelson, C., Srzentic, T. & King, R. (2007) Mere exposure and the endowment effect of consumer decision making. The Journal of Psychology, 141(2), pp. 117125. Winkielman, P., Schwarz, N., Fazendeiro, T.A. & Reber, R. (2003) The hedonic marking of processing fluency: implications for evaluative judgement, in J. Musch & K.C. Klauer (eds) The Psychology of Evaluation. Affective Processes in Cognition and Emotion. Mahwah, NJ: Lawrence Erlbaum Associates, pp. 189217. Wirth, W. (2006) Involvement, in S. Bryant & P. Vorderer (eds) The Psychology of Entertainment. Mahwah, NJ: Lawrence Erlbaum Associates, pp. 199213. Zajonc, R.B. (1968) Attitudinal effects of mere exposure. Journal of Personality and Social Psychology Monographs, 9(2, Pt 2), pp. 127. Zajonc, R.B. (2001) Mere exposure: a gateway to the subliminal. Current Directions in Psychological Science, 10(6), pp. 225228. Zajonc, R.B. (2004) Exposure effects. An unmediated phenomenon, in A.S.R. Manstead, N. Frijda & A. Fischer (eds) Feelings and Emotions: The Amsterdam Symposium. Cambridge, UK: Cambridge University Press, pp. 194203. Zhang, S. & Schmitt, B.H. (2001) Creating local brands in multilingual international markets. Journal of Marketing Research, 38 (August), pp. 313325.
502
About the authors

Jrg Matthes (PhD University of Zurich, Switzerland) is a post-doctoral at the Institute of Mass Communication and Media Research, University of Zurich, Switzerland. His research focuses on advertising effects, the process of public opinion formation, media effects, trust in news media, and empirical methods. Christian Schemer (MA University of Mainz, Germany) is a PhD student at the Institute of Mass Communication and Media Research, University of Zurich, Switzerland. His research focuses on advertising effects, the process of public opinion formation, media effects, media and emotions, and empirical methods. Werner Wirth (PhD University of Leipzig, Germany) is professor for communication and empirical media research at the Institute of Mass Communication and Media Research, University of Zurich, Switzerland. His research focuses on advertising effects, media effects, entertainment, media and emotions, interactive media, and empirical methods. Address correspondence to: Jrg Matthes, Institute of Mass Communication and Media Research, University of Zurich, Andreasstrasse 15, 8050 Zurich, Switzerland. Email: j.matthes@ipmz.uzh.ch
503
Harmonised adspend data for 22 countries
European Advertising & Media Forecast

In the quarterly reports... - Authoritative forecasts of adspend in the major European economies, plus Japan, Canada and the USA - Harmonised comparable adspend data for newspapers, magazines, TV, radio, outdoor, cinema and internet. - Market overviews and expert analysis providing an informed picture of the latest developments - Features such as review of EU legislation affecting the ad industry - Focus on emerging markets every issue adspend and analysis
Produced in association with
Monthly email updates, four quarterly forecast reports and two bi-annual databooks:
In the databooks... - Historical adspend data by medium - Data presented in local currency, US$, Euros and Purchasing Power Parities - Adspend per capita and distribution of adspend by medium - Summaries of harmonised adspend trends by market and by medium - Detailed economic and demographic data In the monthly updates... - Snapshot of all the latest adspend data and forecasts
The European Advertising & Media Forecast gives you all you need for a complete picture of Europes advertising markets
Order online now at www.warc.com/store

The European Advertising & Media Forecast is available on annual subscription, price 785 for AA members (1,465 for non-members) from WARC, Farm Road, Henley-on-Thames, Oxon RG9 1EJ. Tel: 01491 411000 Fax: 01491 418600 Email: info@warc.com
The influence of message format on the effectiveness of print advertisements for tourism destinations
Alain Decrop
University of Namur (FUNDP) CeRCLe (Centre for Research on Consumption and Leisure)
In a context of overabundant ads and saturated consumers, message format is crucial in developing effective advertising campaigns. In this study, four major format components of print advertisements are considered: picture, logo, text and headline. The goal is to investigate the effectiveness of each of these components in triggering a response by the target audience. Three types of response are taken into account: knowledge (information), liking (attraction) and behavioural intention. Four series of hypotheses related to the influence of message format on the effectiveness of print ads are tested for an urban tourism destination through an experimental research design. Findings show that picture and text are the prevailing elements, while logo and headline are of marginal importance. Pictures are especially effective in attracting the consumer and arousing a behavioural intention, whereas text is most powerful in conveying information.
Introduction and problem definition

Designing the message is a major step in developing effective advertising campaigns. It is usually considered once the target audience has been identified and the desired communication objectives have been defined. Four major decisions are involved in designing the advertising message: (1) what to say (content), (2) how to say it logically (structure), (3) how to say it symbolically (format), and (4) who should say it (source). In this paper, the focus is on the third decision, i.e. message format. The format of any advertisement is made up of numerous components. The key elements in print advertising are the headline, visual, subheads, body copy,
505
captions, boxes and panels, slogans, logotypes, seals and signatures (Bove & Arens 1989; Belch & Belch 2007). Each time copywriters create a new ad, they have to make decisions on the mix of these elements: they may be moved, enlarged or reduced, changed or eliminated, until a satisfying design is reached. In this study, four key format components of print advertisements are considered, i.e. the picture (visual), logo (signature), text (body copy) and slogan (headline). A series of studies have shown that most readers look first at the visual, then read the headline, and finally, in the optimum case, read the body copy (Ogilvy 1985). As far as tourism ads are concerned, Urbain (1983) found that sloganphotographtext is the most typical format. The position of the signature (e.g. brand name) is more variable. The aim of this paper is not to investigate the order but rather the effectiveness of each element in triggering appropriate responses by the target audience. Different response levels may be considered when assessing advertising effectiveness. Over the past two decades, measures of effectiveness have been based on recall and recognition tests (e.g. Starch Readership Scores). However, Hanssens and Weitz (1980) have argued that additional ad hoc measures related to purchasing behaviour (i.e. readership and inquiry generation) are needed. It is now well accepted that measures of advertising effectiveness should be connected with conceptual frameworks: advertising objectives are best set if they are based on some underlying theory or model that expresses the laws or principles by which advertising works (Davidson 1994, p. 539). Such general hierarchy-ofeffects models (e.g. Lavidge & Steiner 1961; for a review, see Barry & Howard 1990) have been proposed in the literature to help in determining the communication objectives and in measuring the extent to which they have been reached afterwards. Those models assume that the consumer passes through different stages when responding to marketing stimuli before purchasing a product or a brand. More precisely, the advertiser may seek to produce cognitive (e.g. attention, awareness, knowledge), affective (e.g. interest, liking, conviction) or conative (e.g. intention, trial, purchase) audience responses. In his book about the language of tourism, Dann (1996) acknowledges that most tourism advertising follows the AIDA pattern, capturing Attention, maintaining Interest, creating Desire and getting Action. Johnson and Messmer (1991) have used such a hierarchical structure in order to investigate the effect of advertising on different stages
506
PRINT ADS FOR TOURISM DESTINATIONS: MESSAGE FORMAT
in vacation destination choice. However, the authors limit their scope to two stages, namely inquiry generation and visitation. They regret that few researchers have attempted to directly model the effect of advertising at stages of a hierarchy; the great majority of studies have concentrated on a single dependent variable (p. 18). The current study tries to avoid this limitation by considering three response levels, i.e. knowledge (the ads informational value), liking (attraction value), and intention (behavioural value). In summary, this research aims at contributing to extant literature regarding three aspects: on the one hand, it extends the classical examination of textpicture connections by investigating the role of two additional elements of print message format, i.e. logo and headline. On the other hand, the study measures three response levels referring to the cognitiveaffectiveconative sequence featured in most hierarchy-of-effects models. Finally, it develops an original methodological approach, i.e. a fullprofile conjoint task, to examine the effectiveness of each format element in triggering different types of readers responses. At a more practical level, this study was undertaken to help design the next promotional campaign of an urban tourism destination, namely Namur, which is a medium-sized Belgian city of 100,000 inhabitants and a popular tourism destination for one-day excursions and short trips. Known as the Ardennes gate, Namur offers a mix of natural and cultural attractions (confluence of two major rivers, walking trails through parks and woods, the biggest citadel of Europe built by Vauban in the 17th century, a range of museums, etc.) and recreational venues (large pedestrianised shopping district, casino, pubs and restaurants). Assessing the effectiveness of advertising is of paramount importance to businesses in respect of the huge amounts of money that are spent on it. While it has been neglected for a long time, advertising now plays a vital role in the marketing of tourism destinations. According to the World Tourism Organization, the 2004 marketing budgets of countries such as Malaysia, Australia, Spain, South Africa and Canada amounted to 89.1, 80.2, 67.6, 51.5 and 50.7 million USD respectively. Belgium has to satisfy itself with a promotional budget of around 4 million USD. In the US, advertising accounts for 44% of tourism marketing state budgets. In 20042005, US states planned to spend a total of 602.7 million USD for development and promotion, Hawaii taking the lead with a budget of 69 million USD (Travel Industry Association 2005).
507
Well-executed advertising evaluation research helps managers spend money more effectively, as identifying ways to make even small improvements in communications impact or preventing the repetition of mistakes can lead in the long run to substantial gains in sales and profits (Davidson 1994, p. 538). Optimising the message format is one such improvement. Although much attention has focused on advertising effectiveness in general, it is surprising that little research has centred on the issue of message format. This shortcoming is particularly relevant to the tourism industry: destination managers often find it hard to set up marketing campaigns and to design effective advertisements. Moreover, Davidson (1994) notices that while a number of journal articles have been written on assessing advertising, almost nothing has been published on assessing tourism advertising (p. 543).
Literature review
A few format studies have been carried out during the two last decades for the major advertising media, such as radio (Sullivan 1990), internet (Gretzel et al. 2000; Burns & Lutz 2006), mobile telecommunication (Rau & Chen 2006), and of course print media (see below). Research effort has focused on the verbal content of the ad, although the importance of nonverbal communication has increasingly been acknowledged (Woollacott 1982; Edell & Staelin 1983; Hecker & Stewart 1988). The rise of image advertising is, to a large extent, responsible for this increased interest in non-verbal communication (MacInnis & Price 1987; Stewart et al. 1987; Unnava & Burnkrant 1991). Regarding the format of print ads, most studies focused on the comparison between the two prevailing elements, i.e. text and pictures. A first series of studies has shown that pictures-only proves to be more effective than text-only, especially in being recalled. Based on Starch Readership Scores, Hanssens and Weitz (1980) found that pictures are helpful in improving the effectiveness of industrial print advertisements while this is not the case with text. According to Mitchell and Olson (1981), advertisements designed with pictures have a stronger effect in influencing attitudes than do ads containing text only. For Edell and Staelin (1983), ad recall is stronger when supported by pictures than when purely text-based because processing of pictorial and lexical information is carried out by
508
different cognitive processes, which may result in different amounts of learning. The same positive effect of pictures on recall was highlighted by Unnava and Burnkrant (1991). Miniard et al. (1991) and Kroeber-Riel (1993) suggested the inclusion of involvement as a mediating variable on this effect. The latter authors found that pictures play a significant role in gaining readers attention in advertisements for high-involvement products, such as a tourism destination, despite increased reading of written information. Mitchell (1986) confirmed the effectiveness of the visual component of ads in triggering both cognitive (i.e. the formation of product attribute beliefs) and affective (i.e. the attitude towards the advertisement) responses. Finally, Tscheulin and Helmig (1998) showed that pictures are more effective than text in hospital print advertisements; the type of pictures has both a positive effect on the attractiveness of the ad and on its perceived informational value, whereas the type of text impacts only on the ads informational value. Other authors have demonstrated that advertising effectiveness is stronger when picture and text are combined. Morelli (1970) was the first to support this: the use of pictorial material increases the memorisation effect when pictures are used to illustrate the text. In 1977, Lutz and Lutz reported a stronger recall effect when the text is supported by interactive and non-redundant pictures. That conclusion was confirmed in later studies by Edell and Staelin (1983) and Houston et al. (1987). Finally, Jaeger and MacFie (2001) have shown that ads communicating both pictorial and textual information resulted in different expectations than those raised by text-only information. They further showed that pictures had a greater influence on consumers with a high need for cognition (NFC) than those with a low NFC. In addition to these two major research streams, more specific studies have highlighted the influence of the pictures size (Rossiter & Percy 1983; Percy 1989; Tscheulin & Helmig 1998) and the increased effectiveness resulting from an interactive pictureslogan relationship (Tscheulin & Helmig 1998). A few tourism researchers have also been interested in issues related to message format and print media. Dann (1996) has pointed out that almost without exception photographs are used in tandem with verbal messages in tourism ads and brochures. MacKay and Fesenmaier (1997) focused on the visual/pictorial component of advertising in shaping destination images and expectations. Olson et al. (1986) found that the visual content
509
of ads affects the perception of vacation experiences through the association of certain types of pictures with certain types of experiences (e.g. natural scenery elicits the recollection of a romantic vacation). More recently, MacKay and Smith (2006) found that advertising format is a determinant of elaborative memory, while age is not. In two experiments, they examined age-related differences in memory for ads using text and picture formats. They showed that younger adults recall more of the content of a written description of a destination than do older participants. In contrast, younger and older adults appear to process and encode information from pictures of destinations in a similar manner. In spite of these few papers, format studies are rare in recent tourism advertising research, which focuses on measuring the effectiveness of destination advertising, especially through conversion studies (Woodside & Ronkainen 1994; Woodside & Dubelaar 2003) or tracking studies (see Siegel & Ziff-Levine 1994; McWilliams & Crompton 1997). MacKay and Smith (2006) diagnose that there are still many outstanding issues to be resolved in the area of age, format, and memory for tourism advertising (p. 20).
Research hypotheses
This research aims to test which orientation would be more effective for tourism advertising that is, either a text-oriented or a picture-oriented format. In addition to these two major elements, the logo and the headline are also considered. Three measures of effectiveness are used: knowledge, liking and intention. The hypotheses discussed below involve these four elements of the message format of print ads, namely visual, logotype, text and headline (independent variables) and three levels of response (dependent variables); four series of hypotheses are presented and each series comprises three sub-hypotheses. Visual The first series of hypotheses is concerned with the visual. Pictures are a cornerstone in print advertising as it has been shown that most readers of advertisements look first at the picture before considering other elements (Ogilvy 1985). Pictures and photographs are of paramount importance to tourism destinations. Dilley (1986) has shown that up to 75% of tourism
510
brochure material is taken up by pictures and that the influence of a promotional picture varies directly with its size. Pictures are widely used because they help to induce imagery (MacInnis & Price 1987), to develop knowledge and expectations (MacKay & Smith 2006) and the desire for image verification (Okoroafo 1989). The visual contributes strongly to successful ads by accomplishing three major tasks. First, the visual conveys an abundance of information as it assists with identifying the subject of the ad, clarifying the claims made by the copy and emphasising unique features of the product: adding an illustration to a textual advertising message means communicating additional information (Jaeger & MacFie 2001, p. 190). Second, it is meant to capture the attention of the reader and to arouse his/her interest in other elements of the ad (Belch & Belch 2007). Third, the visual helps to generate an intention to buy the advertised product/brand by filtering prospects, creating a favourable impression of the product and enhancing the trustworthiness of claims made in the copy (Wilmshurst & Mackay 1999). These three functions may be translated into the following hypotheses: H1a: H1b: H1c: The ads informational value is increased when a picture is present compared with when there is no picture. The ads attraction value is increased when a picture is present compared with when there is no picture. The intention effect of the ad is increased when a picture is present compared with when there is no picture.
Logotypes Logotypes (logos) are special designs of the advertisers company name or product name; they are like signatures and trademarks because they provide individuality and immediate recognition in advertising, packaging, point of purchase, and other promotional efforts (Govoni 2004, pp. 117 118). As they help readers recognise familiar products or brands and create identity, they raise attention towards the ad/product (Bove & Arens 1989). However, as they are short and have very specific content (i.e. the brand or company name), they are of limited informational value. Finally, logos usually do not bear any behavioural intention (Belch & Belch 2007). Therefore, the following hypotheses are proposed:
511
H2a: H2b: H2c:
The ads informational value is not increased when a logo is present compared with when it is absent. The ads attraction value is increased when a logo is present compared with when it is absent. The intention effect of the ad is not increased when a logo is present compared with when it is absent.
The third series of hypotheses is concerned with the influence of text or body copy on the effectiveness of destination ads. The text includes all the words used to tell the complete sales story. It has to be clear, simple, direct, appropriately expressed and relevant (Wilmshurst & Mackay 1999) and must be connected with the headline and/or the visual. According to Bove and Arens (1989), body copy is where the sale is closed (p. 264). It must explain how the advertised product/brand satisfies the consumers needs and emphasise the consumers benefit(s) in order to create an intention to buy. To this end, the major goal of the copy is both to inform about the products characteristics and advantages and to provide the greatest sales appeal for the idea that is presented. Therefore, we expect that the amount of text will affect the informational value and the intention effect of the ad, with a longer text being more effective than a shorter text or no text at all. In contrast, the amount of text is not supposed to attract the readers attention. Generally, longer text even proves to have detrimental effects on interest (Belch & Belch 2007). The third set of hypotheses reads therefore as follows: H3a: H3b: H3c: The ads informational value is increased by the amount of text: the longer the text, the higher the informational value. The ads attraction value is not increased by the amount of text. The intention effect of the ad is increased by the amount of text: the longer the text, the stronger the intention.
Headlines The term headline refers to the words that are used in the leading position of an ad. These words are meant to attract attention and to be read first. Research has shown that the headline is generally the first element people look at in a print ad, followed by the visual (Belch & Belch 2007). In a
512
tourism context, Mayo and Jarvis (1981) assert that 66% of people notice the headline and illustration of an ad, while only 15% read the body copy. Therefore, headlines are positioned carefully and usually set in larger type than other textual elements of the ad. Bove and Arens (1989) and Wilmshurst and Mackay (1999) describe the major functions of a headline: to draw attention to the ad, select the reader, lead the reader into the body copy, present the complete selling idea (USP), promise the customer a benefit, and present product news of interest to the reader. These functions show that a headline should influence the attraction and intention effects of the ad. In contrast, as it must be short and appealing, it often lacks clarity and informational content. This leads to the fourth set of hypotheses: H4a: H4b: H4c: The ads informational value is not increased when a headline is present compared to when it is absent. The ads attraction value is increased when a headline is present compared to when it is absent. The intention effect of the ad is increased when a headline is present compared to when it is absent.
Method
An experimental conjoint design has been developed in order to test the four series of hypotheses. Four independent variables were manipulated to build up the advertisement stimuli: the picture, the logo, the text and the headline. For each of these attributes, attribute levels were established (see Table 1). As regards picture, the major purpose was to test hypotheses related to its presence/absence in the ad. When present, we decided to test three types of picture in order to meet our sponsors expectations (i.e. Namurs Tourism Board): a view of nature (park and pond), a view of culture (historic buildings) or a recreational stimulus (nude painting by Flicien Rops). This choice was inspired by existing typologies of tourism attractions and images that often make a distinction between nature, culture and recreation. Among the range of studies that have focused on the content of tourism promotional material (for a review, see Dann 1996), Dilley (1986) found that brochures allocate 32% of their space to landscapes, 36% to culture, and 32% to recreation and services.
513
It should be noted that no assumpTable 1: Format attributes and tion was made as to the relative effecattribute levels used in the study tiveness of these three types of visual. Format attribute Attribute levels No picture was assumed to elicit a better response than others. The logo of Picture Nude painting Namurs Tourism Board was either Historic buildings Park and pond present or absent. The text could be of No picture three possible types: no text, short text Logo Yes or long text, both relating to the same No selling propositions (i.e. well-preText Long text Short text served heritage and the dynamism of a No text young capital city) but with a varying Headline Interrogative headline number of words. Finally, the headline Descriptive headline could be of three possible types: no No headline headline, an indirect/interrogative headline (Do you want to get into the secret?) and a direct/affirmative headline (Namur also has its secrets). These two headlines share the same semantic content which attracts readers into the body copy to learn about the destinations secrets. Belch and Belch (2007) make a distinction between direct and indirect headlines. Direct headlines are straightforward and informative in terms of the message they are presenting and the target audience they are directed toward, whereas indirect headlines are not straightforward but generate curiosity or intrigue so as to motivate readers to become involved with the ad (p. 283). The other features of the advertisements (name and address of the Tourism Board, fonts, size of pictures, number of words used in the headline etc.) were kept unchanged. All attribute levels combine in 72 (i.e. 4 2 3 3) possible advertising formats. These 72 possible combinations were reduced to 16 alternatives according to an orthogonal main-effect design (Addelman 1962). All 16 stimuli were produced to resemble real colour advertisements in full-page size (see Figure 1). This full-profile approach was preferred to pair-wise comparisons because it represents a more realistic evaluation task for respondents. An additional control advertisement (exactly the same as one of the 16 stimuli, i.e. card 3) was added as a reliability check. All 16 stimuli were presented to each subject in a random order in order to avoid priming effects. Respondents were asked to rate each stimulus
514
Figure 1: Examples of designed advertisements
515
separately on a 7-point Likert scale for its informational value (Does this advertisement inform you on what can be done in Namur?), its attraction value (Does this advertisement make you feel attracted by Namur?), and the visiting intention elicited by the advertisement (Does this advertisement make you want to visit Namur?). The use of metric conjoint analysis has now surpassed that of the non-metric approach as ratings are more convenient for respondents and make analyses easier than rankings (Wittink & Cattin 1989). This results in higher reliability and validity. In order to control for priming and possible fatigue effects due to the number of stimuli to rate, we computed correlations between the ratings of the two identical stimuli (cards 17 and 3) and we obtained coefficients greater than 0.90 (with p-values < 0.05) for the three dependent variables. Such results show the reliability of our data. The quota sampling method was used to build a small representative sample of the Belgian population according to gender, linguistic region, and age category. Using a questionnaire, a total of 127 face-to-face interviews were conducted in different Belgian cities. Respondents aged 15 years and over were selected by convenience on the street, on their way to work or at their workplace. It should be mentioned that two versions of the set of 16 stimuli and accompanying questionnaires have been developed: the first one in French for the Walloon respondents and the second one in Dutch for the Flemish participants. All respondents were potential visitors who had not previously visited Namur. The final sample was composed of 66 males and 61 females; 76 Flemish and 51 Walloon respondents; 38 people aged 1524, 27 between 25 and 34, 23 between 35 and 54, and 39 aged 55 or more. Both conjoint analysis and analysis of variance were used in this study. Conjoint analysis helped in estimating the relative importance given by the respondents to the four attributes of the message format and the utilities attached to each attributes level. In addition, a series of inferential analyses (t-tests and one-way ANOVAs) were conducted on each of the three dependent variables to uncover the main effects of the four independent variables.1 The programs CONJOINT LINMAP (Srinivasan & Shocker 1982), based on linear programming, and SPSS have been used to run the analyses. CONJOINT LINMAP offers a distribution of violations
1
It has not been possible to look into the interaction effects due to the nature of the experimental design, i.e. a conjoint choice task.
516
as a goodness-of-fit measure. Predicted and actual evaluations of stimuli are compared and a percentage of violated pairs is calculated. Respondents with more than 50% violations were discarded, leaving us with, respectively, 123, 126 and 118 usable respondents for the analyses on the three dependent variables.
Results
Before testing the four series of hypotheses described above, Table 2 presents the aggregated relative importance weights of the four ad components for the three dependent variables. It shows that the visual is largely prevailing in the attraction and intention effects of print advertisements for a tourism destination. To a smaller extent, the headline also helps to attract and persuade consumers. On the other hand, the text is by far the most powerful tool for conveying information. Finally, the relative importance weight of the logo is very low, whatever the desired effect. We now come to the first series of hypotheses, which are related to the visual. Table 3 shows that the type of picture has a significant influence on the levels of attraction and intention responses. In contrast, the type of picture does not influence the informational value of the advertisement.
Table 2: Relative importance weights (%) for the three dependent variables
Picture Attraction Information Intention 78.77 21.18 57.36 Logo 1.03 1.50 3.58 Text 5.40 72.96 30.19 Headline 14.80 4.36 8.87 Total 100 100 100 n 123 126 118
Table 3: The influence of the type of picture on the effectiveness of print ads
Differences Attraction Information Intention Between Within Between Within Between Within Sum of squares 6.349 1.499 4.161 10.396 6.180 1.838 df 3 12 3 12 3 12 Mean square 2.116 0.125 1.387 0.866 2.060 0.153 F 16.946 1.606 13.448 p-value 0.000 0.241 0.000
517
Therefore, H1b and H1c are supported while H1a is not. This contradiction the results of Tscheulin and Helmig (1998) and Jaeger and MacFie (2001) can be explained by the nature of the product used in this study. The visual may show less informational value in advertisements for intangibles such as a tourism destination than for tangibles. In tourism, pictures are an established means of inducing imagery (subjective knowledge) and emphasising the symbolic attributes of the destination rather than its actual features. Therefore, pictures assist with the evaluation and persuasion process more than with the information process (MacInnis & Price 1987; MacKay & Fesenmaier 1997). The part-worth utilities for the two significant effects are displayed in Figure 2. For both attraction and intention, we see that having a picture is much more preferable than having no picture at all, with the view of nature and the nude painting being more preferred and a view of culture being less preferred than the average (i.e. 0). It is surprising that the nude painting is less effective than the view of the park and pond, especially in arousing a visiting intention. This may be the result of a lack of credibility related to the sex-oriented content of the picture. Manfredo et al. (1992) have stressed the relevance of source credibility, expertise and trustworthiness in determining the effectiveness of image advertising for tourism destinations.
Figure 2: Estimation of part-worth utilities for the attraction and intention effects according to the type of picture
20 15 10 5 0 5 10 15 20 25 Nude Culture Nature No picture Attention (n = 123) Intention (n = 118)
518
The tests of the second series of hypotheses related to the logo do not lead to any significant result: the presence/absence of a logo influences neither the informational value, nor the attraction level, nor the visit intention triggered by the advertisement. These findings are in line with H2a and H2c but do not support H2b. This contradiction may be the consequence of the marginal importance weight of the logo for the attraction effect (i.e. 1.03). In conjoint experiments, attributes with small relative importance weights should be interpreted with caution (Srinivasan & Shocker 1982). Weights near to zero indicate that respondents have ever or not used the attribute in their product evaluations, be it ratings or rankings. Therefore differences between attribute levels are seldom significant. Table 4 shows that the type of text does not influence the ads attraction power. However, significant differences are found with respect to the informational value and the behavioural intention conveyed by the print ad. These results are in line with the third series of hypotheses described above. When looking at part-worth utilities for the informational value of the advertisement, Figure 3 shows that having no text is far less preferred than having a text and that long text is considered more informational than short text. The same logical interpretation applies to the visiting intention elicited by the advertisement. However, the differences between the utilities attached to the three attribute levels are smaller, as illustrated by Figure 3. Finally, testing the fourth series of hypotheses, which are concerned with the headlines of print ads, leads to no significant result: the type of headline does not influence any of the three dependent variables.
Table 4: Influence of type of text on effectiveness of print ads
Differences Attraction Information Intention Between Within Between Within Between Within Sum of squares 0.178 7.669 13.052 1.504 2.966 5.052 df 2 13 2 13 2 13 Mean square 0.089 0.590 6.526 0.116 1.483 0.389 F 0.151 56.407 3.815 p-value 0.861 0.000 0.050
519
Figure 3: Estimation of part-worth utilities for the information and intention effects according to the type of text (n = 126)
30 25 20 15 10 5 0 5 10 15 20 25 30 Intention (n = 118) Information (n = 126)
Long text
Short text
No text
Therefore, H4a is confirmed while H4b and H4c are not supported. Again, it could be argued that, in conjoint measurement, significant differences seldom occur for attributes with small importance weights, as the relative importance of an attribute is directly determined from the range between its extreme levels (for qualitative features). Moreover, these nonsignificant results may indicate that the headline is not serving its function as an advance organiser for those exposed to the ad. However, this studys experimental design does not facilitate testing of the relationship between the headline and the other elements of the ad. Other series of analyses were conducted on each of the three dependent variables to control for the three sociodemographic variables (i.e. gender, age category and linguistic region). However, only one significant difference has been found for the interaction effect between the linguistic region and the type of text for the informational value (F5,26 = 50.443). French-speaking respondents find long text more informational than short text or no text, whereas the opposite relationship is true for Dutchspeaking people.
520
Discussion and implications

