Assignment3 Campaign Critique for Youtube

American Marketing Association is collaborating with JSTOR to digitize, preserve and extend access to Journal of Marketing.

http://www.jstor.org

Building Brand Community Author(s): James H. McAlexander, John W. Schouten and Harold F. Koenig Source: Journal of Marketing, Vol. 66, No. 1 (Jan., 2002), pp. 38-54 Published by: American Marketing Association Stable URL: http://www.jstor.org/stable/3203368 Accessed: 27-07-2015 04:47 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

James H. McAlexander, John W. Schouten, & Harold F. Koenig

Building Brand Community A brand community from a customer-experiential perspective is a fabric of relationships in which the customer is situated. Crucial relationships include those between the customer and the brand, between the customer and the firm, between the customer and the product in use, and among fellow customers. The authors delve ethnographi- cally into a brand community and test key findings through quantitative methods. Conceptually, the study reveals insights that differ from prior research in four important ways: First, it expands the definition of a brand community to entities and relationships neglected by previous research. Second, it treats vital characteristics of brand com- munities, such as geotemporal concentrations and the richness of social context, as dynamic rather than static phe- nomena. Third, it demonstrates that marketers can strengthen brand communities by facilitating shared customer experiences in ways that alter those dynamic characteristics. Fourth, it yields a new and richer conceptualization of customer loyalty as integration in a brand community.

or decades, marketers have sought the Holy Grail of brand loyalty. Just as the legendary grail of Arthurian quest held the promise of extended life and renewal,

marketers attribute to brand loyalty and its sister icon, cus- tomer retention, the promise of long-term profitability and market share (Bhattacharya, Rao, and Glynn 1995; Reicheld and Sasser 1990). Unfortunately, marketing's knights-errant face a daunting problem: They have not fully understood what the grail looks like or where it can be found. As a result, marketers have devised strategies and designed pro- grams to build loyalty with limited information about their real impact or ultimate consequences (Dowling and Uncles 1997; Fournier, Dobscha, and Mick 1998).

To address this problem, we sought out places where we could find loyal customers, and we studied the processes that led to their loyalty. What we found were consumers forging and strengthening a variety of relationships. Many recent quests for the loyalty grail have ventured into the area of relationship marketing (see Garbarino and Johnson 1999; Gruen, Summers, and Acito 2000; Price and Arnould 1999). As a new imperative in marketing practice (Berry 1995; Deighton 1996; Gundlach, Achrol, and Mentzer 1995; Web- ster 1992), a focus on customer relationships is presented as an avenue to competitive advantage (Berry 1983, 1995; Kalwani and Naryandas 1995; Peppers and Rogers 1993). The relationships we observed were in many ways different from those that have been the focus of most prior research (see Fournier 1998). We found consumers and marketers jointly building communities. In exploring those communi- ties, we discovered new ways of understanding loyalty.

Marketplace Communities A community is made up of its member entities and the rela- tionships among them. Communities tend to be identified on the basis of commonality or identification among their mem- bers, whether a neighborhood, an occupation, a leisure pursuit, or devotion to a brand. What seems relatively self-evident about communities is the extent to which they are instrumental to human well-being. Through communities, people share essential resources that may be cognitive, emotional, or mate- rial in nature. Among all the things that may or may not be shared within any given community-things such as food and drink, useful information, and moral support-one thing seems always to be shared: the creation and negotiation of meaning.

With no more than a cursory look at contemporary soci- ety, we can identify communities whose primary bases of identification are either brands or consumption activities, that is, whose meaningfulness is negotiated through the symbol- ism of the marketplace. Scholars have grappled conceptually and empirically with such communities and have examined some of the dimensions that shape them (see Arnould and Price 1993; Celsi, Rose, and Leigh 1993; Fischer, Bristor, and Gainer 1996; Granitz and Ward 1996; Holt 1995; Moore, Mazvancheryl, and Rego 1996; Muniz and O'Guinn 1996, 2001; O'Guinn 1991; Schouten and McAlexander 1995).

Our understanding of marketplace communities begins with what Boorstin (1974, p. 89) describes as consumption communities, which he characterizes as "invisible new com- munities ... created and preserved by how and what men consumed." He observes that in the emerging consumer cul- ture that followed the industrial revolution, the sense of community in the United States shifted away from the tight interpersonal bonds of geographically bounded collectives and into the direction of common but tenuous bonds of brand use and affiliation:

The modern American, then, was tied, if only by the thinnest of threads and by the most volatile, switchable loyalties, to thousands of other Americans in nearly every- thing he ate or drank or drove or read or used. Old- fashioned political and religious communities now became only two among many new, once unimagined fellowships. Americans were increasingly held to others not by a few

James H. McAlexander and Harold F. Koenig are Associate Professors of Marketing, College of Business Administration, Oregon State University. John W. Schouten is Associate Professor of Marketing, School of Business Administration, University of Portland. The authors thank the people of Harley-Davidson Inc., Bozell Worldwide, and DaimlerChrysler for their time and support in this project. The authors also thank the associates and colleagues of Ethos Market Research and Diagnostic Research Interna- tional for their contributions to this research. The authors thank James C. Anderson, James R. Brown, Kent Eskridge, Key-Suk Kim, and the four anonymous JM reviewers for comments and suggestions on previous drafts of this article.

Journal of Marketing Vol. 66 (January 2002), 38-54 38 / Journal of Marketing, January 2002

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

iron bonds, but by countless gossamer webs knitting together the trivia of their lives. (Boorstin 1974, p. 148)

Visit Camp Jeep or a HOG (Harley-Davidson) rally. Par- ticipate in a Saturn Homecoming. Go to a DeWalt contrac- tors night at the local lumberyard. In each of these settings, and others, the so-called invisible consumption communities described by Boorstin (1974) suddenly become visible. Although we found Boorstin's concept of consumption communities attractive, in our own field research we dis- covered phenomena, such as subcultures of consumption (Schouten and McAlexander 1995), that more closely resembled his "iron bonds" than his "gossamer webs." We apparently were seeing a different kind of community.

Another kind of collective, a brand community, is defined by Muniz and O'Guinn (2001, p. 412) as "a specialized, non- geographically bound community, based on a structured set of social relationships among users of a brand." The communi- ties they describe in their insightful work are a better fit for the types of relationships we encountered in the field than were Boorstin's consumption communities. Muniz and O'Guinn's study of brand community, along with other work in the realm of consumer collectives (Holt 1995; Schouten and McAlexan- der 1995), indicates that intercustomer relationships figure importantly in the loyalty equation. In our research, we have found the emphasis on social relationships among customers to be correct but not entirely complete. Other entities and rela- tionships weave through the fabric of community.

A Broader View of Brand Community Muniz and O'Guinn (2001) envision a brand community as a customer--customer-brand triad. We suggest an extension of their model as well as a shift of perspective. Construing brand community as a social aggregation of brand users and their relationships to the brand itself as a repository of meaning (see Aaker 1996; Aaker 1997; Gardner and Levy 1955; Grubb and Grathwohl 1967) overlooks other relationships that supply brand community members with their common- ality and cultural capital (Holt 1998). Customers also value their relationships with their branded possessions (see Belk 1988; Holbrook and Hirschman 1982; Wallendorf and Arnould 1988) and with marketing agents (see Doney and Cannon 1997; Dwyer, Schurr, and Oh 1987) and institutions (see Arnould and Price 1993; Belk 1988; Bhattacharya, Rao, and Glynn 1995; Brown and Dacin 1997; Gruen, Summers, and Acito 2000; Morgan and Hunt 1994; Price and Arnould 1999) that own and manage the brand. Granting community- member status to the branded product and to the marketer sit- uates both the customer-brand dyad (the traditional focus of brand loyalty scholars) and the customer-customer-brand triad (Muniz and O'Guinn's [2001] elemental brand commu- nity relationship) within a more complex web of relationships (see Figure 1). We take the perspective that brand community is customer-centric, that the existence and meaningfulness of the community inhere in customer experience rather than in the brand around which that experience revolves.

The Dynamic Nature of Brand Community Research on consumption and brand communities identifies several dimensions on which they differ, including geo-

FIGURE 1 Key Relationships of Brand Community

Traditional Model of Customer-Brand Relationship

Customer Brand

Muniz and O'Guinn's (2001) Brand Community Triad

Brand

Customer a a Customer

Customer-Centric Model of Brand Community

Brand Product

Focal

Customer

Customer Marketer

graphic concentration, social context, and temporality. Typ- ically, these dimensions are treated as static identifiers in typological discussions (see Fischer, Bristor, and Gainer 1996; Granitz and Ward 1996; Tambyah 1996). Scholars have yet to fully examine these dimensions of brand com- munity as dynamic continua or shifting mosaics, yet this is necessary if they are to understand how the amorphous con- sumption communities described by Boorstin (1974) some- how coalesced into the visible, vibrant, and multifaceted brand communities that we encountered in the field.

Geography is one dimension on which communities dif- fer. Although brand communities have been defined as non- geographically bounded (Muniz and O'Guinn 2001), they may be either geographically concentrated (Holt 1995) or scattered (Boorstin 1974). They may even exist in the entirely nongeographical space of the Internet (Granitz and Ward 1996; Kozinets 1997; Tambyah 1996). Studies have tended to be situated statically on the dimension of geo-

Building Brand Community / 39

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

graphic concentration, even if they consider multiple geo- graphies. What can be learned if this dimension is treated dynamically? For example, how does a normally scattered brand community respond to temporary geographic concen- trations, such as in the case of a brandfest (McAlexander and Schouten 1998)?

Related to, but not dependent on, geographical concen- tration is the dimension of social context. Interactions within a brand community may be rich in social context or nearly devoid of it (Fischer, Bristor, and Gainer 1996). Communi- cation may be predominantly face to face, mediated by elec- tronic devices, or a function of corporate mass media (Boorstin 1974). Community members may have a great deal of information about one another, including such data as age, sex, attractiveness, and personal history, or they may know nothing of one another but pseudonymous "handles" and openly demonstrated topic knowledge (Granitz and Ward 1996). Studies of brand communities tend to be situ- ated statically with respect to social context. There is little understanding of movement along this dimension. What happens, for example, when mass-mediated brand commu- nities have the opportunity for context-rich relationships?

Yet another dimension of communities is their temporal- ity. Some are stable or enduring (Schouten and McAlexander 1995). Others are temporary or periodic (Arnould and Price 1993; Holt 1995; McGrath, Sherry, and Heisley 1993). The temporal stability of a community can be an asset to marketers inasmuch as longevity equates with a long-term, stable mar- ket. Still, even situational communities have been observed to share meaningful consumption experiences (Arnould and Price 1993; McGrath, Sherry, and Heisley 1993). What hap- pens to a temporary brand community after its situational rel-

evance ends? Is a brand community's temporality affected by changes in other dimensions, such as increased social context?

Another dimension that adds complexity to the study of communities is their basis of identification. Communities may be based on such wide-ranging commonalities as kin- ship ties, occupational connections, religious beliefs, or leisure pursuits. In the context of any person's life, some communities may overlap and interlock significantly. Others may represent separate arenas of activity. Scholars of brand community often neglect the effects of multiple community memberships. Interesting questions arise when the possibil- ity of interlocking community ties is recognized. For exam- ple, does commitment to a brand community increase (or decrease) if other members of a person's extended family, neighborhood, or work community also belong?

Research Program This multiple-method research program began with ethnog- raphy through which findings or themes emerged and devel- oped in a nonlinear fashion. We added quantitative support through a pretest/posttest design and structural equations analysis of a key conceptual model. Finally, we returned to ethnographic work for longitudinal and integrative perspec- tives (for a research time line, see Figure 2). Taken together, these methods and findings harmonized in a way that may best be described in terms of Price and Arnould's (1998) "conducting a choir" metaphor.

Ethnography A focus on two brands, Jeep and Harley-Davidson, and two ethnographic research programs served as the foundation of this research. An account of the methods used in the work

FIGURE 2 Research Time Line

?. ... .. ... ....,..?. . .-X r:::...::: .:: .:::...::. ., ,..*,,..., .*,...,, --,,* -,,,. ,,. **,...,, --, ".. .. .., ,...,,.,.,,*...*,...,,,... ?. .. . .. . .. ?.. .. . ... ......... ., ?. ..::...X :---:':..,: .::: ..::: ., -- ,,.., ' .,*,.., ,**.?-,,--- , . ". ., ",,* * ., X ::---:X

, :.:.:.:.:.:.: ... .. ......... . i i~i~~i~:::: ::: :::::::: :::::: ::.::::::::::::::::!::::::.. ? ..... ......... .... ? .. :.:.:.:..:?:.:.:..:..,: .. . .... ......... . ... :. :.:.: :.::?:.:.:.:;.:?:.:.::. .:.:.::.:.:,: ::-:.:.;.::?:.:.:.:.... ..................... ...... . . . . . ............ ............... .-.-.--.-.-.-.--, .. ... ? .................. .. - -.? . - -. ..... '. ?? ..?. .. I ...I .?...it-.?:? . * :: .:: :: ::::. ...:.......,. ., ........X ...... ....:. ?.. . - -- . .. .. ::::::::::::::: . .. . *, " :-::- :--::,. ...::...::...::...::...::...:: .. .. ........, ......, ..............,............,.....'. ,.. ........................................." ?.......?... ..?......... .. ....... . ., "... . ....... ...... ....." .. .. ....'. ... . ..... - ... ... .. ... .. ........ ... ... .. .. ......?............... .?. .............. ... .....,..... ................ .. . ..??. .? - . ..........- ..,, I.. :::":,,::*::,,.::,,:,,::*:',".,,'.,'..,'*,.**.., ..." .*." '. , .::,:, :,*.: X ,...*. '-'. ,"'. , , ..... ......) .......?................?. .::?::. :: :--:--: . .. . .. .. .. ... -... .. ... .. .. . ... .. - . ... .. - - ?....... ., . ... . , - I. . ? ? :-- .::,,..-..''''... ' " .''' ...- .. ..... ...::: : : : " . .. . - ..... . .. . .. .. . .? . . - . . . - . -- '- ?., . . ., . ............, ..., .. .. . .::: . . . .. . . .. " *.. . . .. . . . . . .. . . .. . . . . .?... ............. ........ ....... ...,..... . ..... . .. - . .. '.. .. . .-I- I-. . . . ...,........... ,?'.., ...... - ::::- " *. . . . . . .* . ::::::, ,,*.. ..- .-.'* - ,,,...... ---- . --.... .... ?....,...?.......... .------------ .... ... . .?--... ?...,? ----- .. '-- -- . . ? -.- . . . . . . . . .-- : .. : :. ?.:: :! . ," ....:: .o u av d o M f - ............ "*' ,. ..."----..`-..- ...... .. ... -... ::::::,....l.... -. -. .-. .- .- .- -.- -. .. .... ..... ....... .....- ........I .. .........-.' 1.?..:.::?-:..::.::.:.,."*.? ...?--,, - - ... ...,- ?. .......?...-- ..... .. -"*.. -* .... .-, -."I - ... .- .. . - ............ I....-.... ....- ...?..... ...:::?:...,*,,....'''''-.... ..,,.,,,,..'.......--.....-?..--..,?...1 . .. ......-............,,...-....?.?....? .... .. .. .... ?... .........- .. .' -... ..? .....- ..... . ......- . ?...- ..-.......?....... ."-2.. ...?..,.....-. ...? .. .... - I?I

.. .. . .. . .. . .. . .. . .. . .. . . .. . ,... . .. . .. . .. . .. . .... . . ?. " ?. .. ... . .. . . . . I 1 . . . . . - 1 - :. . .. -" -,: : .: . :- . - , -. . . . . . ... ... -,..:::: - ..::::.... ,::::::::.-, ... .. ..........?.........