This study confirms the prevalence of the visual and the body copy in print advertisements. However, those two components of message format have different functions. On the one hand, the use of pictures is especially effective in attracting consumers and arousing buying intentions. This is in line with previous findings that the visual component of ads affects both beliefs and attitudes (Edell & Staelin 1983; MacInnis & Price 1987), and may lead to both cognitive and affective responses (Mitchell 1986). Visuals are of particular importance for intangibles such as tourism destinations: since tourism is uniquely visual, photographs are considered paramount to successfully creating and communicating an image of the destination (MacKay & Fesenmaier 1997, p. 540). The visual component of ads helps in building an image through the formation and modification of beliefs and thereby differentiating the market offering. In addition, more than for any other product, images related to tourism destinations have a major informational and tautological role: the often heavily romanticized photographs featured in tourism promotion have sufficient semiotic autonomy as to tell prospective visitors what pictures to take with their own cameras (Chalfen 1980, p. 27). On the other hand, the use of text is very effective for conveying information and, to a lesser extent, for arousing a buying intention. An obvious implication of these findings is that advertisers should pay particular attention to visual design in persuasive advertising campaigns and to textual design in informative campaigns. This is not to say that they should not be used in combination: the copy text and the illustrations are intended to communicate different but reinforcing messages (Jaeger & MacFie 2001, p. 190). In contrast, with picture and text, the two other elements of the message format that were manipulated in this study, i.e. the logo and the headline, do not appear to play major roles in print advertisements for tourism destinations. Another managerial recommendation is to use conjoint analysis as a powerful technique for pretesting ads. The partial utilities that are generated for each attribute level can be used to rank different stimuli based on their total utility and to make simulations based on shares of preference for different potential ads. For example, in this study, the ad with the view of park and pond, a long text, the logo and an interrogative headline proved to be most effective for persuading people to visit Namur (26.64% share of preferences).
521
Of course, this study has a few limitations. First, results are based on a particular type of product, i.e. a tourism destination. Communication for such tourism products is characterised by the need to induce imagery through prevailing pictorial content. However, Dann (1996) notices that the language of tourism very rarely speaks through pictures alone: where photographs are featured, almost without exception they appear in tandem with a verbal message (p. 188). This would indicate that this studys findings could be generalised to other product categories. However, for tangible and less differentiated products such as fmcgs, we expect that the visual would be less important and effective, especially in conveying information. In contrast, the logo/brand name and headline would be more important and effective, especially in attracting and persuading consumers to buy the product. This study also has a number of technical and methodological limitations. Investigating the interaction effects between the four format variables was not possible in this study due to the experimental design chosen (i.e. a conjoint task). Such interactions would be worthy of investigation. More particularly, the headlinetext and headlinepicture relationships deserve closer scrutiny as headlines are often used to alert the reader to the main theme of the copy or picture. Moreover, the three dependent variables of our model have been measured with one single item each, in order not to overwhelm respondents who were asked to rate 16 stimuli in the conjoint task. Using more items to measure each variable and/or using other dependent variables (e.g. ad comprehension or time to process the ad) would have been relevant. Finally, we neither checked subjects established image of Namur before the experiment, nor measured their enduring involvement in tourism and their situational involvement for tourism ads, depending namely on whether or not they were already planning their next excursion or city trip.
References
Addelman, S. (1962) Orthogonal main-effect plans for asymmetrical factorial experiments. Technometrics, 4(1), pp. 2146. Barry, T.E. & Howard, D.J. (1990) A review and critique of the hierarchy of effects in advertising. International Journal of Advertising, 9(2), pp. 121135. Belch, G.E. & Belch, M.A. (2007) Advertising and Promotion: An Integrated Marketing Communications Perspective. New York: McGraw-Hill. Bove, C.L. & Arens, W.F. (1989) Contemporary Advertising. Homewood, IL: Irwin.
522
Burns, K.S. & Lutz, R.J. (2006) The function of format: consumer responses to six on-line advertising formats. Journal of Advertising, 35(1), pp. 5363. Chalfen, R. (1980) Tourist photography. After Image, Summer, pp. 2635. Dann, G.M. (1996) The Language of Tourism: A Sociolinguistic Perspective. Wallingford: CAB International. Davidson, T.L. (1994) Assessing the effectiveness of persuasive communications in tourism, in J.R. Brent Ritchie & C.R. Goeldner (eds) Travel, Tourism, and Hospitality Research. A Handbook for Managers and Researchers. New York: Wiley, pp. 537544. Dilley, R. (1986) Tourist brochures and tourist images. Canadian Geographer, 30(1), pp. 5965. Edell, J.A. & Staelin, R. (1983) The information processing of pictures in print advertisements. Journal of Consumer Research, 10(1), pp. 4561. Govoni, N. (2004) Dictionary of Marketing Communications. Thousand Oaks: Sage. Gretzel, U., Yuan, Y.-L. & Fesenmaier, D.R. (2000) Preparing for the new economy: advertising strategies and change in destination marketing organizations. Journal of Travel Research, 39(2), pp. 146156. Hanssens, D.M. & Weitz, B.A. (1980) The effectiveness of industrial print advertisements across product categories. Journal of Marketing Research, 17(4), pp. 294306. Hecker, S. & Stewart, D. (1988) Nonverbal communication: advertisings forgotten elements, in S. Hecker & D. Stewart (eds) Nonverbal Communication in Advertising. Lexington, MA: Lexington Books, pp. 38. Houston, M.J., Childers, T.L. & Heckler, S.E. (1987) Picture-word consistency and the elaborative processing of advertisements. Journal of Marketing Research, 24(4), pp. 359369. Jaeger, S.R. & MacFie, H.J.H. (2001) The effect of advertising format and mean-end information on consumer expectations for apples. Food Quality and Preference, 12, pp. 189205. Johnson, R.R. & Messmer, D.J. (1991) The effect of advertising on hierarchical stages in vacation destination choice. Journal of Advertising Research, 31 (December), pp. 1824. Kroeber-Riel, W. (1993) Bildkommunikation: Imagerystrategien fr die Werbung. Mnchen: Vahlen. Lavidge, R.L. & Steiner, G.A. (1961) A model for predictive measurements of advertising effectiveness. Journal of Marketing, 25(6), pp. 5962. Lutz, K.A. & Lutz, R.J. (1977) Effects of interactive imagery on learning: application to advertising. Journal of Applied Psychology, 62(4), pp. 493498. MacInnis, D. & Price, L. (1987) The role of imagery in information processing: review and extensions. Journal of Consumer Research, 13(4), pp. 473491. MacKay, K.J. & Fesenmaier, D.R. (1997) Pictorial element of destination in image formation. Annals of Tourism Research, 24(3), 537565. MacKay, K.J. & Smith, M.C. (2006) Destination advertising: age and format effects on memory. Annals of Tourism Research, 33(1), pp. 724.
523
Manfredo, M., Bright, A. & Haas, G. (1992) Research in tourism advertising, in M. Manfredo (ed.) Influencing Human Behavior. Champaign, IL: Sagamore, pp. 327368. Mayo, E.J. & Jarvis, L.P. (1981) The Psychology of Leisure Travel. Boston: CBI. McWilliams, E.G. & Crompton, J.L. (1997) An expanded framework for measuring the effectiveness of destination advertising. Tourism Management, 18(3), pp. 127137. Miniard, P.W., Bhatla, S., Lord, K.R., Dickson, P.R. & Unnava, H.R. (1991) Picturebased persuasion processes and the moderating role of involvement. Journal of Consumer Research, 18(1), pp. 92107. Mitchell, A.A. & Olson, J.C. (1981) Are product attribute beliefs the only mediators of advertising effects on brand attitudes? Journal of Marketing Research, 18(1), pp. 318322. Mitchell, A.A. (1986) The effect of verbal and visual components of advertisements on brand attitudes and attitude toward the advertisement. Journal of Consumer Research, 13(1), pp. 1224. Morelli, G. (1970) Pictures and competing pictures as mediators in paired-associate learning. Perceptual Motors and Skills, 30, pp. 729730. Ogilvy, D. (1985) Ogilvy on Advertising. New York: Random House. Okoroafo, S. (1989) Branding in tourism, in S. Witt & L. Moutinho (eds) Tourism Marketing and Management Handbook. London: Prentice Hall, pp. 2326. Olson, J., McAlexander, J. & Roberts, S. (1986) The impact of the visual content of advertisements upon the perceived vacation experience, in W. Joseph, L. Moutinho & I. Vernon (eds) Tourism Services Marketing: Advances in Theory and Practice. Cleveland, OH: Cleveland State University, pp. 260269. Percy, L. (1989) The influence of pictures on advertising effectiveness, in B. von Keitz (ed.) 3. Symposium zur Kommunikationsforschung: Institut fr Kommunikationsforschung von Keitz GmbH. Saarbrcken: Physica. Rau, P.-L.P. & Chen, D. (2006) Effects of watermark and music on mobile message advertisements. International Journal of HumanComputer Studies, 64(9), pp. 905914. Rossiter, J.R. & Percy, L. (1983) Visual communication in advertising, in R.J. Harris (ed.) Information Processing Research in Advertising. Hillsdale, NJ: Erlbaum. Siegel, W. & Ziff-Levine, W. (1994) Evaluating tourism advertising campaigns: conversion vs advertising tracking studies, in J.R. Brent Ritchie & C.R. Goeldner (eds) Travel, Tourism, and Hospitality Research. A Handbook for Managers and Researchers. New York: Wiley, pp. 559564. Srinivasan, V. & Shocker, A.D. (1982) LINMAP (Version IV): a FORTRAN IV computer program for analyzing ordinal preference (dominance) judgments and for conjoint and tradeoff analyses. Journal of Marketing Research, 19, pp. 601612. Stewart, D., Hecker, S. & Graham, J. (1987) Its more than what you say: assessing the influence of nonverbal communication in marketing. Psychology and Marketing, 4(4), pp. 303322. Sullivan, G.L. (1990) Music format effects in radio advertising. Psychology and Marketing, 7(2), pp. 97108.
524
Tscheulin, D.K. & Helmig, B. (1998) The optimal design of hospital advertising by means of conjoint measurement. Journal of Advertising Research, 38(3), pp. 3546. Unnava, H.R. & Burnkrant, R.E. (1991) An imagery-processing role of pictures in print advertisements. Journal of Marketing Research, 28, pp. 226231. Urbain, J.D. (1983) Sur lespace du touriste: un voyage en Tunisie. Elments pour une smiotique de lespace touristique des Franais. Espace Gographique, 2, pp. 115124. Wilmshurst, J. & Mackay, A. (1999) The Fundamentals of Advertising (2nd edn). Oxford: Butterworth-Heinemann. Wittink, D.R. & Cattin, P. (1989) Commercial use of conjoint analysis: an update, Journal of Marketing, 53(3), pp. 9196. Woodside, A.G. & Dubelaar, C. (2003) Increasing quality in measuring advertising effectiveness: a meta-analysis of question framing in conversion studies. Journal of Advertising Research, 43 (March), pp. 7885. Woodside, A.G. & Ronkainen, I.A. (1994) Improving advertising conversion studies, in J.R. Brent Ritchie & C.R. Goeldner (eds), Travel, Tourism, and Hospitality Research. A Handbook for Managers and Researchers. New York: Wiley, pp. 545558. Woollacott, J. (1982) Messages and the media, in M. Gurevitch, T. Bennett, J. Curran & J. Woollacott (eds) Culture, Society and the Media. London: Methuen, pp. 91112.
About the author