::: ... : .

i: #..ii.ii.iil :,ii...ti:..I.:ita i.t.. ::::.::.:.::.:. :.::::::::: I .... ..., ::::!:::;:::::,..:!:::,?:. :::::::::::::::::::::::: ::::::::::::::::::::::::...::!:::::::::: :::..:.:..::.:.,,::::,: .......... "*.'..''.''..''...*'."..'.'"'?--.- ---??.....,................,..............:........,....:......,..,. .... ? .... ? .......

............ .. .... .. ' .... . ..-............. : . ." :::, ::: . : .:. .:. ::::,*.:":,..::::::::::::::::::::::::::::::::::: : . . ": ...... * ........ * ....... :::: :::.:. ::x -. . :::;:::::X ?..?;;; - ?j:: ::::": "i":,* : X " .. ........... - ... :.,::::.:::.:-4::-:'j:ii::: i:j.X::. ,:.:.:.:", ............. .. -.:'::::::::':. :::;:?::]??]???::?ii~ii?:i;?:iii?.'.'. -.-,.::-.-.,::::,- :X ::: ?:,.:,.:.:.::::: :.::::::::::: :::::.,:: ..?.......,,..... . ., ::::,::::::::,:::? :::: :::::::::::.::::.,.' '? -'-** --" .. .......... .. . . .. ... ....... ... ... . :.:,:,:.:,:.:.:.:.:,:,:.: . - I. I. I.,. .,.. I. ... ..,8 i . . . . ???t?:?ss:: ?::??- I .. . ...,..:...... ..::::::::::::?::::::::::::::: ::::i.::: :::::::'::..:::: ?:?:?: ? j?:::j::::j:i.. . ::::: .. ? :1 :1 : 1 :. I. I.........., . .. .......... ...7::..:..*. :..:..::.::.::?.:..::.: : . ... .. ... . . .. .:::.:.;.:. :.*.::.: -;? ??: ? ? i?;?i .:.:::::::: ::::::: ::: :::::::: :? .. .. . ................,. ..........:: :::::::::::::::::!?: :::::::.. . . . . .. .. - . .. . . ?. .. .:::l ::l:: : ::.- :::.. :::::::?: ::.' ..,..::;:.:....,.::::. .-.. .-:: ....

.... ? . .... ..... .. .:.::.:..:.?::.:..,....,.? .?,..,.,,.,..,.., ..*..,.-,.,?.-.,? ." ..?..:. . ........... . . ................... ...,....... . . - . I ?.... .... ... ... ? - .. .. ,

""" ::: . . . ... ... .... ::.::-: '.1... .. ".????;????'????;;????????;??i: ii:-:j: ;?:*::::::?i:::::::::::::::.- ............ .??? ??????????????:??]????????????i?i????????i????????????????i??? ?????? .%.- . ... - -- ..?::--:--::-::-::?::-::-::-::-::-:1:1,* '' '"" . 1 1 . . . . - ..:?:.:,.:.:.:.:.:.:.::.,.:.:.:......:........ ... '.-?-.-.-.-" ,-. .I.-...- ... ?..... , .....,.

:;::::::::::::::::::::::::::::::::, . .. . .. , - ::i:?:?:?:??:?:-.':-:?? :?:::??:?:?:? ??:::?:i:,:":*, - - ....... . ... - . ...... .- ......? . ... " .....', ? I

..:.. j ~ :::: '' -- *i:'?::*?:?: ::::: :.-j:..` - .. - - -. '..::: ::::j:?:j:j?:::: :i: :j.*.:%..k.. . .. .. . - -. .. .... .. . . .. .. ... ....... .... .. :. ..I . ...:..:.;:.:.::.:..:.::.:.::.,-??;?];?;?? ??::??;?i?????;??????i?;]???;" ?????; ??????;?????????????????:??:j::i...?:]i ,. . . i? ??. : ' .. . . .... ..,...?...........-.- .?...........?.-.--.??..-....? , .- . '' ,?. I I ? ?

40/ Journal of Marketing, January 2002

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

with Harley-Davidson owners already exists (Schouten and McAlexander 1995). Except where noted, the following description pertains to the study of Jeep owners.

This study conforms in many ways to the technique of situated or autoethnography (Denzin 1997; Denzin and Lin- coln 1994), in that the ethnographers became fully situated as members of the group being studied (see Schouten and McAlexander 1995). With corporate access and support, we began as naive, neophyte off-road drivers with rented Jeep vehicles. By the third year of our involvement, two authors had acquired Jeep vehicles. As suggested by Stewart (1998), this research employed prolonged fieldwork in two brand communities (Jeep and Harley-Davidson), which gave us the opportunity to experience brand community from multi- ple perspectives and observe ways in which ownership expe- riences and motives evolved.

Unlike pure autoethnography, this study was influenced by accountability to corporate marketing decision making. Sponsorship of the Jeep-related research by DaimlerChrysler and Bozell (Jeep Division's principal advertising agency) forced us periodically to withdraw from the situated- consumer perspective and engage in analysis of the owner as "other" in order to consider marketing implications of our emergent findings. Consistent with the current psychological notion of fluid, dynamic, and contextual self structures (Markus and Wurf 1987), shifting our roles as researchers turned out to be almost as easy as shifting context.

Ethnographic Fieldwork We began our ethnographic research at brandfests (McAlexander and Schouten 1998) and then broadened the field context to include sites that were not event related. Eventually, through product adoption we became fully situ- ated in the experience of Jeep ownership and gained per- sonal points of reference on which to reflect our data and analyses. The brandfests, including Jeep Jamborees, Camp Jeep, and Jeep 101, hosted significant numbers of brand owners and potential owners engaging in brand consump- tion and the celebration thereof. Briefly, Jamborees are regional rallies with a focus on off-road trail driving; Camp Jeep is a national rally that, in addition to off-road driving opportunities, offers lifestyle and product-related activities; and Jeep 101 is a touring off-road driving course coupled with product-related activities and displays. Although the participant mix varied somewhat by event, all the brandfests attracted a wide range of owners (and friends), from veteran off-roaders to neophytes and from first-time owners to those with family heritages of Jeep ownership. These events are described more completely in McAlexander and Schouten's (1998) work.

Fieldwork at brandfests consisted mainly of participant observation augmented by photography, videotaping, and informal interviews. Near the end of an event, it frequently became convenient to engage in more formalized depth interviews. Our data-recording methods mirrored the activi- ties of many other participants who sought to capture their experiences to share with "the folks back home." We gath- ered data throughout the events, including at orientations, at meals, on the trails, and at social gatherings. Typically, for the first day or two of an event, we maintained an unobtru-

sive presence as participants. Accurately representing our- selves as relative newcomers to the off-road experience, we placed ourselves in positions to empathize with other first- timers and to receive instruction and socialization from experienced owners. Our conversations covered such topics as ownership history; off-road experience; current thoughts (e.g., anxieties, excitement, periodic boredom, opinions about the event); plans for Jeep usage; and personal data, including names, occupations, family situations, and occa- sionally addresses. On-trail relationships developed a famil- iarity that enabled us to observe changes in attitudes and social relationships over the course of an event.

To challenge our emergent findings from event ethnog- raphy, we broadened our scope to include informants who did not choose to attend brandfests. We interviewed six informant households in a western city (selected for diver- sity in ownership of the basic Jeep platforms). This research is also informed by ongoing ethnographic involvement with Harley-Davidson owners beyond that reported previously. Relationships with informants established eight years earlier provide us longitudinal perspectives of these consumers' experiences in brand communities. We also conducted depth interviews with marketing managers and consumers of the brands DeWalt (a marketer of power tools) and Mentor Graphics (a producer of software used in high-technology design). These interviews helped us determine whether the findings are resonant across diverse industries and markets rather than idiosyncratic to the industries we studied.

Interpretive Analysis Analysis of the ethnographic data occurred at several levels. We conducted a first level of analysis individually through- out the process of data collection, continually reflecting them against previous data and emerging themes. We con- ducted a second level of analysis at intervals throughout each day in group analysis and synthesis sessions. These sessions involved the principal researchers and frequently included research assistants and/or agency personnel who attended and monitored the events. At these sessions, each person discussed his or her activities, findings, and insights from the foregoing period. We used previous findings to elaborate, test, and delimit emerging themes as well as to suggest directions for further inquiry. The third level of analysis involved intensive study of recorded data. The goal of this analysis was to scrutinize themes more thoroughly and elaborate those that stood up to scrutiny.

Ethnographic Findings Brandfests, in essence, provide for geotemporal distillations of a brand community that afford normally dispersed mem- ber entities the opportunity for high-context interaction. These conditions prevail for all the types of customer- centric relationships that make up a brand community. In this section, we present findings that both support and extend Muniz and O'Guinn's (2001) brand community work, especially with respect to their central issues of con- sciousness of kind, shared rituals and traditions, and sense of moral responsibility. We then expand on the dynamics of geography, social context, temporality, and interlocking

Building Brand Community 1 41

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

communities. Last, we offer a new conceptualization of cus- tomer loyalty as integration in a brand community.

Brand Community Emerges During fieldwork at Jamborees and Camp Jeep, we observed the building of brand community. The Jeep owners who came to these events were quite diverse. Our observations, sup- ported by event registration data, revealed that roughly half of event participants had never taken their vehicles off the highway. They belonged to no Jeep-related clubs, felt no sub- cultural affiliation, and had few interactions with other Jeep owners beyond incidental encounters on the roads or parking lots of their daily living environments. Even among the vet- eran off-roaders, important differences existed. For example, certain drivers espoused the "tread lightly" ethos promoted by conservation groups and Jeep corporate communications, whereas others demonstrated with mud-covered vehicles a more cavalier stance toward the environment and an ethos of conquest over nature. Notwithstanding their differences, these diverse Jeep owners created meaningful temporary communities within the broader brand community.

All the characteristics of brand community discussed by Muniz and O'Guinn (2001) soon manifested themselves: consciousness of kind, shared rituals and traditions, and a sense of moral responsibility. In our observation, however, these characteristics of brand community did not all and equally exist before the brandfest events.

Consciousness of kind, for many community members, was tempered by a fear of not belonging. Interviews at the inaugural Camp Jeep revealed that one potential barrier to participation among some owners was a fear of not fitting in. Put simply, the images they held of other customers in this particular consumption context were inaccurately based on prejudicial stereotypes that often emphasize ways in which these "others" are different from themselves. Jake, a new Grand Cherokee owner, said that he "almost didn't come" to Camp Jeep because he expected the event to attract a pre- dominance of "barbarians" and "hard-core four-wheelers." He spoke of his fear of feeling "like some geeky yuppy on the sidelines." That fear was quelled somewhat upon his arrival as he observed other Grand Cherokees, which he pre- sumed belonged to people similar to himself. Similarly, Amanda, the upscale wife of a retired surgeon, attended the 1996 Camp Jeep reluctantly. On the first day of the three-

day event, she explained, "I just don't see myself as a 'Jeep Person,'" in a pejorative tone. At the close of the event, we spoke with her again, and she reported having experienced a quantum shift in attitude. She lauded Chrysler's efforts in creating an enjoyable event. Her experience gave her added appreciation for her own Jeep and its capabilities. Moreover, having interacted pleasurably with many people, she no longer maintained a me-versus-them attitude about "Jeep people."

Anxieties about belonging were dispelled in part by the outreach of experienced participants acting out of a sense of moral responsibility. For example, on-trail instruction for neophytes at Jamborees was provided by knowledgeable participants who spontaneously stepped forward to assist them through difficult stretches of trail. In Colorado, a long- time Jeep owner spent time at an intimidating stream cross-

ing, loudly guiding drivers along the "correct" route through the rough water. He encouraged inexperienced drivers and reassured them about the capabilities of their vehicles. He seemed to relish the recognition and status that came with his superior knowledge and skills. The benefits of brand socialization between more and less experienced owners are symbiotic. New owners benefit from the expertise and social approval of veterans. At the same time, veterans benefit from the status accorded them in their assumed leadership roles. Moreover, the community as a whole benefits as exchanges of knowledge cement relationships through reci- procal exchanges of value (Gouldner 1960; Sahlins 1972).

As community members, marketers also contribute to the process of community building by creating the context in which owner interaction occurs. For example, participants shared their driving experiences through ritual storytelling facilitated by activities, such as barbecues and roundtable discussions, hosted by the marketers. Topics ranged from trails and adventures to aftermarket equipment. One rela- tively mundane conversation centered on the ritual "Wran- gler wave" given by Jeep Wrangler owners when they pass each other on the road. The wave is a tradition that many owners learn in the course of everyday driving, but we saw this in-group knowledge verified and reinforced through dis- cussion at a Jeep 101 gathering.

Marketers may also take an active role in establishing the shared rituals, traditions, and meanings that foster con- sciousness of kind. Tools such as a "History of Jeep" exhibit at Camp Jeep and Jeep 101 helped underscore a sense of similarity, authenticity, and exclusivity among participants. Promotional materials that depict the product in use may help establish shared aspirations. One informant at Camp Jeep told us that he had come to the Colorado event from his midwestern home largely because of inspiring advertise- ments that depicted a Jeep like his in a pristine mountain environment. His excitement was heightened by the regis- tration packet, which provided guidelines for enjoyment of the event. Certain techniques of off-road driving, such as starting the vehicle in gear to avoid tire slippage, were repeated by staff driving instructors many times over the course of a Jeep Jamboree and were codified in a booklet of off-road driving tips.

Marketers, as well as owners, have incentives to exercise moral responsibility to brand and community. For example, the ethic of nondestructive driving was a matter of intense indoctrination at Jamborees. The event sponsors heavily promote the "tread lightly" program as a cornerstone value of Jeep and Jeep ownership, and the message is passed along by owners. To the owners and marketers alike, "treading lightly" serves the dual purposes of protecting the environ- ment and preserving access to off-highway trails.

The community-building activities of event participants, including marketers, appeared to be remarkably efficient. Even owners who came to events dwelling on how different they felt from others often left after two or three days believ- ing they belonged to a broader community that understands and supports them in realizing their consumption goals. In this discussion of event-intensified brand communities, we have dwelt primarily on human-to-human interactions. In these geotemporally concentrated situations, the relation-

42 / Journal of Marketing, January 2002

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

ships between owners and their vehicles and between cus- tomers and the brand also benefit from enriched context. Through ritual consumption, participants routinely reported newfound intimacy with and understanding of their vehi- cles. Likewise, their many positive experiences under the auspices of the Jeep brand conveyed its potency as a symbol of their values and lifestyle preferences. To better under- stand how these transformations occur, we turn again to the dynamic dimensions of brand community.