Alain Decrop is associate professor of marketing and head of the department of Business Administration at FUNDP (University of Namur). He is a member of CeRCLe (Centre for Research on Consumption and Leisure) and visiting professor at the Catholic University of Louvain (KUL). He holds masters degrees in history and economics, and a PhD in business administration from the University of Namur. His main research interests include consumer behaviour, advertising effects and tourism marketing. Address correspondence to: Alain Decrop, Department of Business Administration, University of Namur, Rempart de la Vierge, 8, B-5000 Namur, Belgium. Email: alain.decrop@fundp.ac.be
525
Bold thinking for busy marketers

BRANDING NIALL FITZGERALD
8 11:4 2/04 Page 50
Life and death in the world of brands

By NIALL FITZGERALD
Arrogance, greed, complacency, inconsistency and myopia are the causes of brand death that should appear on the cororners certificate. Here is a brilliant treatise from the CEO of one of the worlds most celebrated brand owners, Unilever, on the everlasting potential of brands and how companies wilfully but unknowingly commit brandicide
NG NDI B R A BRIGGS PADDY
grace ll f She d fell from n ory o The st great bra ow a H
ion satisfact omer businesses of cust to er levels y consum trying of usly n highest our man conscio reputatio en d in all e also on bran orate ed. Wh book (pub we wer the corp d and valu the terS INAL e Identity s re ance es visited SEM Olin de of ensu orat BRIGG utiv was enh N HIS gn, Corp multitu ects Wally of corDDY Shell Shell exec on a 9), s proj By PA the and desi in 198 form d, in ussions stment fact senior rent line in for disc na inve by the wn lished three diffe endorse p of Chi ely aided ritory the dec e one of kno ribes cture: the has a grou it scale ady ed of onc desc large ion was hug was alre g. Around stru which The spe of Shell, ected task d porate the corporat panies e and their tion ll bran Hong Kon over 100 com st resp p nam ); the reputa ies which s and grou the She ected in were well lds mo the outside it the compan Motors rates as activitie s with the wor seems to and resp ld there ed Shell eral y ope may , n as fast endorse (e.g. Gen compan the worcountry-bas strategy. ch is verbrands have bee tity re a ds whi to the ilar ilar iden industry pany to sim g. a sim d whe of bran r or world n shockin the oil oil com am the othe brande a series pursuing m in veteran each an rbon stre to nor r) and ion The has bee cle Shell cribes how throughelated to gration hydroca the way Unileve corporat . arti inte des a tity be unr ion (e.g tical In this BRIGGS the full for oil all put in the nds which al iden Y corporatthic, in and visu the bra fairly manages oration petrol is PADD seeds of e expl noli ic Until n ly more nam mo olith n from t whe rs car. one general a mon More the first s were sow uses n was the poin a custome e and out . and how er ll. ple of of operatio brand nam Shell. throughhis exam chose She business eth problem ade ago tank this for one lls s ll ns, tog For a dec recentlyted under of She s the She the case rs in the e Olin than decisio rnal events, uent structur per cent ld bear tation of conduc always been a-merge ewhat 90 wor subseq mitous exte a once s meg issue som repu aorthan this has nt year creatout the The the of e extr al with cala status In rece clouded panies have multithrough he wrote. to a quit visu e e the and sed have com nd nam name, symboli its name this time sector e of the rily, a mor ped the rocked le bra is ly ilab Shell extent by h around and som n temporae. BP drop name as was high unassa e logo dinary . Researc Shell ed, ofte structur r corporat they have that Shell cial h imagery not only that the branded from thei (althoug Castrol ed commer the ld also rich disshow Amoco they cou d but gnised ll was uity as d-eq ally swiftly . respecte most reco ld. She stry and soon bran wor the the l equ their title indu est was Tota from benkept in the the oil and rnal the larg the symbol leader in name) the FinaElf ious exte there are sures, ld in d e obv s e mea the wor bran carded indeed! as the le nam As Olin by som types in dy days As well ng a sing well. also, of all [Shell] . Hea of havi efits as name ificant business 500 lists efits ben single sign tune rnal A For ng? inte became employd it, go wro king for describe al identity for staff whole it all , t wor the visu ere did1980s I was , increasinglyy and rallying pointify with Wh iden as a g and late ing ever benefits could In the Hong Kon were mak of the mercial brand ees ll in rprise. huge com having a re we as benefitsidentity She ente em) the e the na whe ll Given accrue from tity syst k this in Chi to maximis the She rately of thin al iden can so accu the effort strength that (and visu you would s had to offer lls unique ly Olin name ter 2004 as She Wal striving a marketingand er Win that re, familiar d. By et Lead dAwa d reputation Mark Bran describe
e.qxd - blue styl over June

s runs y Brigg that advis Padd ltancy ity. consu rate ident corpo es on bran
20/05/05
14:38
Page 1
ISSUE NUM
50
BER 29
N E W
Market Leader
Long Live the Brand
, K I N G T H I N
T E R E N D I F F
t was 2 April 1993 Marlboro Friday. Not a lot happened that day. In fact, the only thing that happened was an announcement from Philip Morris that it was cutting the price of a packet of cigarettes by 20%. But that was enough. Stock prices fell off the edge of a cliff. Heinz, Quaker Oats, Coca-Cola, Pepsico, Procter and Gamble, RJR Nabisco and all quoted advertising stocks plunged wildly, with Philip Morris itself being the hardest hit. There was genuine panic on Wall Street, and way, way beyond. Look carefully at most of what is written and said about brands and it is clear that, at heart, most experts and commentators believe that the art of branding is little more than an immensely clever technique for persuading consumers to pay more for their goods and services than any rational person needs. Brand values, so the argument goes, are first artificially constructed, then imposed from above on a gullible, consuming public. In this view of the world, brands certainly have a real value but only to their owners, never their users. The art of branding is a black art a means of persuading the innocent to part with large sums of money in return for something called image which has no material usefulness at all. The markets admire this greatly, of b i k B
I Subscribe to Market Leader, the journal where true market leaders share some of the thinking that helped them reach the top and stay there. I A single-minded strategic perspective on marketing, brands and business from brand strategy to globalisation, and social responsibility to new technology. I High-calibre writing from the worlds leading companies, consultancies and business schools Unilever, Diageo, HSBC, McKinsey, London Business School, and dozens more.
lective senses, rationality will return, choice will once again become based solely on price, and the age of the brand will be over for good. With these largely unspoken assumptions as background, it becomes a great deal easier to understand Marlboro Friday. If the whole branding business is such a con trick, at some point the whole thing is bound to come crashing down to earth. The moment one manufacturer significantly cuts the price of one of its brands, hysteria erupts. The games up! The bubbles burst! Time to jump ship!
SUMMER
2005
There were, of course, a great many people who, though astonished by the rhetoric, remained staunchly unmoved. Just as the actions of those h i k d di b i
Market Leader is published quarterly by WARC in association with the Marketing Society
Mark t L d
ISSUE NUM BER 29 SUMMER 2005
I V E S P E C T P E R S
MEASURING
MARKETING
How to e demonstrat marketings profitability

Robert Shaw
l Alex Batchelor nagement too et tion as a ma tions Les Bin g Brand valua g communica ng? David Lan ng marketin out marketi Evaluati to know ab nt y analysts wa do Cit
What
? off the four Ps Time to kilS.l DEV & DON E. SCHULTZ N

CHEKITA
n succ Tescos fashio to cut-price: RNER JOHN HOE From catwalk k nt work acquisitions do ergers andJOHN WARD Why m
ess
Market Leader is refreshingly high minded while remaining firmly on planet earth. It does what it says on the cover: lead. And thats what all top marketers should do.
Tim Ambler
Senior Fellow, London Business School
SUBSCRIBE TODAY
www.marketleadermagazine.com +44 (0)1491 411000
Internet advertising effectiveness