Dynamic Dimensions of Brand Communities At an Ouray, Colo., Jeep Jamboree, we interviewed partici- pants from many parts of the country, including California, Utah, Iowa, and Illinois. Some had come with friends or members of local clubs, but many had only passing acquain- tanceships, if that, with other Jeep owners. One couple from Illinois was illustrative. Their prior connections to brand community were limited to the purchase transaction with their dealer, direct mail from Jeep, and advertising. They learned about the Jamboree from a brochure. They had been to Colorado for winter ski trips but had never experienced the Rockies during the summer. The Jamboree gave them an excuse to do so and to become familiar with the four-wheel drive capabilities of their Grand Cherokee. This was their first off-pavement excursion. At the beginning of an all-day trail ride, we found ourselves lined up behind them. We were intrigued by their enthusiasm, especially that of the wife, who carried a large camcorder as though it were an appendage of her body. By midway through the day, we were calling each other by first names, sharing personal sto- ries, and exchanging addresses with the expectation of future interaction. An unforgettable part of the experience was watching their confidence in their Jeep grow as they surmounted the challenges of the rough and sometimes pre- cipitous road. Their appreciation for their vehicle was con- tagious. This was not an isolated case. Similar bonds were formed among strangers at every event we attended, and at every event people were vocal with their enthusiasm for Jeep and their Jeep vehicles.

Just as Jeep owners got to know one another by virtue of sharing the same geotemporal space, they also got to know, or at least interact with, agents of DaimlerChrysler and the Jeep brand. At Camp Jeep they were able to converse with Jeep engineers in roundtable discussions. At Camp Jeep and Jeep 101, uniformed "camp counselors" were on hand to deliver hospitality that ranged from free beverages to Jeep product information to off-highway trail recommendations. The message in this: Behind the product and the normal cor- porate communications are real people who understand and care about their customers.

Compared with the normally diffuse nature of a brand community, the temporary geographic concentrations pro- vide a rich social context for communication. In close prox- imity, people got to know one another in ways that would be difficult or impossible through electronic or mass media. Face-to-face contact reduces opportunities for personal mis- representation. A novice driver would have difficulty boast- ing of off-road driving prowess while nervously hanging on to every word of a trail coach or spotter. More to the point, high-context interactions speed up the processes of social-

ization. They enable consumers to see, feel, and hear demonstrations of product use. They provide more informa- tion on which to base judgments about people's credibility, sincerity, and concern for one another. In some cases, sus- tained interpersonal interactions can lead to relationships that transcend mere common interest in a brand and its applications.

By the end of a Jamboree or Camp Jeep, weekend partic- ipants may have formed friendships necessitating good-byes. Inherent in the good-byes there may be a desire to reconnect with people or form similar friendships at future times in similar situations. At our first Camp Jeep, we were struck by the enthusiasm that we shared upon encountering a couple whom we initially met at our first Jamboree. At the second Camp Jeep, we found participants seeking friends whom they had met the previous year at the inaugural Camp Jeep event. This included us, as we reestablished ties with informants we had met at both the previous Camp Jeep and the earlier Jam- borees. What began as a temporary or situational community had developed an apparent momentum toward greater tem- poral stability. This increased sense of community longevity appeared to be a direct result of the qualities of relationships facilitated by the temporary geographic concentration and the contextual richness of the events.

The Jeep brand community is not the primary commu- nity affiliation for most, if any, of our informants, nor does it exist independently of other social groups. Consider the phenomenon described by Pamela and Greg, Cherokee owners, in this West Coast in-home interview excerpt:

Pam: A number of people have bought 'em (Jeep Chero- kees) since we've bought....

Greg: Yeah, it's sort of like it's just grown in a circle of people. The people that we initially-before chil- dren-we used to hang out a lot with, three or four of them have Jeeps of some kind or....

Pam: They all lived in this little cul-de-sac, and they all owned Jeep Cherokees! I mean you didn't move there unless you [laughing] had a Cherokee!

Other informants also spoke of friends and family who owned Jeeps. Waiting in line to drive at Jeep 101 in Boston, we struck up a conversation with Tom and Cindy, who were approximately 35 years of age, and asked if any of their friends or family also owned Jeeps. Tom described a three- year chain reaction of Jeep purchases that began with his own and ended with three of his five siblings as well as his father also owning Jeep vehicles of one model or another. Tom's experience is a vivid illustration of what Olsen (1995) calls the "lineage factor" in brand loyalty.

By encouraging participants to bring friends or family to Jeep 101, the corporate sponsors provide opportunities for sharing brand enthusiasm and experiences. Among Jeep owners, we have observed many cases of conversion to the brand among family members, including siblings, spouses, children, and parents. Larry, a Boston optometrist we inter- viewed at Jeep 101, called to his 15-year-old son during our interview and introduced him as "the next generation of Wrangler owner" to the son's enthusiastic agreement. The Goldsmiths, a New York family and owners of their third Grand Cherokee, came to Jeep 101 to assess the Wrangler as

Building Brand Community 143

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

a possible car for their daughter, who was about to turn 16; by the end of the event, they were planning the purchase right down to the choice of color. In an interview with a father-son pair in New York, we learned that the father, approximately 45 years of age, had traded in his luxury sedan for a Jeep Grand Cherokee on the urging of his son, a college freshman who had owned a used Jeep Cherokee for about a year.

A particularly fervent expression of missionary zeal comes from an interview with Barbara, a woman approxi- mately 50 years of age who now owns her second Jeep Cherokee:

I really annoy a lot of my friends because I am constantly talking about Jeep. I absolutely love my Jeep and I will continue to love my Jeep. To me there's just no substitute for Jeep. I could walk around with a sign: "I - love - my - Jeep!" [in a chanting cadence, then laughter] I do this to my friends. I really do. All the time.

Integration in the Brand Community An ethnographic perspective helps us understand how and why community grows through and in the form of customer- centered relationships. Events like Jeep Jamborees, Camp Jeep, and HOG rallies bring together people, or parties of people, who often share no other connection than an interest in a brand and its consumption. Given the opportunity for context-rich interaction, in which previous communication was either nonexistent or limited to mass or electronic media, participants share extraordinary consumption experi- ences (see Arnould and Price 1993; McAlexander and Schouten 1998). Sharing meaningful consumption experi- ences strengthens interpersonal ties and enhances mutual appreciation for the product, the brand, and the facilitating marketers. Virtual ties become real ties. Weak ties become stronger. Strong ties develop additional points of attach- ment. Our analysis suggested that consumer-centric rela- tionships with different entities in the brand community might be cumulative or even synergistic in forming a single construct akin to customer loyalty. Put another way, more and stronger points of attachment should lead to greater integration in a brand community (IBC). Similar to the con- struct of brand loyalty in that it conveys an emotional and behavioral attachment to a brand (Ehrenberg 1988; Jacoby and Chestnut 1978), IBC is a more comprehensive concept grounded in consumers' total-life experience with a brand as most broadly construed.

As compelling as we found our ethnographic evidence to be, we nevertheless desired to triangulate our data. Quan- titative testing enabled us to confirm qualitative interpreta- tions and answer several empirical questions.

Quantitative Triangulation Ethnographic work revealed four customer-centered rela- tionships that appeared to be integral at the macro level to the brand community and at the micro level to individual IBC We observed that the brandfests, which provided much of the context for our study, have an impact on each of those

relationships. Our desire for quantitative triangulation and additional empirical exploration led us to examine the fol- lowing hypotheses in the context of Camp Jeep:

H1: Integration in the Jeep brand community (IBC) is a func- tion of the customers' perceived relationships with their own vehicles, the brand, the company, and other owners.