The effect of design on click-through rates for banner ads
Helen Robinson, Anna Wysocka and Chris Hand
Kingston University
Online advertising has experienced phenomenal growth since its inception in 1994. This empirical study investigates the impact of seven creative characteristics of banner ads on the effectiveness of online advertising using a multiple regression model. A random sample of 209 banner ads was drawn from a sampling frame of advertisers, provided by an advertising agency specialising in internet advertising for the gaming industry. The findings of this study are broadly consistent with past research into online advertising efficiency, indicating that the creative characteristics of effective banner ads in the online gaming arena include: a larger size, absence of promotional incentives and the presence of information about casino games. In contrast, banner features such as animation, action phrase and presence of company brand or logo were ineffective in generating clickthroughs. Contrary to expectations, long messages on banners were associated with higher click-through rates.
Introduction
Since the first banner ads appeared in 1994, the internet advertising industry has experienced exceptional growth. The Interactive Advertising Bureau (IAB) heralded 2003 as the most successful year in association history, recording full-year revenue as almost US$7.3 billion (IAB 2003). Existing academic research encompasses a broad spectrum of studies on internet advertising effectiveness measured by direct response and branding metrics. The few studies that have included click-through rate (CTR) as a measure of online advertising effectiveness have examined the relationships between CTR and a series of factors such as web user motives, audience targeting, exposure frequency, copy content and a limited number of design elements of banner ads. Advances in technology have made
527
online gambling one of the fastest-growing industries in recent years. While in 1999 there were only 300 online casinos worldwide, this number had grown to 1800 by 2004 (Constable 2003; Catty 2004). This empirical investigation broadens the existing academic knowledge of online advertising efficiency and, more specifically, contributes to the understanding of which creative elements of banner ads impact clickthroughs. In addition, this study is the first to examine internet advertising effectiveness in the context of the online gaming industry. It is not the intention of this study to enhance the effectiveness of the advertising of online casinos. Rather, it centres on whether the effects of banner design on click-through differ between this market and others. Online gambling is an example of an industry where those who participate online may not have considered participating offline. This contrasts with most other examples of e-commerce (e.g. purchasing groceries, clothes, books). Hence, an online casino has to build credibility from its banner ads and website alone. Consequently, there might be reason to believe that different banner characteristics would increase effectiveness more than those previously found in other industries. Nevertheless, given ongoing concerns and much recent debate over the growth of the gambling sector and relaxed restrictions on broadcast advertising (from September 2007 in the UK, for example), an understanding of the effectiveness of online advertising is of interest to both the casinos and their regulators.
Previous research
Internet advertising, and in particular banner effectiveness, in this relatively new medium has received considerable attention from academics and practitioners. The study of online advertising effectiveness has been conducted through two alternative paradigms. The first, widely used in academic research, argues that banner adverts should be considered as a form of marketing communication used to raise brand awareness. The second, predominantly used in empirical research, contends that the internet is a direct marketing medium, and hence a banner ad is likened to a coupon in print media. Accordingly, the success of the banner ad should be measured through the return rate or CTR for the internet (Chandon et al. 2003).
528
INTERNET ADVERTISING EFFECTIVENESS
Dreze and Hussherr (2003) found that internet users avoided looking at ads while online, and hypothesised that internet users might perceive banner ads in their peripheral vision. Similarly, Janiszewski (1998) claimed that peripheral vision allows individuals to recognise objects that are located outside their focal point of attention. Several studies suggest that click-through effectiveness may depend on web user motives. Briggs and Hollis (1997) argue that the primary factor in generating click-throughs is the nature of the audience and what the inherent interest in the product category may hold for them. Further studies have concluded that banner ads that complement the users web motives may be more effective (Raman & Leckenby 1998; Rodgers & Thorson 2000; Rodgers 2002; Danaher & Mullarkey 2003). Practitioners and academics agree that while repetition reduces clickthroughs, it builds brands (Broussard 2000; Dynamic Logic, 2000; Gugel 2001; Chatterjee et al. 2003; Danaher & Mullarkey 2003; Dreze & Hussherr 2003). A considerable body of research has demonstrated that successful targeting of online ads improves CTR (Briggs & Hollis 1997; Sherman & Deighton 2001; Chandon et al. 2003; Chatterjee et al. 2003). Relevance of the advertisement to the site may also be a determinant of click-through rates. Chang-Hoan (2003) found that those more involved in a product were more likely to click through, but also that users who were interested in the site and shown advertisements for products and services related to those on the site also achieved a higher click-through rate. The relationship between the banner size and CTR is conflicting. Baltas (2003) found that bigger ads are more effective in attracting attention and (hence more likely to) trigger response. Chandon et al. (2003) found positive interactions on five banner sizes, although no significant difference between the two bigger sizes. In contrast, Dreze and Hussherr (2003) tested the standard banner against two other banner sizes and found the smaller ads performed just as well as the large ones, in accordance with Cho (2003) who found no significant relationship between banner size and clicking. Comparably, Rettie et al. (2004) established that banner size impacted click-through and post-impression measures differently. Research on the impact of price and promotions on click-throughs has revealed that none of these various stimuli (gift, rebate or free offers) has a direct effect on CTR (Chtourou et al. 2002; Baltas 2003; Rettie et al.
529
2004). Chtourou et al. (2002) found that the mention of price reduced direct response. Rettie et al. (2004) discovered that banners which mentioned neither price nor promotional offers had the most effect on clickthrough and post-impression rates. Conflicting evidence exists on the impact of branded banners on clickthrough rates. Research conducted by Baltas (2003) and Chandon et al. (2003) revealed a negative impact and suggests that unbranded banners might stimulate greater curiosity, leading to click-through. Similarly, Dahlen (2001) found that familiar (as opposed to unfamiliar) brands received double the click-through. In contrast, Briggs and Hollis (1997) argued that the practice of running unbranded banners ... surely runs counter to the concept of brand building through ad banner exposure, a view supported by Dynamic Logic (2002). Dreze and Hussherr (2003) concluded that, in terms of artistic influences, audiences were most affected by the banner message rather than how the message was conveyed. They concluded that artistic execution overall had little effect on both click-throughs and traditional memorybased effectiveness measures, with the exception of animation, which influenced aided brand recall. Baltas (2003) found that banners with lengthy messages and multiple frames (animation) received fewer clicks. He reasoned that these two factors increase the complexity of an ad and hence have a negative effect on the viewers attitude towards and response to the banner. It has also been suggested that animated banners may be more difficult to remember than static ones (Burke et al. 2005). On the other hand, Chandon et al. (2003) and Lothia et al. (2003) concluded that animation improved click-through rates. Overall, existing research on the relationship between creative characteristics of banners and CTR reveals conflicting results. An early study by Hofacker and Murphy (1998) suggests that creative factors tend to increase response rates. The inclusion of clichd messages such as click here or click now, along with animations and cryptic messages, was suggested as methods of increasing response rates. More recently, evidence has shown little support for a positive relationship between artistic variables and direct response (Dreze & Hussherr 2003). Moore et al. (2005) indicate that there may be a conflict between getting a user to attend to a banner while maintaining a favourable attitude towards it. Their research suggests that the use of incongruent colour schemes in banners compared
530
to the rest of the website results in more attention being given to that banner, but that the attitude towards the banner was more negative than towards a banner more congruent with its website.
Research design
The aim of this empirical study is to investigate the impact of seven design characteristics of banner ads on the click-through rate using data from 209 real banner advertisements hosted on a single website in the context of online gaming. The research builds upon past research in the area of online advertising efficiency and, in particular, a study conducted by Baltas (2003). The study investigates relationships between the CTR and the seven design elements of banner ads discussed earlier: banner size, message length, promotional incentive, animation, action phrase, company brand/logo and casino games. The design element games is also included to reflect the focus on the online gaming industry. This study examines the impact of creative elements of banner ads on CTR, the direct response metric used for measuring internet advertising effectiveness. Data were provided by an advertising agency (specialising in internet advertising for the gaming industry) that employs the Adtrack system, which records impressions, clicks and CTR for each banner on a daily basis. The database comprised an inventory of 70 advertisers and approximately 1500 banners that had appeared between 2001 and 2004. A sample of 14 advertisers and their 209 banners was extracted. The original inventory of 70 advertisers was filtered to 58 advertisers who ran campaigns for at least a three-month period. In order to reduce the sample to a manageable size, 14 advertisers were selected randomly from the list of 58 remaining advertisers and, finally, the CTR data for the resulting 209 banners from the 14 advertisers for a three-month period were extracted, forming the data set. The three-month period was chosen to ensure a substantial database for the ensuing statistical analysis and also to control for the moderating impact of the length of an ad campaign on the CTR. Click-through rates provide the marketer with considerable advantages, as click-throughs record voluntary behaviour in the actual medium environment. Furthermore, the data are collected unobtrusively; are based on observed behaviour rather than self-reporting; and are free from researcher
531
bias and recorded on a census of consumers rather than a sample (Chatterjee et al. 2003). A standard multiple regression model was used to examine the relationship between the dependent variable (CTR) and the seven creative banner characteristics to determine which had a significant effect on CTR.
Research findings
Prior to running the regression, the CTR variable was found to be highly skewed, with few banners obtaining high rates of click-through (the histogram is shown in Figure 1). An initial run of the regression model produced non-normal errors. In order to overcome this problem, CTR was transformed by taking logarithms (a common remedial measure for nonnormal errors in econometrics, for example see Gujarati (1995)). The source website employs a total of seven distinct sizes of banner that were included in the original model as six dummy variables (the standard
Figure 1: Histogram of click-through rates

200
150
Frequency
100
50
0 0.00%
5.00% CTR
10.00%
15.00%
532
banner size, 468 60 pixels, being the base category). One dummy was dropped due to a collinearity problem. A more conventional approach might be to include banner height and width in the model (as one reviewer suggested). However, given the design of the source website, it was impossible to separate the effects of banner size and placement. A banner of a particular size can appear only in a given place on the web page (e.g. long, thin banners appear only at the top and the bottom of the page). Consequently, it was not possible to separate the effect of placement and size; therefore generalising the results to other banner sizes would be potentially misleading. The effect of message length was also captured by a series of dummy variables. Again, at first glance, the number of words would be the most obvious measure to use. It was decided to code message length as a series of dummies for two reasons. First, the message lengths were clustered around a few lengths. Second, the effect of adding a word or two to the message on click-through is likely to be small; larger differences are likely to be found between messages of different lengths. These differences would be more apparent from a comparison of the dummy variable coefficients. An alternative approach would be to use the total number of words on the banner, which would permit testing for polynomial relationships, as long as there was sufficient variation in message length across banners. The effects of the presence of a logo, a promotional offer, the use of a clichd phrase and whether the games offered were listed on the banner were captured by a series of dummy variables. The study also investigated whether animation on banners attracts greater click-through. Rather than use another dummy variable, the number of transitions between frames was used (i.e. 0 = static, 1 = two frames or 1 transition, and so on). The multiple regression model takes the form of the following equation: log CTR = b0 + b1 Bsize 120 120 + b2 Bsize 120 240 + b3 Bsize 175 60 + b4 Bsize 234 60 + b5 Bsize 360 240 + b6 Smsg + b7 Mmsg + b8 Lmsg + b9 Frames + b10 Promotion + b11 Clich + b12 Logo + b13 Game + e
533
where: log CTR is the dependent variable, b0 is the constant, b1 to b13 are the regression coefficients for the corresponding independent variables as follows: Bsize 120 60 Bsize 360 240 = banner size dummies (base category = standard banner size 468 60) Smsg = short message (15 words) dummy Mmsg = medium-length message (610 words) dummy Lmsg = long message (1115 words) dummy (base category = very long messages, 16+ words) Frames = number of transitions if animated (= 0 if static) Promotion = promotional incentive dummy Clich = stereotypical action phrase dummy Logo = company brand/logo dummy Game = casino games offered at an online casino dummy e = residual. Inspection of a histogram of the residuals, a plot of the residuals against predicted CTR (shown in Figures 2 and 3 respectively) and collinearity diagnostics (shown in Table 1) suggested that the assumptions of normality, homoscedasticity and no multicollinearity were met. The model explains 68% of the variation in log CTR. As the dependent variable was transformed, the regression coefficients can not be directly interpreted as the effect of a unit change in the independent variable on CTR. Instead, the effect of each variable is calculated using (antilog b) 1, which gives a measure of the magnitude of the variables effect on CTR, shown in the magnitude column of Table 1 (Gujarati 1995).
534
Figure 2: Residuals histogram

40
30
Frequency
20
10
0 3 2 1 0 1 2 3 Regression standardised residual
Figure 3: Scatterplot of dependent variable against residuals (heteroscedasticity test)

3 Regression standardised predicted value
2 3 2 1 0 1 2 3 4 Regression studentised deleted (press) residual
535
Table 1: Regression results

Description (Constant) Banner 120 120 Banner 120 240 Banner 175 60 Banner 234 60 Banner 360 240 Short Mgs (15W) Med Mgs (610W) Long Mgs (1115W) Animation Promo incentive Clich Logo/brand Games b 0.111 0.296 0.339 0.414 0.289 1.784 0.526 0.438 0.635 0.061 0.792 0.016 0.148 0.361 Std error 0.293 0.168 0.199 0.198 0.165 0.181 0.252 0.178 0.169 0.036 0.113 0.114 0.199 0.123 t 0.381 1.762 1.701 2.089 1.753 9.861 2.084 2.454 3.748 1.698 7.012 0.141 0.740 2.924 Sig. 0.704 0.080 0.091 0.038 0.081 0.000 0.038 0.015 0.000 0.091 0.000 0.888 0.460 0.004 Magnitude (%) 10.506 25.621 28.752 33.900 25.099 495.362 40.904 35.467 47.006 5.918 54.706 1.613 13.757 43.476 Tol. VIF
0.828 0.813 0.823 0.830 0.584 0.441 0.341 0.434 0.858 0.824 0.864 0.773 0.674
1.208 1.230 1.215 1.215 1.713 2.268 2.933 2.306 1.166 1.218 1.157 1.294 1.485
R2 Adj. R2 F
0.681 0.659 31.668 (sig. = 0.000)
Note: Tol. = tolerance; VIF = variance inflation factor
Compared to the standard banner size (468 60 pixels) only two banner sizes have a significant impact on CTR (at the 5% level). Small banners (175 60 pixels) receive 33.9% fewer clicks than standard banners. In contrast, large banners (360 240 pixels) receive 495.4% more clicks than standard banners. However, it should be remembered that the size measure used for this study also captures a banners position on the webpage, hence not all of the increased click-through can be attributed to the size of the banner. Describing or listing the games available on the site also significantly increases CTR by 43.5%. It allows the website user to find a site that appeals to them more quickly as it reduces the time spent searching for a particular game. The regression model also produces some surprising results. Banners with short, medium and long messages receive significantly fewer clicks than very long messages. This result may reflect the finding that larger banners obtain a higher CTR; long messages tend to appear on larger banners (if only because they fit). Promotional incentives generate significantly fewer clicks than banners without such incentives. Apparently, the
536
offer of free bets is not enough to entice online gamblers to try a particular online casino; indeed, such offers would seem to have the opposite effect than was intended. The use of clichd phrases has no significant effect on click-through rates; neither does the presence of logos or brand emblems, perhaps reflecting the lack of brand awareness in this subset of the gambling industry.
Conclusions and recommendations

This research indicates that the design elements of effective banner ads include: a larger size, long message, absence of promotional incentives and the presence of information about casino games. In contrast, creative characteristics established as being ineffective in attracting a direct response include: stereotypical action phrases and the company brand/logo. These findings are generally consistent with past research in the area of online advertising effectiveness. The largest banner ads are more effective in generating click-through than the smaller-sized banners, replicating findings by Baltas (2003), Chandon et al. (2003) and Chtourou et al. (2002), and supporting the industry mantra that bigger is better (Briggs, 2001a; Dynamic Logic, 2004). However, this study contradicts some earlier findings that banner ads carrying a shorter message are more effective than those with a lengthy message. Animation, on the other hand, demonstrated an insignificant effect on users direct response (at the 5% level), a point of interest to media agencies since animated banners are more costly to produce. Therefore, Baltass (2003) contention that a wordy animated message increases the complexity of the ad, and hence receives a negative response from the audience, receives only partial support. Moreover, the discovery that banners containing no promotional offer attract more attention than those with such incentives, concurs with past studies (Chtourou et al. 2002; Baltas 2003; Rettie et al. 2004). The traditional attention-capturing characteristics of banner ads (such as action phrases and company brand/logo) play an insignificant role in eliciting a direct response. These findings are consistent with previous studies on the response to branded vs non-branded banners (Baltas 2003; Chandon et al. 2003). While some commentators have suggested that
537
unbranded banners stimulate greater curiosity, hence generating more clicks, others have argued that unbranded banners contradict the concept of brand building through banner exposure (Briggs & Hollis 1997). The evidence from the current study indicates that the presence of games in a banner ad positively impacts on CTR. Clearly, online gamblers find this aspect of the communication important to their decision making, and thus banner ads for the online gaming sector should comprise casino games in order to increase direct response. The findings of this study have strong implications for online communications; the evidence stresses the importance of fresh and innovative message tactics, given the ineffectiveness of the more conventional design tools employed to stimulate user interest. These results also suggest that the effectiveness of banner ad campaigns for online gaming (as measured by CTR) is governed by the same principles of artistic execution as those found in other, non-gambling-related sectors of the internet. This study used CTR to evaluate the effectiveness of banner ads. This metric is one of the oldest methods of evaluating the success of online advertising campaigns and is an appropriate tool for assessing direct marketing objectives. However, it has been criticised for not including the additional effects of online advertising such as branding (Briggs 2001b) and may be considered by some as a limitation of this study. Further research could investigate the branding-building effects of banner ads. Another limitation of the study is that it included a sample of banner ads hosted on a single website, namely an online gaming portal. Future research could extend the scope of the research to a number of websites.
Acknowledgements
The authors are grateful to participants at the 34th European Marketing Academy Conference, Milan, the 2005 Academy of Marketing Conference, Dublin, the editor and two anonymous referees for helpful comments.
References
Baltas, G. (2003) Determinants of internet advertising effectiveness: an empirical study. International Journal of Market Research, 45(4), pp. 505513.
538
Briggs, R. (2001a) Richer, bigger, more interactive advertising research studies. Measuring Success Series, 1(3), US Interactive Advertising Bureau, http//www.iab.com. Briggs, R. (2001b) The role of creative execution in online advertising success. Measuring Success Series, 1(4), US Interactive Advertising Bureau, http//www.iab.com. Briggs, R. & Hollis, N. (1997) Advertising on the web: is there response before clickthrough? Journal of Advertising Research, 37(2), pp. 3345. Broussard, G. (2000) How advertising frequency can work to build online advertising effectiveness. International Journal of Market Research, 45(4), pp. 439457. Burke, M., Hornof, A., Nilsen, E. & Gormon, N. (2005) High-cost banner blindness; ads increase perceived workload, hinder visual search, and are forgotten. ACM Transactions on ComputerHuman Interaction, 12(4), p. 423. Catty, J.P. (2004) Valuing online gaming enterprises. World Online Gambling, June, pp. 1216. Chandon, J.L., Chtourou, M.S. & Fortin, D.R. (2003) Effects of configuration and exposure levels on responses to web advertisements. Journal of Advertising Research, 43(2), pp. 217229. Chang-Hoan, C. (2003) Factors influencing clicking of banner ads on the WWW. Cyber Psychology & Behaviour, 6(2), pp. 201215. Chatterjee, P., Hoffman, D.L. & Novak, T.P. (2003) Modelling the clickstream: implications for web-based advertising efforts. Marketing Science, 22(4), pp. 520541. Cho, C. (2003) The effectiveness of banner advertisements: involvement and clickthrough. Journalism and Mass Communication Quarterly, 80(3), pp. 623645. Chtourou, M.S. & Guerin, F. (2001) What makes people like, and click on, an internet banner? European Society for Opinion and Marketing Research, Worldwide Audience Measurement Conference Proceedings, pp. 147167. Chtourou, M.S., Chandon, J.L. & Zollinger, M. (2002) Effect of price information and promotion on click-through rate for internet banners. Journal of Euromarketing, 11(2), pp. 2340. Constable, N. (2003) This is Gambling. London: Sanctuary Publishing. Dahlen, M. (2001) Banner advertising through a new lens. Journal of Advertising Research, 41(4), pp. 2330. Danaher, P.J. & Mullarkey, G.W. (2003) Factors affecting online advertising recall: a study of students. Journal of Advertising Research, 43(3), pp. 252267. Dreze, X. & Hussherr, F.X. (2003) Internet advertising: is anybody watching? Journal of Interactive Marketing, 17(4), pp. 823. Dynamic Logic (2000) Frequency plays a significant role in lifting awareness: four or more exposures doubles the impact of online branding, in Beyond the Click: Insights from Online Advertising Research, http//www.dynamiclogic.com. Dynamic Logic (2002) Bigger ads do not guarantee effectiveness, in Beyond the Click: Insights from Online Advertising Research, http://www.dynamiclogic.com. Dynamic Logic (2004) Consumer perceptions of various ad formats, in Beyond the Click: Insights from Online Advertising Research, http://www.dynamiclogic.com.
539
Gugel, G. (2001) Caught in the web: from ad weary to ad wearout. European Society for Opinion and Marketing Research, Worldwide Audience Measurement Conference Proceedings, pp. 1121. Gujarati, D. (1995) Basic Econometrics. London: McGraw-Hill. Hofacker, C.F. & Murphy, J. (1998) World wide web banner advertisement copy testing. European Journal of Marketing, 32(7/8), pp. 703712. Interactive Advertising Bureau (2003) The tide has turned for interactive advertising: IAB celebrates most successful year in association history. Press release online at http://www.iab.com/news/pr 2003 12 23.asp (accessed 31 March 2004). Janiszewski, C. (1998) The influence of display characteristics on visual exploratory search behaviour. Journal of Consumer Research, 25(3), pp. 290302. Lothia, R., Donthu, N. & Hershberger, E. (2003) The impact of content and design elements on banner advertising click-through rates. Journal of Advertising Research, 43(4), pp. 410418. Moore, R.S., Stammerjohan, C.A. & Coulter, R.A. (2005) Banner advertiser web site context congruity and color effects on attention and attitudes. Journal of Advertising, 34(2), pp. 7184. Raman, V.R. & Leckenby, J.D. (1998) Factors affecting webad visits. European Journal of Marketing, 32, pp. 737748. Rettie, R., Grandcolas, U. & McNeil, C. (2004) Post Impressions: Internet Advertising without Click-through. Kingston Business School, Kingston University. Rodgers, S. (2002) The interactive advertising model tested: the role of internet motives in ad processes. Journal of Interactive Advertising, 2(2), available online at http://jiad.org/vol2/no2/rodgers. Rodgers, S. & Thorson, E. (2000) The interactive advertising model: how users perceive and process online ads. Journal of Interactive Advertising, 1(1), available online at http://jiad.org/vol1/no1/rodgers. Sherman, L. & Deighton, J. (2001) Banner advertising: measuring effectiveness and optimising placement. Journal of Interactive Marketing, 15(2), pp. 6064.
About the authors

Helen Robinson is a principal lecturer in the School of Marketing at Kingston University, UK. Previously, she worked in the advertising industry, for JWT in London. Her current research interests include advertising and media planning, marketing communications and internet marketing. Her work has been published in a number of marketing journals. Anna Wysocka graduated from Kingston Business School with a masters degree in marketing, moving on to follow a career as a practitioner in the field. Chris Hand is a post-doctoral researcher in the School of Marketing, Kingston Business School. His research interests include the effects of eCommerce on consumer behaviour, the entertainment industries and the
540
application of quantitative methods in marketing. His research has appeared in journals such as the Journal of Brand and Product Management, Environment and Planning A and the Journal of Cultural Economics. Address correspondence to: Helen Robinson, Kingston Business School, Kingston Hill, Kingston upon Thames, Surrey, KT2 7LB, UK. Email: h.r.robinson@kingston.ac.uk
541
The Journal of Advertising Research

Theres advertising theory. And theres advertising practice. With JAR you get the best of both worlds.
www.JournalOfAdvertisngResearch.com
Directions in marketing communications research

An analysis of the International Journal of Advertising
Douglas West
University of Birmingham
This paper provides a content analysis of the inputs and outputs of the International Journal of Advertising (IJA) during the period 19922006 and was sparked by the 25th anniversary of the journal (19822006). A total of 348 papers were surveyed using a content analysis in order to provide researchers and readers with a better sense of the contribution of the IJA over the past 15 years. The analysis reveals a journal largely focused upon topics involving practice and effects with increasingly sophisticated statistical techniques employed. Single-authored papers appear to be in decline in favour of two- to three-authored papers, and author institutions widely domiciled across North America, Asia, the UK, Europe and Australasia.
Introduction
The year 2006 was a significant landmark for the International Journal of Advertising (IJA). The publication of issue 4 in 2006 marked 25 years of the IJA (19822006), and the purpose of this paper is, in some small part, to celebrate this achievement in the time-honoured fashion of a journal review. Additionally, the paper takes a reflective view of marketing communications research as demonstrated by a content analysis of the authors, institutions and papers published in the journal over the 15 years up to 2006 (19922006). The format of the paper is to examine the nature of the inputs (authors and institutions) and outputs (the nature of the papers published in the IJA).
543
Journal studies
As noted by Pasadeos et al. (1998), studies of scholarly literature can be categorised along six dimensions. Publishing productivity studies include an assessment of the contributions of particular authors and institutions (e.g. Barry 1990); comprehensive reviews set about establishing heuristics or paradigms on the conclusions reached in a large number of studies on a particular topic (e.g. Arndt 1986); meta-analyses utilise data-based conclusions on the findings from multiple studies (e.g. Crouch 1996); methodological investigations focus upon the research methods used across a number of studies within the same topic or in the same discipline (e.g. Kolbe & Burnett 1991; Pitt et al. 2005; van der Merwe et al. 2007); and specific journal investigations provide an in-depth review of one or more publications (e.g. Leong 1989; Malhotra 1996). Citation analyses are more concerned by the references/footnotes listed in papers (e.g. Baumgartner & Pieters 2003) and co-citation networks (Pasadeos et al. 1998). This study offers a combination of publishing productivity, comprehensive review and methodological investigation of a specific journal.
Method
On balance, a content analysis was deemed preferable to a survey of the Editorial Advisory Board. A content analysis of publications provides unobtrusive ex post facto evidence of the predilections of authors, reviewers and editors. A largely objective content analysis approach was also possible as most of the variables did not require judgemental coding. These elements were top publishing authors, top publishing author institutions, number of authors, domicile, occupation (academic or practitioner), number of words, sample and primary analytic technique. Author gender proved slightly more difficult to ascertain in cases where there was ambiguity over the first name; this was largely resolved by a careful reading of profiles, which generally included references to he or she, and a close inspection of Asian names by a Chinese research assistant. The grouping of topics was the most subjective aspect of the study given the potential multiplicity of categories. For example, Ewing et al.s (2002) study of the generalisability and equivalence of advertising response scales might be listed as cross-national, television medium, copy
544
DIRECTIONS IN MARKETING COMMUNICATIONS RESEARCH
testing and global campaigns. To address the problem, it was decided to categorise each paper by the major topic classifications and to address the issue of specific topics by a word content analysis of paper titles. Five topic classifications were adapted from Yale and Gilly (1988) and coded by a research assistant. These were: advertising practice (advertising agency, advertising management issues (e.g. budgeting, reach, sales and response), ambient, brand, campaign planning, creative management, direct marketing, e-business, media, and general industry concerns), marketing, new media, pricing, PR, promotions, sponsorship, web, wireless; advertising effects (appeals, attention, awareness, attitudes, consumer behaviour, effects, information processing, involvement, neuroscience, perception, persuasion, preferences, recall, recognition, response); social (children, culture, deception, economy, ethics, food, gender, pharmaceuticals, politics, race, socio-economic, regulation); content (appeals, comparison, copy, corporate, creative, image, information); and method (emphasis on research method). The process began with a full briefing of the research assistant, who then coded a random sample of 20 papers that were checked by the author. A small number of ambiguous categorisations were highlighted by the research assistant and these were resolved by further careful reading by both the assistant and researcher. Several new sub-codings were introduced in an iterative fashion as fresh topics emerged over the period, most notably concerning new media. None of this affected the main codings, as for example papers focused on digital media remained part of the media and thus advertising practice. Given the expertise developed in the process, the first 10% of papers coded were re-evaluated in light of the experience developed. Lastly, all papers were independently reviewed by the author for final classification. As an aside, it had been intended to apply Rust and Cooils (1994) proportional reduction in loss (PRL) measure for inter-judge agreement/reliability, but surprisingly the process proved to be more of a dialogue. WordStat was then used to examine the specific words used in paper titles to identify changes in topic over the period and provide for the more specific identification of minor rather than major topics. The past 15 years were chosen rather than the entire 25, as the purpose of the study was not to trace the evolution of the IJA and place it in its historical context but rather to trace the recent key trends in marketing communications research. Research notes and observations were not included in the study.
545
Inputs
As noted by Inkpen and Beamish (1994) and Malhotra (1996), any study of research published in a journal would be incomplete without an analysis of authorship. The top publishing authors those with more than three publications over the period 19922006 can be seen in Table 1 (figures are unadjusted for multiple authorships). The table is led by Michael Ewing, with eight publications, followed closely by Tim Ambler. There is then a small group of three consisting of Berthon, Pitt and Prendergast with five each, six authors with four publications each, and nine with three.
Table 1: Top publishing authors in the IJA 19922006
Author Ewing, Michael Ambler, Tim Berthon, Pierre Pitt, Leyland Prendergast, Gerard Gendall, Philip Hansen, Fleming Number of publications 8 7 5 5 5 4 4 Author Hoek, Janet Laroche, Michel Lee, Wei-Na West, Douglas De Pelsmacker, Patrick Dibb, Sally Duffy, Martyn Number of publications 4 4 4 4 3 3 3 Author Number of publications 3 3 3 3 3 3
Fay, Michael Kaynak, Erdener Meenaghan, Tony Palihawadana, Dayananda Simkin Lyndon Svennevig, Michael
Author institutions publishing in the IJA during the period 19922006 can be seen in Table 2. Leeds University tops the list with 21, followed by the University of Texas at Austin with 17 and Hong Kong Baptist University with 15. These three were closely followed by the Copenhagen Business School, Massey University and the University of Warwick, each with 14. Looking at the trends over the three five-year periods (not shown in the table), the top institutions from 199296 were Massey, Warwick, Hawaii and Baruch College; 19972001 saw the rise of Leeds, Pennsylvania and the London Business School, with a continued strong presence from Warwick; 200206 saw the rise of the Copenhagen Business School, Hong Kong Baptist, Texas at Austin and Monash. Of course, the presence of several of these institutions largely reflects the domicile of the IJAs top publishing authors.
546
Table 2: Top publishing institutions in the IJA 19922006

Institution Number of publications 21 17 15 14 14 14 11 10 Institution Monash Pennsylvania State Concordia Auckland Curtin Baruch College, City Calgary Florida Manchester Number of publications 10 10 9 7 7 6 6 6 6 Institution Miami New York (NYC) Nottingham Antwerp Bentley College California, San Diego Chinese, Hong Kong Hawaii Otago Number of publications 6 6 6 5 5 5 5 5 5
Leeds Texas, Austin Hong Kong Baptist Copenhagen Business School Massey Warwick Georgia London Business School
A breakdown of the number of authors, their domicile, gender and occupation can be seen in Table 3. Taking the period as a whole, both single authors and joint authors accounted for just under two-fifths of the papers published. Three-author papers provided slightly under a fifth, and work authored by four or five was insignificant. Taking the three five-year periods separately, it can be seen in the table that there is a sizeable shift away from single authorships (24%) towards joint (45%) during 200206, and a continual rise in papers authored by three people (from 14% to 26%). Unfortunately, past inconsistencies in recording practices make studies of author domicile (location of institution) somewhat difficult, so it was not possible to ascertain domicile in all cases and was especially difficult in the period 19972001. As such, Table 3 shows that North America provided the largest share (overwhelmingly US scholars) followed by Asia, the UK, Europe and then Australasia. As far as can be reliably ascertained, crossnational research accounted for around 10% of studies, with some consistency over the period. The rest of the world (in essence South America and Africa) provided an insignificant number of papers. It can be seen in Table 3 that the breakdown by gender was broadly three-quarters male and one-quarter female over the entire period (2% of authors could not be verified). Without apparent explanation, the balance shifted slightly in the direction of female over the period 19972001, with women accounting for slightly over a quarter, but this fell back slightly over 200206.
547
Table 3: Number of authors, their domicile, gender and occupations 19922006 by percentages and five-year periodsa
19922006 [n = 348] Number of authors One Two Three Four Five Total Domicile North America Asia UK Europe Australasia Rest of world Cross-national Not available Total Gender Male Female Not available Total Occupation Academic Practitioner Both Not available Total
a
199296 [n = 116] 42 41 14 3 0 100 26 11 12 10 5 0 8 28 100 78 21 1 100 79 10 10 1 100
19972001 [n = 126] 46 31 17 5 1 100 5 14 13 6 5 2 10 45 100 69 27 4 100 82 10 7 1 100
200206 [n = 106] 24 45 26 5 0 100 23 18 15 14 10 1 10 9 100 77 23 0 100 85 2 11 2 100
38 38 19 4 1 100 18 14 13 10 7 1 10 27 100 74 24 2 100 82 8 9 1 100
Where appropriate, figures have been rounded up.
Finally, turning to occupation in terms of academic versus practitioner, the papers in the IJA were largely authored by academics, who provided over four-fifths of all publications. Taking the three five-year periods, solely authored practitioner publications appear to be in decline, with only 2% of all publications during 200206. On the face of it this might be a result of the raising of the entry barriers with increasingly rigorous review processes that many practitioners, despite being capable, simply lack the time to contemplate. However, a personal observation since taking up the post of Editor in mid-2004 is that hardly any sole-practitioner papers are submitted.
548
One hopeful sign on this general issue (which also concerns many other journals and academics) is that joint academicpractitioner papers, which had dipped slightly over 19972001, rose to just over 10% during 200206.
Outputs
The average length of the 348 papers included in the study for 19922006 are given in Table 4. It can be seen that the mean hovered just over 5000 words between 1992 and 2001 but increased to over 6000 in the period 200206 for an overall mean of around 5500 words for the entire period (median of 6411 and standard deviation 1743). Following Yale and Gillys (1988) categorisation (see above or the notes to Table 4), the main topic covered was marketing communications practice, which accounted for around half of all the papers published. Studies on effects came next with just over a fifth of the published output, followed by social topics. It should be noted that social topics have taken an increasing proportion of the IJAs output, reaching nearly a quarter and equal to effects in the last five years of the period. Content studies slipped considerably in their share over the period and papers devoted to method were largely insignificant. Around 70% of papers utilised empirical data over the entire period. As shown in Table 4, the primary unit of analysis was practitioners, with just under a fifth of published studies, just ahead of secondary data and adults. Student samples provided a tenth of studies, closely followed by advertisements and then a small number of studies focusing on children. Looking at the trends over the three periods, the rise of student samples to just under a fifth has been the most notable. This contrasts with the general decline in the number of studies analysing advertisements. One other striking feature is the large number of other (e.g. discursive and observational) and non-empirical work published between 1997 and 2001. Such work had fallen back to under a fifth by 200206. Of those studies employing empirical data, the primary method employed was the survey, which accounted for just under two-fifths of the total, followed by content analysis, laboratory/experimental studies and secondary, with case and qualitative studies very much in the minority. Looking at trends, again the period 19972001 provided some stark contrasts, with rapid falls in content analyses and relatively large rises in
549
Table 4: Outputs 19922006 by percentage and five-year periods