H2: Customers will report more positive relationships with their own vehicles after participating in the brandfest.

H3: Customers will report more positive relationships with the Jeep brand after participating in the brandfest.

H4: Customers will report more positive relationships with Jeep as a company after participating in the brandfest.

H5: Customers will report more positive relationships with other Jeep owners after participating in the brandfest.

H6: The overall level of integration in the Jeep brand commu- nity will increase as a result of participation in the brandfest.

Quantitative Methods This field study was constrained in some ways by corporate requirements. The construction of the questionnaire was influenced by the need to address a broad assortment of top- ics for different corporate audiences, the desire to maximize response rate, the need to manage a budget, and other con- flicting objectives. The intent of this portion of the study was to triangulate the qualitative findings and to some extent quantify effects of the marketing program. The scales we used were derived largely from ethnographic data. On the basis of informant interaction, they have face validity.

We collected quantitative data in connection with two consecutive national Camp Jeep events. The first year, we conducted a post-event survey of participants. It included questions related to the participants' experience at the event and measured variables related to our model of brand com- munity. Analysis of this data served as a pretest for the design and execution of the survey at the second event. The assessment of the second Camp Jeep was structured as a one-group pretest/posttest quasi-experimental design (Cook and Campbell 1979) of the impact of the event. In this study, the experimental treatment was participation in the event.

Because no control group was associated with this design, the interpretation of our results must acknowledge the possi- bility of several different biases (Cook and Campbell 1979). We designed the measurement to minimize the potential impact of these biases. For example, we timed the survey and devised the instrument to reduce the impact of testing bias. The pretest survey had a cutoff return date approximately five weeks prior to the event and asked a wide range of questions that were not specific to the research topic (e.g., proprietary segmentation questions). The post-event questionnaire was similarly constructed to make it difficult for the respondent to glean our research purpose (e.g., including a series of ques- tions evaluating specific event activities) and was sent approx- imately three weeks after the event. Consequently, the repeated measures occurred about eight weeks apart, a time period we believed was sufficiently long that respondents could not read- ily recall the first set of questions or their responses to them. Moreover, the quantitative findings need not stand alone; they are triangulated with ethnographic fieldwork conducted at the sites of these and other brand community events.

44 / Journal of Marketing, January 2002

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

Sampling Plan

Camp Jeep was national in scope. It was promoted through a national print campaign. In addition, recent Jeep pur- chasers were sent direct-mail invitations to the three-day event. Before the event, all registrants were sent a package of materials, which included information about activities, suggestions for items to bring to the event (e.g., sunscreen, appropriate clothing), and an identification-coded pre-event questionnaire. They were directed to return the question- naire with other registration information in a prepaid return envelope. Participants were encouraged to submit this infor- mation by a date approximately five weeks prior to the event to reserve space in desired activities. Approximately 40% (n = 453) of preregistered participants returned the question- naire on time.

Three weeks after the event, an identification-coded post-event questionnaire was sent with a cover letter and other materials to all respondents who had returned the pre- event questionnaire. Respondents were also sent a reminder card. A total of 259 usable questionnaires were returned, for a response rate of 57% of initial respondents. Because we were able to match returned post-event questionnaires with the pre-event questionnaires, it was possible to provide com- parisons between consumers who did and those who did not return the second questionnaire. We found no meaningful differences between these groups.

Measurement On the basis of the model developed from our ethnographic research, we measured four customer-centric relationships: owner-to-product, owner-to-brand, owner-to-company, and owner-to-other owners. All items were measured with a five point Likert-type scale, anchored by (1) "strongly disagree" and (5) "strongly agree."

The customer-product relationship was measured with four items that attempt to capture owners' feelings about the product they own (see Table 1). The selection of these items was informed by our ethnography, Belk's (1988) work on extended self, and the earlier pretest. The customer-brand relationship was measured with two scales. One represented important brand-related values or associations that have been promoted by the marketer.1 The other scale consisted of measures that are typically associated with brand loyalty. The brand association scale consisted of three items. The brand loyalty scale consisted of four items (Table 1). The customer-company relationship was measured with two items that attempt to capture the feelings owners have about the organization that sponsored the event. This scale reflects the company's concern for customers, a theme that came up consistently in our on-site ethnography and interviews among participants (Table 1). The customer-customer rela- tionship was measured with three items that attempt to cap- ture the feelings owners of the product have about other owners (Table 1).

Analysis and Results The survey items that correspond to the four customer- centered relationships were tested in a four-factor confir- matory model (details are described in the Appendix). An acceptable fit was achieved after three items were elimi- nated; this change had a negligible impact on the substan- tive content of the affected dimensions, and reliability and average variance extracted (AVE) for each dimension were good (Table 2, Appendix). The next step was to assess whether these four constructs were an adequate reflection of a single higher-order construct. For this pur- pose, a second-order factor structure was tested (see Fig- ure 3), and an acceptable fit for this model was found (Table 3). The confirmatory factor analysis (CFA) and the second-order model provide evidence of two important points: First, each dimension has good measurement prop- erties and is distinct from the other dimensions, and sec- ond, the dimensions can be combined to form one higher-

TABLE 1 Community Integration Scale Items

Dimension Items

Product 1. I love my Jeep vehicle. 2. I am proud of my Jeep vehicle.a 3. My Jeep vehicle is one of my favorite

possessions. 4. My Jeep vehicle is fun to drive.

Brand 1. I value the Jeep heritage.a 2. <proprietary> 3. <proprietary> 4. I would recommend Jeep to my friends.a 5. If I were to replace my Jeep vehicle, I

would buy another Jeep. 6. Jeep is of the highest quality. 7. Jeep is the ultimate sport-utility vehicle.

Company 1. The Jeep division understands my needs. 2. The Jeep division cares about my opinions.

Other 1. I have met wonderful people because of owners my Jeep.

2. I feel a sense of kinship with other Jeep owners.

3. I have an interest in a club for Jeep owners. aRemoved from final scale. Notes: Items identified as proprietary reflect specific brand-related

values. To protect the competitive position of participating organizations, these individual items are not reported.

TABLE 2 Measurement Properties

Number Construct of Items Reliability AVE

Owner-product 3 .90 .74 Owner-brand 5 .88 .58 Owner-company 2 .88 .79 Owner-other

owners 3 .70 .61

IWe also measured a customer's identification with the brand (as reflected by a desire to wear branded clothing or purchase branded accessories). Additional analysis substituting these measures of the customer-brand relationship produced similarly supportive results.

Building Brand Community / 45

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

FIGURE 3 Two-Stage Model of IBC

Owner-Product I

Owner-Product

Owner-Product 3

Owner-Brand 2

COmperBrany Owner-CBran2d 3 Relationship Owner-Brand 3

Owner-Brand 4 T12

Owner-Brand 5 IBC

Owner-Company I Owner- Company

Owner-Company 2

----os

Owner-Owners 13

Owner-Owners 3

Owner-Owers

Owner-Owners

Own r-Owners 3 T~ 14

order construct. From this point, the focus shifted to examining the brandfest's impact on participants' attitudes.

We performed two tests to begin to understand the rela- tionship between the pre- and post-event scores. The first test was an overall assessment that modeled pre-event IBC and post-event IBC as unobservables in a structural equation model (see Figure 4). The structural parameter between the pre- and post-event brand community was left as a free para- meter to be estimated in one model and fixed at zero in a second model. When the parameter was estimated, the coef- ficient was positive and significant, and the fit of the model was significantly better than the model estimated without the parameter (see Table 4).