19922006 [n = 348] Word count Mean Median Standard deviation Total Topica Practice Effects Social Content Method Other Total Sample Practitioners Secondary data Adults Students Advertisements Children Other/non-empirical Total Method Survey Content Lab/experiment Secondary Case Qualitative Other Total Analysis Descriptive Means Correlation/regression ANOVA/MANOVA Factor, discrim., cluster CFA/SEM Qualitative Other None Total 5,586 5,314 1,820 1,921,826 49 21 17 6 1 6 100 17 16 14 10 9 2 32 100 37 15 9 6 5 3 25 100 20 14 12 12 2 2 2 5 31 100 199296 [n = 116] 5,233 4,855 1,904 607,134 45 20 12 13 2 8 100 16 15 15 8 15 1 30 100 35 21 2 2 5 8 27 100 18 19 13 3 3 2 3 7 32 100 19972001 [n = 122126] 5,329 5,152 1,645 650,135 58 19 14 1 0 8 100 17 15 6 6 5 2 49 100 35 11 9 6 8 2 29 100 22 14 11 10 2 2 1 1 37 100 200206 [n = 106] 6,269 6,411 1,743 664,557 41 25 25 4 2 3 100 19 18 20 18 7 2 16 100 43 14 14 10 3 5 11 100 22 11 13 17 4 4 2 10 17 100
Notes: a = advertising effects (appeals, attention, awareness, attitudes, consumer behaviour, effects, information processing, involvement, neuroscience, perception, persuasion, preferences, recall, recognition, response); social (children, culture, deception, economy, ethics, food, gender, pharmaceuticals, politics, race, socio-economic, regulation); content (appeals, comparison, copy, corporate, creative, image, information); and method (emphasis on research method)
550
laboratory/experimental studies and secondary, with significant falls in qualitative. It should be emphasised that each paper was examined to elucidate the primary method. For example, a large proportion of the authors applying surveys used qualitative methods to underpin the work whereas this was coded solely as survey rather than qualitative. Turning to analytic technique, by far the most popular has been descriptive statistics (e.g. frequencies, percentages and cross-tabulations), accounting for a fifth of all analyses over the period 19922006 (see Table 4). Comparison of means came next, closely followed by correlation/ regression and analysis of variance (ANOVA) and multivariate analysis of variance (MANOVA). Factor, discriminant and cluster analysis, confirmatory factor analysis (CFA) and structural equation modelling (SEM), and qualitative analyses comprised relatively few of the analyses undertaken, with around 2% of studies respectively. In terms of trends, descriptives tended to largely maintain their share, whereas papers using simple comparisons of means declined, with rises in ANOVA/MANOVA, factor, discriminant, cluster and CFA/SEM. Akin to method, it should be noted that only the primary method has been reported here and a large number of papers utilised multiple techniques. To provide further insight into the specific topics covered in the IJA over the period, a frequency analysis of the words used in paper titles was undertaken. Only words that were deemed to add to the analysis were included, therefore advertising (appearing 212 times) and marketing (31 times) were removed, along with United States (21), China/Chinese (17) and UK (13), and all extraneous words irrelevant to topics were filtered out of the analysis. It can be seen in Table 5 that studies examining effects dominated paper titles for most of the period, with a frequency of 47 between 1992 and 2006. The list is inevitably a little jumbled, being the output of a straight word count. However, it is clear that advertising practice and effects are heavily implied in the words used, as indicated by effects, television and brands, along with some social issues indicated by tobacco and children. As for trends, the most striking elements indicated are the rise and relative fall of branding and the rise of food, the web and promotion. Inevitably, any firm conclusions are circumspect given keywords do not necessarily convey the meaning of the paper.
551
Table 5: Top keywords in paper titles, 19922006 and by five-year periodsa

19922006 Effect(s) Television/TV Brand(s) Agency(ies) Attitudes Consumer Media Food Market Sponsorship Information Perceptions Tobacco Web
a
199296 47 26 25 24 19 18 16 14 12 12 10 10 10 10 Effect(s) Agency(ies) Television/TV Attitudes Tobacco Consumer Media Information Perspective Expenditures 12 12 11 9 6 5 5 5 5 5
19972001 Brand(s) Effect(s) Agency(ies) Television/TV Sponsorship Perceptions Attitudes Media Market Children Responses Social 14 9 8 6 6 6 5 5 5 5 5 5
200106 Effect(s) Television/TV Consumer Brand(s) Media Food Web Promotion 26 9 9 8 6 8 5 5
The cut-offs are arbitrary but reflect points where the lists are both manageable and meaningful
Conclusion
This review of the inputs and outputs of the IJA over the period 19922006 was sparked by the 25th anniversary of the journal. A total of 348 papers were surveyed using a content analysis, which provides researchers with a better sense of what they have been contributing to the journal over the past 15 years. Journals can only select papers from those submitted, therefore this review says as much about what interests researchers as it does about the predilections of editors and reviewers. The IJA emerges as a journal largely focused upon topics involving practice and effects based upon samples of practitioners, secondary data and adults. However, student samples have been regularly published and have considerably grown in number during 200206. Single-authored papers appear to be in decline in favour of two-authored papers, with a strong trend towards papers written by three authors. This may reflect both the highly productive nature of teams and the relative ease of working with new technologies. Author institutions are widely domiciled across North America, Asia, the UK and Europe, with a lesser but significant presence from Australasia. While there has been a rise in the use of more sophisticated statistical techniques (e.g. discriminant and structural equation
552
modelling), papers employing basic statistical techniques continue to make up about a fifth of the output, perhaps as these often communicate managerially friendly information. Papers are getting longer and recent emerging topics include food, new media and promotions. It is hoped that this review of work published in the IJA over the period 19922006, undertaken to celebrate its 25 years of publishing, will have provided readers and authors with a better sense of the journal and a small sense of the current trends in marketing communications research in general, despite the limitations of the single journal sample. But what of the future? It is the job of authors to decide what topics they wish to pursue and submit to the IJA for, as with all other journals, the inputs and outputs of the journal are the product of the overall research community. Having said that, it should be pointed out that the journal has a public policy/social marketing orientation, therefore topics such as marketing communications effects related to issues such as obesity and alcohol consumption are particularly welcomed, along with papers addressing the full spectrum of marketing communications tools (i.e. not just advertising). Finally, in terms of method, it appears that qualitative work and field experimentation are under-represented in the IJA.
Acknowledgements
The author would like to thank Yan Mei for her research assistance and Matthew Coombs, Colin McDonald, and Stephen Whiteside for their valuable advice on previous versions of this paper.
References
Arndt, J. (1986) Paradigms in consumer research: a review of perspectives and approaches. European Journal of Marketing, 20(8), pp. 2340. Barry, T.E. (1990) Publication productivity in the three leading US advertising journals: inaugural issues through 1988. Journal of Advertising, 19(1), pp. 5260. Baumgartner, H. & Pieters, R. (2003) The structural influence of marketing journals: a citation analysis of the discipline and its subareas over time. Journal of Marketing, 67(2), pp. 123139. Crouch, G.I. (1996) Demand elasticities in international marketing: a meta-analytical application to tourism. Journal of Business Research, 36(2), pp. 117136.
553
Ewing, M.T., Caruana, A. & Zinkhan, G.M. (2002) On the cross-national generalisability and equivalence of advertising response scales developed in the USA. International Journal of Advertising, 21(3), pp. 323343. Inkpen, A.C. & Beamish, P.W. (1994) An analysis of twenty-five years of research in the Journal of International Business Studies. Journal of International Business Studies, 25(4), pp. 703713. Kolbe, R.H. & Burnett, M.S. (1991) Content-analysis research: an examination of applications with directives for improving research reliability and objectivity. Journal of Consumer Research, 18(2), pp. 243250. Leong, S.M. (1989) A citation analysis of the Journal of Consumer Research. Journal of Consumer Research, 15(4), pp. 492497. Malhotra, N.K. (1996) The impact of the Academy of Marketing Science on marketing scholarship: an analysis of the research published in JAMS. Journal of the Academy of Marketing Science, 24(4), pp. 291298. Pasadeos, Y., Phelps, J. & Bong-Hyun, K. (1998) Disciplinary impact of advertising scholars: temporal comparisons of influential authors, works and research networks. Journal of Advertising, 27(4), pp. 5370. Pitt, L.F., Berthon, P., Caruana, A. & Berthon, J.-P. (2005) The state of theory in three premier advertising journals: a research note. International Journal of Advertising, 24(2), pp. 241249. Rust, R.T. & Cooil, B. (1994) Reliability measures for qualitative data: theory and implications. Journal of Marketing Research, 31(1), pp. 114. van der Merwe, R., Berthon, P., Pitt, L. & Barnes, B. (2007) Analysing theory networks: identifying the pivotal theories in marketing and their characteristics. Journal of Marketing Management, 23(3/4), pp. 181206. Yale, L. & Gilly, M.C. (1988) Trends in advertising research: a look at the content of marketing oriented journals from 19761985. Journal of Advertising, 17(1), pp. 1222.
About the author

Douglas West is Professor of Marketing at the Birmingham Business School, the University of Birmingham, and editor of the International Journal of Advertising. He has published work in the International Journal of Advertising, the Journal of Advertising, and the Journal of Advertising Research, among others. He is joint author of Direct and Interactive Marketing (Oxford University Press, 2001) and Marketing Strategy: creating competitive advantage (Oxford University Press, 2006). Address correspondence to: Douglas West, Birmingham Business School, University of Birmingham, University House, Edgbaston Park Road, Edgbaston, Birmingham B15 2TT, UK. Email: d.west@bham.ac.uk
554
International Journal of Advertising CALL FOR PAPERS

Special Edition on Advertising, Markets & Demand
Guest Editor
Tim Ambler, Senior Fellow, London Business School
The debate surrounding the effect of advertising in mature, established markets has a tendency to polarise views. Some argue that brand advertising within mature markets (for example, fmcg/cpg) works at brand level within discreet product categories. That is, advertising enables brands to compete against each other within specific categories but it is a zero-sum game that does not affect the overall size of the category in which it operates. Others may concede that brand advertising predominantly concerns brands, but such advertising has a wider and cumulative effect on the overall size of the category. In other words, rather than just creating consumer demand for one brand at the expense of another, advertising is seen as creating demand for and, therefore, growing the size of the overall category. Part of the debate depends on the definition of the market. For example, brand advertising may affect the size of the spirits market but not alcoholic drinks taken as a whole. And then we have to consider category marketing such as the classic drink more milk campaign. Milk as a whole would not usually be considered to be a brand. And what is the effect on markets as a whole, or mature markets as a whole? To what extent is demand as a whole driven by advertising? The issues encompassed by this debate lie at the heart of public policy discussions about the extent to which advertising should be restricted, most notably tobacco, alcoholic drinks and, in relation to obesity concerns, food. This special issue of the International Journal of Advertising (IJA) seeks to gather together the latest research evidence on the effect of advertising in mature markets to help academics, advertising practitioners and all those involved in the public policy process better understand the intricacies of an inevitably complex debate. Empirical and/or theoretical papers are invited by the IJA on the effect of advertising in mature markets overall as well as the effect of advertising in specific mature markets (for example, alcoholic drinks, food etc). Papers selected for this special issue will be eligible for the annual Best Paper Prize of $1000, as voted by the Editorial Board.
Submissions The closing date for submissions is 15 February, 2008.

All submissions must conform to the guidelines of the International Journal of Advertising. For details see www.InternationalJournalofAdvertising.com. Manuscripts should be sent to the guest editor at the following address: IJAspecial@warc.com. For all enquiries contact the Production Editor, International Journal of Advertising World Advertising Research Center,Farm Road, Henley on Thames, Oxfordshire, RG9 1EJ Tel: 01491 418688, Email: ija@warc.com www.internationaljournalofadvertising.com
Smart. Thinking.
Admap
Subscribe to the most useful and stimulating marketing magazine in the world.
Order online now at www.admapmagazine.com
Comments
Edited by Professor John Ford
Old Dominion University, Virginia, USA
The subject for this issues Comments section is international advertising issues and challenges. Two wellrespected international advertising researchers have provided commentaries, which I believe will be of interest to our readers. Charles R. Taylor, from Villanova University, provides a commentary that offers a series of suggestions for improving the quality of international advertising research. His helpful recommendations include: (1) building international collaborations; (2) offering multiple contributions in the same article; (3) providing sounder theoretical foundations; (4) employing theories with broad strategic foci; (5) utilising tools and techniques that allow the examination of associations and causality; and (6) demonstrating the practical benefit of the research for both advertising researchers and practitioners.
The other commentary, by Barbara Mueller of San Diego State University, looks at corporate social responsibility and the marketing of food products around the world. Childhood obesity is becoming a global problem, some believe exacerbated by food advertising to children. She makes suggestions regarding the ways in which food marketers can act in a more ethical and socially responsible manner. Barbara also suggests that consumers themselves have certain responsibilities, including: (1) educating themselves more effectively about diet and nutrition; (2) being more active in their lifestyles; and (3) making sound decisions on healthy food choices for themselves and their children. As always, we hope that you will find these essays both informative and valuable. We welcome your comments at ijacomments@warc.com.
Overcoming barriers to publishing international advertising research in top journals Charles R. Taylor
In spite of recognition of the globalisation of business over the past several decades, it seems that international
advertising has never quite been a hot topic among academic researchers. A content analysis of international advertising articles conducted by Zou (2005) showed that only a very limited number of articles were published in some highly prestigious general marketing and business outlets, such as the Journal of Consumer Research, Journal of
557
Marketing, Journal of International Business Studies and Journal of Marketing Research. Moreover, the number of SSCI citations for international advertising articles was somewhat less than one might expect. While many articles that advance international advertising appear in prestigious journals specialising in advertising or international marketing, it is, nonetheless, important to scholars who choose to work in an area to know that their best work can be published in top general journals. The question as to why international advertising research has historically seemed somewhat less than glamorous, is an important one to ask. It seems the answer lies in a combination of factors, ranging from the greater difficulty of applying sophisticated theories and methodologies in crossnational studies to the need for more interaction from scholars from around the world who are interested in the topic. I would like to spend the remainder of this article outlining why I believe the time is ripe to see more truly outstanding cross-national advertising studies published and what can be done to facilitate this process. We are already seeing the beginnings of a trend in which a greater number of papers on international advertising that have strong theoretical grounding and apply advanced methodologies are being published. However, I believe that this trend could be accelerated if researchers avoid some past obstacles at each stage of the research process. I will organise my thoughts on this matter based on the typical sections of an academic paper, starting with author information.
Author information It has been known for some time that more perspectives from all over the world are needed in order to advance our knowledge of international advertising. Clearly, more truly international collaborations, in which co-authors collaborate on more aspects of the research than just data collection, are needed. Joint effort in conceptualising the paper, developing theory, ensuring equivalent data are collected, and writing up implications can lead to richer perspectives that help to advance the field. The good news with respect to collaborations is that it is probably easier to find co-authors than ever before. In addition to more communication via the internet, networking opportunities are available at the meetings of organisations such as the European Academy of Advertising, the American Academy of Advertising and the IAA. Each has a group of consistent attendees who are interested in international issues. I believe it is particularly important for those interested in international advertising to attend meetings outside of their home region in order to hear different perspectives and meet prospective collaborators. Abstract and introduction In conceptualising papers on international advertising, I would urge authors to consider the need to make multiple contributions to the literature in a single paper. I have often heard reviewers of international papers complain that authors are just applying what has already been learned to another context (often a single country), rather
558
COMMENTS
than focusing on theory development and/or or truly insightful managerial implications. Notably, international studies that have public policy implications seem to be in short supply. Studies that make a strong case that they are making multiple contributions are at an advantage in leading journals. Literature review A clear problem in the study of international advertising has been the large number of papers that start out by reviewing the standardisation vs local adaptation debate and then frame the study in this context. This debate now represents a false dichotomy that does not do justice to the insights provided by the extant literature. More importantly, this approach is an inadequate conceptual framework for true theory development, but is often used in this manner. A good literature review should originate as a means of truly summarising the current state of knowledge in an area and should also clearly set the stage for the conceptual framework that will guide the study. Theoretical framework Perhaps the greatest obstacle to international advertising work appearing in the most prestigious journals is the perception that much work in the area has not been theoretically sophisticated. It is true that many international studies have focused on isolated cultural dimensions as opposed to elaborate, sophisticated theories. Developing theory that incorporates cultural aspects will remain important, but it should not be the sole focus of research streams.
Of particular importance in terms of publishing in general journals is applying theories with a broad strategic focus. The good news for international advertising scholars is that, more than ever before, theoretical advancement in related areas is providing new opportunities. For example, global marketing strategy theory (Zou & Cavusgil 2002) and global consumer culture theory (see Arnould & Thompson 2005) are promising frameworks that have components that are applicable to advertising. These perspectives, along with applying insight from strategic management perspectives such as I/O theory and the resource-based view of the firm, hold considerable promise. Additionally, more applications of consumer behaviour theories to the advertising context, such as theory of reasoned action and the elaborations likelihood model, would be worthwhile. Methodology I have gone on record elsewhere (Taylor 2005) suggesting that a wider variety of methods should be used in international advertising research. In the past, content analyses often accounted for a high proportion of published studies. This appears to be changing, as the last few years have seen more survey and experimental research published. Nevertheless, there is room for more studies aimed at examining measures of association and examining possible causality. While there is still a place for innovative descriptive research, it is appropriately being held to a high standard.
559
Another issue is that when cultural dimensions are posited to have an impact on a dependent variable, they should be measured among the subpopulation that is sampled. Because some aspects of culture can be dynamic and some groups within an overall culture can vary, it strengthens a study examining culture if it actually measures the samples cultural characteristics. Results and conclusions The question of the degree of noise in international data sets has been a key obstacle to research getting published. While it has been recognised for some time that cross-cultural studies should apply advanced techniques to ensuring equivalence is established before data are collected, it now should be regarded as essential for them to make use of post hoc techniques such as confirmatory factor analysis or a Raaschebased technique to assure reviewers and readers that the data are clean. In terms of the results section as well as the conclusion, it is paramount to make the general contribution to academic knowledge clear to the reader. Here, demonstrating and reiterating the promised contributions in the introduction is essential. Postlude: cause for optimism There is little doubt that there is increasing awareness of what truly global perspectives can bring to academic research. I am encouraged by hearing Journal of Marketing editor Roland Rust speak about the priority he gave to internationalising his
Editorial Review Board because it was in the best interests of both the journal and the discipline. I believe it is no coincidence that the Journal of Marketing has concurrently seen its SSCI citation impact factor rise to new heights. It is also encouraging to see the increased prominence of the International Journal of Advertising in the field as its pages are sought after by scholars from around the world. In the face of such developments, which show academia recognising the true importance of internationalisation, it is important for those of us interested specifically in the international/cross-national/crosscultural aspects of the field to strive to do quality work that gets noticed by the field at large.
References Arnould, E.J. & Thompson, C.J. (2005) Consumer culture theory (CCT): twenty years of research. Journal of Consumer Research, 31(4), pp. 868892. Taylor, C.R. (2005) Moving international advertising research forward: a new research agenda. Journal of Advertising, 34 (Winter), pp. 716. Zou, S. (2005) Contributions to international advertising research: an assessment of the literature between 1990 and 2002. Journal of Advertising, 34(1), pp. 99110. Zou, S. & Cavusgil, S.T. (2002) The GMS: a broad conceptualization of global marketing strategy and its effect on firm performance. Journal of Marketing, 66(4), pp. 4056.
Charles R. Taylor is the John A. Murphy Professor of Marketing at Villanova University.
560
COMMENTS
Just where does corporate responsibility end and consumer responsibility begin? The case of marketing food to kids around the globe Barbara Mueller
Much has changed in the corporate landscape over the past few decades. A growing number of firms have come to recognise that it takes more than just engaging in ethical business practices and generating profits to navigate the minefield of todays global marketplace. Corporate social responsibility has become the buzzword of the day. Several factors are driving corporate social responsibility: the unending parade of corporate scandals that has rocked the business world, the resulting increase in regulatory scrutiny, the intense 24-hour-a-day worldwide media coverage, and, in particular, consumers increasing expectations of big business. As consumers, we rightfully expect that the manufacturers of the foods and beverages we consume apply the highest standards of ethical behaviour with regard to each element of the marketing mix. We assume that the products we purchase are both fresh and pure, that we are charged a fair price for these goods, and that they are made available to us where we require them. Moreover, we expect that promotional materials on behalf of these products are neither false nor misleading. But, increasingly, we expect marketers to go beyond this to additionally take on responsibility for our consumption behaviour. The
epidemic of obesity around the globe has put the spotlight on marketers of foods and beverages. In particular, those targeting children have come under heavy criticism. The statistics relating to obesity are hard to argue with. A recent European Union report shows that one in five EU schoolchildren is now overweight a total of 14 million a figure that includes at least three million obese (Lobstein et al. 2005). Things dont look much better across the pond. In the United States, since the 1970s, the prevalence of obesity has nearly quadrupled for children between the ages of 6 and 11 (from 4% to 15.3%), and more than doubled for youths aged 1219 from 5% to 10.4% (Kaiser Family Foundation 2004). However, childhood obesity is not unique to western countries. The prevalence of overweight and obese children in China increased 28 times between 1985 and 2002 (Walsh 2006). The finger for this problem has been pointed directly at food marketers. The food industrys global ad budget is a whopping $40 billion. Children around the globe are bombarded with an unprecedented avalanche of food advertisements. While television advertising remains marketers prime tool for selling food to kids (Spake 2003), there are a multitude of other promotional practices designed to make children desire specific products, including websites, inschool marketing and movie tie-ins. The concern is that the bulk of food products being aimed at kids are fast foods, soft drinks and snacks in short, foods high in calories, sugar and fat.
561
Governments and health advocates worldwide are cracking down on the marketing tactics they blame for the explosion in childhood obesity. In 2004, the UK Government published health recommendations giving food and beverage industries until 2007 to act more responsibly, or face formal legislation. In France, marketers of foods or beverages using radio or television are forced to choose between adding a health message to their commercials or paying a tax to fund healthy eating messages. Ireland has banned celebrities from food and beverage ads aimed at children, and requires that sweets and soft-drink spots broadcast in programmes with half the audience under the age of 18 carry a warning that snacking on such foods can damage teeth. Markos Kyprianou, EU Commissioner for Health and Consumer Affairs, also launched a Platform for Action against obesity. While emphasising that he prefers not to legislate, Kyprianou stated that he would not rule out the possibility of EU regulation of the food industry if nothing concrete comes out of the platform (Lobstein et al. 2005). In the US, the Federal Trade Commission recently issued subpoenas to 44 food manufacturers in order to gather information for a forthcoming report to Congress on how companies market so-called junk foods to kids. Congress also plans to hold hearings on the medias marketing of such foods to youngsters. The cause has been taken up by other bodies, as well: the Alliance for Childhood, the Campaign for a Commercial-free Childhood, and the
Center for Science in the Public Interest. The latter two groups have even threatened food marketers with lawsuits over their marketing tactics. While outright bans on marketing food to children have been debated in many countries, it is notable that both Sweden and Quebec, which have banned all advertising to children, still face high obesity rates. Some self-regulatory bodies, food marketers and advertising agencies have begun to take action in response to this changing environment. The International Chamber of Commerce recently issued guidelines on responsible marketing for food and beverage companies. In the US, major food companies as well as the American Advertising Federation, the American Association of Advertising Agencies, and the Association of National Advertisers formed the Alliance for American Advertising. The alliances purpose is to fight efforts to limit the marketing of foods to children, defend the industrys First Amendment rights to advertise to children, as well as to promote its willingness to police itself. The US advertising industrys selfregulatory body, the Childrens Advertising Review Unit which is charged with monitoring ads targeting children under the age of 12 announced intentions to enforce a new level of oversight for food ads. Individual marketers, as well, have taken steps. Kraft Foods announced it would stop running print and broadcast ads for products such as Oreos and Chips Ahoy!, aimed at 611 year olds. The Kellogg Company stated it would phase out advertising its products to
562
COMMENTS
children younger than 12 unless the foods meet specific nutritional guidelines. These voluntary changes will apply to about half the products that Kelloggs markets to children worldwide. Governments are also taking action. Self-regulatory bodies are drafting guidelines. And some food marketers are indeed attempting to take the high road by promoting their products in a more socially responsible manner or ceasing their promotion altogether. But what role does the consumer play in the growing epidemic of childhood obesity? What about consumer responsibility? Without question, exercise habits and overall activity levels play a direct role in weight management and maintenance. In the EU, one-third of children aged 211 participate in less than 60 minutes of activity per day, and half of those aged 611 receive less than two hours per week of physical education in school (International Obesity Taskforce 2002). The Centers for Disease Control reported that 25% of US children engage in no free-time physical activity at all (Jacques 2006). There are several explanations for the observation that children today are less active than those of previous generations among them the increased use of cars, a reduction in the time children spend playing outdoors, and the increase in sedentary activities, such as playing computer games and watching television. Today, young kids everywhere watch an inordinate amount of television. In Germany the average viewing time of children aged 213 is 1:31 hours a day, and among 1419 year olds, viewing totals 1:49 hours per day (Zubayr & Heinz 2006). In China,
youngsters watch an average of 1.8 hours of television per weekday and 4.2 hours on weekend days (Chan & McNeal 2004). Finally, in the US, it is estimated that 818 year olds spend about 3:04 hours per day watching the tube (Rideout et al. 2005). It must be pointed out that youngsters logging that much time on the couch are likely to become overweight, regardless of whether they are exposed to advertisements for chips or carrots. Clearly, one option for parents is to simply turn the television off and encourage their kids to go outside and get some exercise. Another factor related to obesity is nutritional intake which is ultimately a matter of personal choice. The US documentary film Super Size Me, which received tremendous media attention, chronicled the deterioration of filmmaker Morgan Spurlocks health during a month-long experiment eating nothing but McDonalds food. The filmmaker gained 24 pounds during the 30-day period by super-sizing his meals and consuming 5000 calories a day. Receiving significantly less coverage was an independent film entitled Me and Mickey D, which followed Soso Whaley as she spent 90 days on a McDonalds diet. Whaley dropped 36 pounds by eating 2000 calories a day at the fast-food restaurant. Notes Whaley, I couldnt just walk in there and say, Ill take a cinnamon bun and a Coke I know a lot of people are really turned off by the whole thought of monitoring what they are eating, but thats part of the problem. The real key was moderate exercise, and making healthy choices (quoted in Lopez 2005). Clearly, the problems
563
arent with whats on the McDonalds menu, but the choices made from it. At the start of his film Super Size Me, Spurlock asks where personal responsibility ends and corporate responsibility begins. Me and Mickey Ds creator Whaley has a brilliant response to this query: As far as Im concerned, personal responsibility for your own life never ends. Besides, I dont need some corporation to be responsible for me and my actions. The staggering increase in childhood obesity around the globe is cause for significant concern. However, to lay the bulk of the blame at the feet of multinational food marketers is unjust. The prevalence of obesity is multi-causal and combating it requires a multifaceted approach one that includes the marketer and the consumer. Greater corporate responsibility in the form of offering healthier products, promoting them via media to the appropriate age audience, with messages that contain balanced nutritional information that youngsters can understand is certainly called for. However, consumers must also take increasing responsibility by learning more about diet and nutrition, increasing activity levels for the whole family, and ultimately by making healthier food choices, both for themselves and their offspring. As is true for the marketing of other product categories, consumers cannot completely shirk their responsibility.
References Chan, K. & McNeal, J. (2004) Advertising to Children in China. Hong Kong: The Chinese University Press.
Food Facts (2004) Health groups warn: children at risk for junk food marketing. Nutrition and Food Science, 34(1), p. 42. International Obesity Taskforce (2002) Obesity in Europe: the case for action. London: European Association for the Study of Obesity. Available online at http://www.iotf.org/media/euobesity.pdf (accessed 10 September 2007). Jacques, J. (2006) Childhood obesity: causes and considerations. OAC News, 2(1). Available online at http://www.obesityaction.org/ resources/oacnews/oacnews5/coverstory.php (accessed 10 September 2007). Kaiser Family Foundation (2004) The Role of Media in Childhood Obesity. Washington: Kaiser Family Foundation, p. 1. Lobstein, T., Rigby, N. & Leach, R. (2005) EU platform on diet, physical activity and health. London: European Association for the Study of Obesity. Available online at http://www.easoobesity.org/docs/ report81.pdf (accessed 8 September 2007). Lopez, K.J. (2005) Soso, so good: can woman live on Mickey Ds alone? Available online at http://www.nationalreview.com/ interrogatory/whaley_200506230747.asp (accessed 10 September 2007). Rideout, V., Roberts, D. & Foehr, U. (2005) Generation M: Media in the Lives of 8 to 18 Year Olds. Washington: Kaiser Family Foundation. Spake, A. (2003) Hey kids! Weve got sugar and toys. US News and World Review, 17 November, p. 62. Walsh, B. (2006) No longer starving in China. Time, 18 August. Available online at http://time.bloggs.com/global_health/ 2006/08/no_longer_starv.html (accessed 16 February 2007). Zubayr, C. & Heinz, G. (2006) Tendenzen im zuschauerverhalten und fernsehgewohnheiten im jahr 2005. Media Perspektiven, 3, pp. 125137.
Barbara Mueller is a professor in the School of Journalism & Media Studies at San Diego State University.
564
Book Reviews
Edited by Dr Stephanie ODonohoe
The University of Edinburgh
Its not just PR: public relations in society W. Timothy Coombs and Sherry J. Holladay (2007) Blackwell Publishing, 17.99 The authors claim that this book provides a balanced view of public relations as neither flawless nor sinister, with practical debates about the power issues in PR theory and practice. They further claim that the book investigates corporate issues in PR, and how public relations can be used by marginalised and activist groups. I think the book does exactly these things, although at times I was disappointed with some of the content. While I would not use such glowing terms as those employed by Debra Worley and Doug Newson who both claim that the book shines a light on PR practice it does articulate the value of public relations from a PR practitioner perspective. For this reason alone, I would recommend it as a good read for all those in management, whether in the public, private, voluntary or the activist domain. I would certainly recommend it to my students at undergraduate, postgraduate and professional level, and it would be on the additional reading list for student consumption, although I would
recommend that it is read with quality books that give a more critical approach of the profession, such as LEtang and Pieczka (2006) or Moloney (2006). The book is generally well written, with good copy flow (in the early chapters). Chapter 1 provides an excellent overview of the negative perceptions of the PR profession, from which the authors develop their common themes in critiques of public relations. I liked the balance of material presented in the chapter on media content analysis, anti-PR book reviews, references to PR Watch, discussion of the models of public relations and a professional/academic defence of the industry. Chapter 2 provides a discussion on ethics in public relations. It is a wellconstructed chapter providing some interesting views and insights into professional regulations, and three general perspectives of teleology, deontology and the ethic of care. The authors also include more theoretical discussion on such issues as Grunigs models, dominant coalitions, boundary spanners, publics and stakeholders and the power of the PR professional. Chapter 3 starts with a comment on the history of American PR and its corporate-centric focus, before moving to the more interesting debate on
565
stakeholders and power. There is a very well explained analysis of stakeholder power and legitimacy, with some examples from Greenpeace and the American Environmental Protection Agency. The chapter then has 15 pages of historical case studies recalling the power of activism in American history. While I enjoyed this section (as a historian), some readers will ask why, in 2007, we still have to read about the Muckrakers, abolitionism and temperance. Some might also question why the only example of modern activism discussed was that of the American Family Association and their attack on the Ford Motor Company for alleged pro-homosexual activity. Chapter 4 is another chapter which created interest but some disappointment. There are very good discussions on issue management and social marketing theory, with good explanation of models such as Chase-Jones (1979), Crable & Vibbert (1985) and McGuire (1981). These discussions are supported with some interesting case studies. However, in this chapter, the authors move to a less formal writing style in the description of the 200506 Ready.gov website, and then introduce case studies and opinion on Nazi antiSemitic communication and the Nazi anti-tobacco campaign. The style and contrast of writing, and the bizarre mix of case studies, does tend to detract from the books contribution. Sadly, the final chapter on the global effects of public relations was descriptive, offering little new knowledge or
insight. The focus is public diplomacy, global activism and corporate social responsibility. Amazingly, it fails to discuss terrorism and the use of public relations by terrorists and governments. At the very least, I would have expected some discussion on terrorist propaganda, with hopefully some critical analysis on how new public relations has led to peace from conflict. So, in summary, an American PR text with a pro-PR stance. A book I would recommend to practitioners and students, but not to fellow academics. It is interesting and informative in some chapters, but let down by differing styles of writing and some dated case studies.
References LEtang, J. & Pieczka, M. (2006) Public Relations: Critical Debates and Contemporary Problems. Mahwah, NJ: Lawrence Erlbaum Associates. Crable R.E. & Vibbert, S.L. (1985) Managing issues and influencing public policy. Public Relations Review, 11(2), pp. 315. McGuire. W.J. (1981) Theoretical foundations of campaigns, in Rice, R.E. & Paisley, W.J. (eds) Public Communication Campaigns. Newbury Park, CA: Sage. Moloney, K. (2006) Rethinking Public Relations: PR Propaganda and Democracy (2nd edn). London: Routledge.
Andrew Purcell School of Communication, University of Ulster; Fellow of the Chartered Institute of Public Relations.
566
BOOK REVIEWS
The New Influencers: A Marketers Guide to the New Social Media Paul Gillin (2007) Quill Driver Books, 12.50 As a marketing academic, I am consistently frustrated by popular press (or airport books as we snobs in academia call them) about marketing, particularly lately due to the reality that marketing as a business function seems to be morphing into something almost foreign compared with the traditional notions that I learned as a student and now impart as a professor. I find myself having two responses to these books and these changes: (1) am I just getting old and fear change?; and (2) the changes happen so fast and furious these days that I find it impossible to keep up with them all. Im like the old lady trying to relive her double dutch1 days, but unable to find the hole in the feverishly swinging jump ropes. Im left on the sidelines waiting for my opportunity to jump into the fray, into the discussion about these great new tools and tactics, while the younger, thinner kids slide in effortlessly and just start jumping. It is with this previously held level of annoyance for practitioner books about marketing that I began to read The New Influencers by Paul Gillin. And immediately the book began to annoy me. Consumers were referred to as passive on the first page and the ageold misconception that marketing serves merely a message communication
1
Double dutch is the sport of jump roping using two jump ropes going simultaneously.
function crops up early and often. Most practitioner-authors make the mistake of using the term marketing when in fact they are referring to advertising or public relations. They relegate the job of the marketer to communicator and manager of the message while ignoring all the other functions that marketing performs such as setting prices, brand management, distribution strategy, and environmental monitoring. You would be hard pressed to find one of these authors state that Southwest Airlines success is largely due to the fact that they use the same plane for every flight (the 737), which is just as much a marketing strategy decision as the fact that Southwest has good customer service policies and happy employees. Unfortunately, for me and my ilk, this general ignorance with the full functioning of marketing as a business function seems to be largely our (marketing academics) fault. The marketing function is consistently marginalised at both corporate and the business education levels, in favour of finance, accounting, and information systems, which dont seem to suffer the lack of understanding to which marketing consistently falls prey. Accountants may be dweebs but marketers make kids start smoking. In the ultimate stroke of irony, the field of marketing seems to suffer from bad PR. Having made my preconceived negative biases known, I must admit that this is a good book, despite its limited view of the role of marketing in the firm as well as some other minor flaws that I will of course also point out. In
567
fact, I enjoyed the book so much that I contacted its author and asked him if he would be willing to collaborate on future projects. I think more collaboration between marketing pundits and marketing academics would be beneficial to both sides. We would learn how to jump in faster and they would learn to widen their vision to include the whole scope of marketing. This book focuses on what Gillin calls the new social media, which refers to blogging, podcasting, social networking, and wikis. Within these new social media, there are Influencers and Enthusiasts. Gillins basic thesis is that marketers must learn to listen to these groups in order to better understand their customers, better tailor messages to relevant stakeholder groups, and better manage negative press or product failures. In traditional marketing circles, we call these people opinion leaders and have been professing for years that these consumers should be cultivated, listened to, plugged in to new product development, and otherwise considered partners in the process. That this new breed of opinion leaders have taken to the internet and have the potential to create blog swarms only magnifies the necessity for marketers to get and remain plugged in to this highly influential group of consumers. I fear marketing academics will read this book and say weve known that for years, whats new about it? My response to us (Im including myself, see previous comments) is that if weve known about it for so long why havent we been able to communicate it effectively enough
to MBAs and executive education participants so that they are actually implementing these tools? Gillin relates an amusing anecdote from Stowe Boyd, a well-known blogger, who said he took a social media dog and pony show on the road for an American Marketing Association press tour but dropped out early after it became clear that his audience wasnt interested (p. 205). He blames the lack of interest on marketers preoccupation with traditional lead generation and a feeling of being overwhelmed at the thought of jumping in to the world of blogs and podcasting. I would argue that the audience members werent interested because of the language barriers I discussed earlier. Marketers arent just about message delivery as is stated on p. 204; their responsibilities reach far wider. Finally, is the new social media an American-centric phenomenon? As I read the book, I kept in mind that I was reviewing the book for an international audience. It struck me more so than if I were reviewing for an American publication that almost all the examples and interviewee choices were American. It made me wonder if whether this social media phenomenon was decidedly American in terms of reflecting certain cultural values such as need for community, reaching out to strangers because of fractures within our physical social networks (high divorce rates, high mobility rates, etc.), or our overwhelming need to talk. I made a presentation to a knowledge management forum entitled Wikis are from Venus, Blogs are from Mars, outlining what I saw as very gendered
568
BOOK REVIEWS
communication styles reflected in these two tools. While Gillin makes the point several times that thou shalt comment, it is still up to the communicator (the blogger in this case) to engage the comments. He or she can choose not to. This communication pattern would be seen by Deborah Tannen and others as masculine because the conversation is still dominated by the speaker with the reader/responder still relegated to optional status.
In summary, I found this book very informative, easy to read and digest, and most importantly for me as an academic, stimulating in terms of how the field of marketing could be better served if only marketers, both academic and managers, were to embrace the potential of these New Influencers. Susan Dobscha Associate Professor of Marketing, Bentley College, Massachusetts
EX LIBRIS Malcolm White