The second test was a repeated measures analysis that used the four components as well as the composite measure of IBC as dependent variables. Separate analyses were run for each variable. On the basis of an examination of pre- event frequencies, it was clear that some owners already had strong positive feelings. Because these owners had nearly "pegged the scale" before attending the brandfest, it would be difficult for the brandfest to have a significant, positive impact on their attitudes. To account for this ceiling effect, IBC and the four component relationships were sorted by the magnitude of the pre-event scores, split at the median, and analyzed separately as "high" and "low" groups.

The majority of the analysis of variance (ANOVA) results showed violations of the homogeneity of variance assumption (Levene 1960). We estimated the data sets that

exhibited homogeneity of variance violations using the gen- eralized least squares (GLS) procedure (Judge et al. 1985) in the SHAZAM econometric package (White 1977), and we estimated the remainder using ordinary least squares (OLS). Using the approach suggested by Cohen and Cohen (1983), we calculated the percentage of within-subjects variance explained by time for each regression.

For owners who were less positive before the event, all five repeated measures analyses showed significant changes between pre- and post-event scores, with R2 values ranging from .25 to .57. Three of the five upper-half analyses were significant, with R2 values ranging from .04 to .08 (Table 5). Pre- and post-event means for each group are also included in Table 5.

Discussion Both the quantitative and qualitative analyses offer support for hypotheses regarding the multifaceted nature of brand community and the way in which brandfests may build it. With respect to hypotheses H2-H6, we also found that the strength of prior owner bonds wields significant moderating influence in the experience of the brandfest. A discussion of the hypotheses is followed by additional ethnographic findings.

H1: The Construct of Brand Community H1 was confirmed strongly. Quantitative analysis supports the integrity of a construct of IBC as the cumulative impact of four types of customer-centered relationships. In addi- tion, although the individual relationships are depicted as dyadic, our ethnographic data reveal that they do not func- tion entirely independently of each other. Rather, they develop interdependently in ways that are mutually reinforcing.

The following excerpt from an interview at Jeep 101 provides an example of how brand community may begin to form in the early stages of purchase and ownership. The informant, Susan, is a first-time Jeep owner attending Jeep 101 with her fiance, George. Throughout the purchase process, Susan referred to her interpersonal relationships for assistance and input. Interaction with friends, car dealers, and the Internet (with the assistance of her brother) all helped her arrive at her purchase decision. Subsequently, postpurchase communication with her dealer and from the company further strengthened her connection to the brand:

I've been very happy. I get a lot of communications from Jeep, which I've been so impressed with. Usually you buy a car and then you're a forgotten soul. Its kinda like they want you to be part of the family. As soon as I got the invi- tation for Jeep 101, I registered. I was very excited. But I was nervous. I didn't think I would end up driving. I was very relieved to see someone in the car with you, 'cause it gave you the confidence to do what you're supposed to. Otherwise, I had visions of abandoning the truck on the hill and saying, "I can't do it!" [laughs]. I thought I might wimp out, but I didn't [smiles].

Susan viewed "Jeep," a corporate entity, as a caring institution, a family, that provided her a sense of belonging and importance, not feelings of being a "forgotten soul." In

46 / Journal of Marketing, January 2002

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

TABLE 3 Standardized LISREL Estimates for Second-Order Modela

Indicator Owner-Product Owner-Brand Owner-Company Owner-Other Owners

First-Order Loadings (Ay) Own-Prl .894b Own-Pr2 .830 (15.29) Own-Pr3 .863 (16.40) Own-Brdl .761 (9.89) Own-Brd2 .753 (8.78) Own-Brd3 .706b Own-Brd4 .759 (9.86) Own-Brd5 .757 (9.84) Own-Orgl .795 (9.98) Own-Org2 .986b Own-Own1 .713 (10.82) Own-Own2 .900b Own-Own3 .712 (10.81)

Second-Order Loadings (Tyk) First-Order Construct Community Integration Product .925 (13.41) Brand .939 (10.13) Organization .595 (8.53) Other owners .802 (11.19)

Goodness-of-Fit Statistics X2 (61 d.f.) = 223.61 (p = .0) Goodness-of-fit index = .867 RMSR = .060 Normed fit index = .872 CFI = .903

aThe t-values are in parentheses. bFixed parameter.

FIGURE 4 Model to Assess the Association Between Pre-

event and Post-event Measures

0e5,1 E Pre-event Owner-Product

86,2 2

--

Pre-event Owner-Brand X2 IBC

03 E3 Pre-event Owner-Company X4

0E4

4 Pre-event Owner-Owners

0E5 Post-event Owner-Product X5

8E6

-

Post-event Owner-Brand )6 IBC

00e Post-event Owner-Company ).

OE8 ---'

Post-event Owner-Owners

a situation that inspired fear as well as excitement, "Jeep" was there in the form of a real person to support her in achieving a meaningful personal triumph. Her experience left her wanting more of the same kind of interaction in the

future. She feels and values a sense of community among Jeep owners:

I would love to do a Jamboree.... If we had something locally I would definitely go. Or even somewhere on the East Coast, you know, Virginia, Florida, I'd go. Defi- nitely. And it really makes you feel like a part of a family.

Susan's relationship with the Jeep brand means that she is not alone in the important and sometimes daunting arena of automobile ownership. Moreover, she appeared to be blind to the lines between the corporate and dealer organi- zations, seeing instead a unified entity that she can trust to take care of her and her Jeep vehicle. A combination of attentive dealer service, periodic marketing communica- tions, and face-to-face interaction with company agents at Jeep 101 have strengthened Susan's perceived relationships with the entity of "Jeep." She has learned to understand and appreciate her own Jeep vehicle better. She feels like she belongs to a benevolent family of Jeep that includes both owners and marketers. This community gives her confi- dence and inspires her to seek out additional opportunities, such as Jamborees, to join in shared consumption of the brand.

Susan's story is far from unique among our informants. Repeatedly in our ethnographic work, we have encountered

Building Brand Community / 47

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

similarly synergistic development of customer relationships. Products are purchased and consumed in the context of social and business relationships, which in turn influence feelings about the products specifically and brands more generally. Each relationship connects to all the others through the central nexus of consumer experience, creating

the holistic sense of a surrounding community. Moreover, each relationship acts as a personal linkage to the brand community. The more each relationship is internalized as part of the customer's life experience, the more the customer is integrated into the brand community and the more loyal the customer is in consuming the brand.

TABLE 4 Comparison of Two Models of Pre-event and Post-event IBC

Fit Statistics for the Model Fit Statistics for a Model of Connecting Pre- and Post- Independence of Pre- and event IBC Measures Post-event IBC Measures Change in Value

IBC Correlated error terms used 4 of 4 possible 4 of 4 possible x2 25.763, 15 d.f. 92.852, 16 d.f. 67.089,1 d.f. P(2,1 )b .634 (7.34)a .0 RMSR .0142 .0800 .0658 Goodness-of-fit index .971 .902 .069 Normed fit index .965 .872 .093 CFI .985 .890 .095

aThe t-value is in parentheses. bStandardized parameter estimate.

TABLE 5 Means for Pre-event and Post-event Scores and Repeated Measures Regression Results

Lower Halfa Upper Half IBC Pre-event mean 43.42 55.73

Post-event mean 49.64 57.23 F(1,98) = 80.28 F(1,97) = 6.37c

p < .001 p < .025 R2 = .450b R2 = .062

Product Pre-event mean 11.22 14.67 Post-event mean 12.51 14.22

F(1,98) = 33.44 F(1,97) = 8.55C p < .001 p < .005

R2 = .254 R2 = .081

Brand Pre-event mean 17.20 22.00 Post-event mean 19.38 22.51

F(1,98) = 49.27C F(1,97) = 4.54c p < .001 p < .05

R2 = .335 R2 = .045

Company Pre-event mean 5.68 8.16 Post-event mean 6.94 8.28

F(1,98) = 55.23c F(1,97) = .70c p < .001 p = n.s.