Theres nothing as fantastic in my book as a book you cant put down. We all have our own personal favourites (mine is Wilkie Collins Woman in White), but how often is your personal favourite a business book? Is there anyone out there who has found themselves reading Malcolm Gladwells Tipping Point or Blink into the wee small hours? Anyone who has missed a meal to devour James Surowieckis The Wisdom of Crowds in one bite? Heard of anyone getting caught under the bedsheets, torch in hand, while finishing Freakonomics? Fine books though they are, I thought not. But there is a different type of business book that has the stuff it takes to be a personal favourite, for me anyway. They are books that I am always dipping into, that I can happily get lost in the thick of, and that I can browse for hours regardless of the passing hours. Books which help me and inspire me and excite me, and they do all of this for me at work. Given the kind of deep emotional engagement Im talking about, it might surprise you to know that Im referring to the humble reference book, or rather to a number of humble reference books. Reference books tend to get a bad press if you ask me. They are the workhorses of the book world. On the football pitch they would be the players player, not the gifted inside forward. Of course among their number would be counted the trusty dictionary and the well-thumbed thesaurus, but the ones Ive got in mind are much more special than this. First among my favourites is The Chronology of Words and Phrases: A Thousand Years in the History of English, by Linda and Roger Flavell. Unlike the magisterial Oxford English Dictionary, which starts from a particular word and then tracks down its
569
debut, The Chronology of Words and Phrases starts from the other end of the telescope. It starts from historical events. It then pinpoints the genesis of particular words, peculiar to their peculiar times. Take pandemonium for example, a word that makes its first appearance as Satans palace in Miltons Paradise Lost, and means dwelling place of all demons. A product of its time: from 1642 to 1651, England and John Milton were gripped by the turmoil, despair, chaos, the pandemonium you might say, of the English Civil War. Then theres the wonderful A Word in your Shell-Like: 6,000 Curious and Everyday Phrases explained, edited by Nigel Rees. Its packed full of famous catchphrases, ad slogans, and everyday phrases. Did you know, for example, that Sherlock Holmes never actually said Elementary, my dear Watson? Proverbial Wisdom from Around the World, edited by David Crystal, is my third choice. Theres much to enjoy here, including the stark warning from Denmark that She is a foolish woman who blames her own cabbage. But what links these three books together is that they remind us that words are magic. Words, mysteriously and magically, give shape to ideas. In my line of work, that makes these dull reference books unputdownable page-turners.
References Collins, W. (2003) The Woman in White (first published 1860). London: Penguin Books. Crystal, D. (2006) As They Say in Zanzibar: Proverbial Wisdom from Around the World. London: Collins. Dubner, S.J. & Levitt, S.D. (2006) Freakonomics: A Rogue Economist Explores the Hidden Side of Everything. London: Penguin Books. Flavell, L. & Flavell, R. (1999) The Chronology of Words and Phrases: A Thousand. Years in the History of English. London: Kyle Cathie. Gladwell, M. (2000) The Tipping Point: How Little Things Can Make a Big Difference. London: Little, Brown and Company. Gladwell, M. (2005) Blink: The Power of Thinking Without Thinking. London: Allen Lane. Milton, J. (2003) Paradise Lost (first published 1667). London: Penguin Books. Rees, N. (2004) A Word in Your Shell-Like: 6,000 Curious and Everyday Phrases Explained. London: Collins. Surowiecki, J. (2004) The Wisdom of Crowds: Why the Many Are Smarter Than the Few and How Collective Wisdom Shapes Business, Economies, Societies and Nations. London: Abacus.
Malcolm White is chairman of the Account Planning Group (APG), and a co-founder of the progressive communications agency krow.
570
Global economy and adspend prospects

Colin Macleod
Research Director, WARC
Within weeks of the OECD calling for an interest rate cut in the US, the Federal Reserve duly obliged in the middle of September with a halfpoint reduction the first in four years and the largest since 2002. The move was clearly designed to put an end to the crisis in the US sub-prime mortgage industry which has been threatening to spill over into the wider economy and is already beginning to impact elsewhere in the world. Meanwhile, the US housing market continues to remain depressed. The number of houses built in July fell to a ten-year low, while the number of repossessions had almost doubled in 12 months. The labour market has also taken a turn for the worse with the news that the number of new jobs created in August was 4,000 fewer than in July the first month-on-month decline for four years and forecasters are now downgrading their predictions for US economic growth this year. After increasing at an annual rate of 4% in the second quarter, the OECD now sees third-quarter growth slipping to 2%, and the final quarter to just 1.5%. For 2007 as a whole, the OECD is now forecasting 1.9% growth, compared with its May prediction of 2.1%. After a strong first quarter, eurozone GDP in the second was up just 0.3% quarter-on-quarter with growth in both France and Germany being well below expected levels. In Italy, the situation was worse, with the economy expanding by just 0.1%, compared with expectations of around 0.4%. Across the eurozone, growth should pick up in the second half of the year, but the so-called credit crunch has meant that banks here, and elsewhere, are becoming increasingly cautious when it comes to lending money. As a result, consumer spending, and company profits, are likely to
571
be hit. Business investment in the region has also weakened, slipping back in the second quarter after a 2% increase in the first. The slowdown in the economy has meant that the predicted 1.6% constant price growth in US adspend this year, forecast back in the spring, has now been cut to 0.7% in WARCs European Advertising and Media Forecast. By contrast, the increase in the German VAT rate at the start of the year has not had the expected negative impact on the economy and the 2007 adspend forecast here has been revised from a 0.7% decline to growth of 1.2%. The US economic climate should improve next year, and, helped by the presidential election, adspend growth there is forecast at 2% at constant prices.
572
GLOBAL ECONOMY AND ADSPEND PROSPECTS
Table 1: Annual % change in adspend at constant prices (US, Canada, Japan)

1996 1997 1998 1999 2000 US Canada Japan 5.6 3.6 5.8 4.9 7.2 2.8 6.6 8.6 4.3 6.5 1.3 1.8 2001 2002 2003 2004 2005 2006 2007 (f) 2008 (f) 0.0 0.1 5.9 1.2 3.6 0.2 4.9 3.4 3.2 0.3 2.7 1.6 1.4 6.0 0.1 0.7 3.9 1.1 2.0 3.4 0.8
10.0 10.8 3.0 0.3 8.0 1.1
Table 2: Annual % change in adspend at constant prices (UK, Germany, France, Italy)
1996 1997 1998 1999 2000 UK Germany France Italy 5.3 0.1 1.0 5.5 7.9 1.9 2.5 9.2 7.3 4.2 1.0 8.1 4.7 4.6 9.8 12.1 9.3 6.4 8.3 14.8 2001 2002 2003 2004 2005 2006 2007 (f) 2008 (f) 5.0 0.7 9.4 10.2 6.5 3.3 5.3 6.1 2.3 5.9 1.5 0.5 6.4 0.3 2.6 5.2 1.7 0.7 1.3 1.2 0.8 2.0 1.9 1.4 1.5 1.2 1.6 0.9 1.4 0.8 1.5 1.3
Source for all tables: WARC European Advertising and Media Forecast (www.warc.com/EAMF) Adspend figures exclude direct mail
573
Figure 1: Annual % change in adspend at constant prices (US, Canada, Japan)

% 10 8 6 4 2 0 2 4 6 8 10 12 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 (f) 2008 (f) US Canada Japan
Figure 2: Annual % change in adspend at constant prices (UK, Germany, France, Italy)
% 15 12 9 6 3 0 3 6 9 12 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 (f) 2008 (f) UK Germany France Italy
574
World Advertising Research Center

everything youll ever want to know about the advertising business
...all in one place.

See for yourself at
www.warc.com and take a 7 day free trial
International Journal of
CALL FOR PAPERS

Guest Editors
Ex 1 F tend eb ed ru de ar ad Advertising y 2 lin 00 e 8
Special Edition on the Advertising and the Brain

Dr Gemma Calvert, Reader in Cognitive Neuroscience, University of Bath, UK Dr Peter Kenning, Professor of Marketing, University of Muenster, Germany Dr Carl Marci, Chief Science Officer, Innerscope Research, USA Advertising works in two ways: it may trigger some immediate response and/or change the respondents brand memories in some way that influences later behaviour. This special issue of the International Journal of Advertising (IJA) seeks to draw on new findings from both neuroscience and psychology (N&P) to help academics, agencies and advertisers better understand how advertisements are processed, how brand memories are stored and the subsequent behavioural effects. The objective of the IJA special issue is to investigate how to deploy neuroscience and insights from psychology to produce advance MARCOMS theory and practice. Empirical and/or theoretical papers are invited by the IJA on any aspect of N&P in relation to advertising with a greater weighting towards neuroscience. For guidance, here are some suggested topics: 1. 2. 3. 4. 5. 6. N&P and how advertising works (e.g. contribution to brand equity). N&P and selective perception. Effectiveness of MARCOMS and N&P. N&P and the effectiveness of media planning. Negative messages (such as political advertising) compared to positive and N&P. N&P communications processing in different media (e.g. print, TV, radio, sales promotions). 7. N&P communications research methods best practice. 8. Case studies illustrating the practical use of N&P. 9. Interaction effects between media and N&P. 10. Ethical issues and implications in the use of N&P in the MARCOMS business. Please note that papers that are selected for this special issue will be eligible for the annual Best Paper Prize of $1000, as voted by the Editorial Board. Submissions All submissions must conform to the guidelines of the International Journal of Advertising. For details see www.InternationalJournalofAdvertising.com.
Manuscripts should be sent to the guest editors at the following address: IJAspecial@warc.com. The closing date for submissions is 1 February , 2008.
For all enquiries contact the Production Editor, International Journal of Advertising World Advertising Research Center,Farm Road, Henley on Thames, Oxfordshire, RG9 1EJ Tel: 01491 418688 Email: ija@warc.com www.internationaljournalofadvertising.com

Advertising Research

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Advertising Research

Uploaded by

Copyright:

Available Formats

ADVERTISING

The Quarterly Review of Marketing Communications

Volume 26 Number 4 2007

419 425 447

EDITORIAL ADVISORY BOARD

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

The academicianpractitioner gap in advertising

Directions in marketing communications research: an analysis of the International Journal of Advertising

The academicianpractitioner gap in advertising

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

Academic organisational structures

Practitioner knowledge utilisation

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

Summary and new directions

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE ACADEMICIANPRACTITIONER GAP IN ADVERTISING

About the authors

Subscription Order Form

I enclose a cheque for _____ made payable to WARC

Expiry date: Job Title

Fax No. Date

The Web Motivation Inventory

Frank Alpert The University of Queensland

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE WEB MOTIVATION INVENTORY

Internet motives: instrument evaluation

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

Table 1: Internet motives instrument evaluation

Instrument name Dimensions

User perceptions of commercial websites

Uses and gratifications perspective of websites

Journal of Business Research

Gratifications and concerns

Journal of Advertising Research

Internet motives & electronic media

Papacharissi & Rubin

THE WEB MOTIVATION INVENTORY

Table 1: Internet motives instrument evaluation (continued)

Instrument name Dimensions

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

Motivations for communication technology use

*SD = semantic differential scale, LIK = Likert scale

THE WEB MOTIVATION INVENTORY

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE WEB MOTIVATION INVENTORY

INTERNATIONAL JOURNAL OF ADVERTISING, 2007, 26(4)

THE WEB MOTIVATION INVENTORY

p * * * * * * * * ***