R2 = .360 R2 = .007

Other owners Pre-event mean 8.07 12.16 Post-event mean 10.57 12.46

F(1,98) = 132.13c F(1,97) = 2.19c p < .001 p = n.s.

R2 = .574 R2 = .022

aBased on a median split of the pre-event scores. aR2 presented is the percentage of within-subjects variance explained by time. cWe used GLS estimation to account for differences in variance between the pre-event and post-event scores. Notes: n.s. = not significant.

48 / Journal of Marketing, January 2002

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

H2: The Brandfest and the Customer-Product Relationship For Jeep owners who felt less of a bond to their vehicles before the brandfest, H2 was confirmed strongly; brandfest participation led to more positive relationships with their Jeep vehicles. However, for the owners who felt more posi- tive toward their vehicles, we found a statistically significant reduction in their enthusiasm for their Jeep vehicles.

The increase in positive feelings toward their vehicles of owners with weaker product bonds (and who usually have less product-related experience) can be explained by the nature of their brandfest experience. Many of them have their first off-road experiences, learn new skills, and, more important, learn about previously unknown or unappreciated capabilities of their Jeep vehicles. At the conclusion of a Colorado Jeep Jamboree, Vivian, a single woman approxi- mately 50 years of age, offered these remarks that indicated the development of an interdependent (Fournier 1998) customer-product relationship:

Researcher: Did your experience change the way you think about your Jeep?

Vivian: I love my Jeep now. I liked it before, but now I feel like we're a team.

Researcher: You've formed a relationship with the vehicle that didn't exist before?

Vivian: That's right [continues nodding her head]. Researcher: Could you explain how this happened? Vivian: Well, mine sits in the garage a lot because I have a

second vehicle. I've had the Grand Cherokee for a year, and it's just kind of my luxury car and so I didn't know it as well as I have since this weekend. [Now] I just feel like it's very dependable. And I feel like we could probably go anywhere and I'm not going to be fearful.

In contrast to Vivian and other relative neophytes, par- ticipants who were already more highly engaged with their own vehicles tend already to be very familiar with them. Vigilant for information about engineering improvements in new models or newly available performance modifications, the more experienced owners may feel some dissatisfaction from comparing their current vehicles with others, which they regard as even marginally superior. For example, we asked Jody, a 24-year-old Wrangler owner, what he thought of the new Wrangler model (on display and in use at Jeep 101). He described at length the advantages of its new fully independent coil-spring suspension over the leaf-spring sus- pension of his own vehicle. When we asked if he thought he would trade up to the new model, his response was that he would if he could afford it. Exposure to new and improved products within the context of a brand community at a brandfest may be instrumental in motivating trade-ups and the purchase of new vehicles and accessories.

H3: The Brandfest and the Customer-Brand Relationship For Jeep owners who exhibited weaker connections to the brand before the event, H3 was confirmed strongly; brand- fest participation led to more positive relationships with the

Jeep brand. Those who felt stronger brand connections before the event also showed strengthening in the customer- brand relationship, but to a lesser degree. The main differ- ence in attitude shifts seemed to be a simple matter of room for improvement; those in the upper measurement category of the customer-brand relationship had already pretested at such high levels that there was less room left in the mea- surement scale to accommodate upward change.

Changes in the customer-brand relationship among cus- tomers who previously had weaker bonds to the brand were sometimes dramatic. For example, David, who was approxi- mately 25 years of age and was the owner of a Ford Explorer, had this to say after his Jeep 101 driving experience:

Dave: I just had the most amazing time. My dad's friend is the dealership, you know, the Ford guy? So I don't want to mock him or anything, but (the Wrangler) kills my Explorer. My Explorer's a Limited. It has air suspension, all this special crap. And it just destroys my Explorer.

Researcher: Would you feel comfortable on this course with your car?

Dave: I don't think so. I think it might tip, actually.

David left the Jeep 101 resolved to buy a new Jeep, despite a long-standing family friendship with the Ford dealer who had made him a deal on a top-of-the-line Explorer sport-utility vehicle. Among non-Jeep owners who attended Jeep 101, mostly as guests of Jeep owners, we interviewed several who appeared to have become converts to Jeep.

Other evidence of improved customer-brand relation- ships manifested in participants' interactions with brand dis- plays and in purchases of branded products. Camp Jeep and Jeep 101 participants report deriving considerable benefit from Jeep heritage displays (which include Jeep advertise- ments through the decades, a video about Jeep's military beginnings, and antique Jeep vehicles), engineering displays (including life-size cutaway models of new Jeep vehicles), and product displays (which include new Jeep models and accessories). At these and other such events, we observed brisk sales of branded accessories. Informants proudly wear branded apparel from the events when they return home and for long afterward (see Cornwell 1990).

H4: The Brandfest and the Customer-Company Relationship For those who felt less tied to the company before the event, H4 was confirmed strongly; brandfest participation led to more positive relationships with the Jeep corporate entity. The owners who scored higher on this scale before the event showed no significant change in this relationship. As for H3, the main difference in attitude shifts between these owner groups appears to be relative room for improvement.

Corporate image can play an important part in customer reactions to a company's products (Brown and Dacin 1997). Our analysis, which examines perceptions of the company as a caring and approachable entity, demonstrates that a brandfest can effectively influence that image. Participants report how impressed and respectful they are that Jeep would go to the expense of hosting Jeep 101 or Camp Jeep without introducing hard-sell (or even soft-sell) sales tactics.

Building Brand Community / 49

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

Consistent with gift-giving theory (Sherry 1983), the com- pany has given customers a gift, the brandfest, without push- ing for direct reciprocity, thereby creating a sense of indebt- edness or generalized goodwill on the part of the customer.

More important, our research emphasizes the value of going beyond image-building endeavors to establishing real relationships between the company and customers. Customers crave audience with the people behind the brands. The ethnographic evidence for this observation is staggering. At Camp Jeep, the engineering roundtables consistently drew maximum-capacity participation. At Harley-Davidson brand- fests, where intense corporate interaction with customers is made policy, some corporate officers have achieved celebrity status and occasionally are swamped by requests for pho- tographs or signatures. Brandon J., an employee who rou- tinely travels to attend the company's events, has developed friendships with people from coast to coast who seek him out at events to chat or for after-hours socializing. Employees can provide customers a human manifestation of the com- pany at a time when many corporations are perceived as impersonal and unfeeling bureaucracies.

H5: The Brandfest and the Customer-Customer Relationship For customers who felt less connected to other owners, H5 was confirmed strongly; brandfest participation led to more positive relationships with other Jeep owners. Owners who pretested higher on this scale showed no significant differ- ence in their feelings toward fellow owners. The latter result can be explained, as for previous hypotheses, in terms of high base-level responses. These typically more experienced owners likely have had more interaction with other owners. Less experienced owners probably have had less cause to form brand-related interpersonal relationships, and their overall image of other users is more likely to suffer from stereotypes or misconceptions before the brandfest.

For an example of how a brandfest can facilitate lasting customer-customer relationships, consider the case of Katie, a 22-year-old Michigan woman and first-time Jeep Wrangler owner. She had attended her first Jamboree with her sister-in-law Kim in Michigan the previous year. At the Michigan event, she met two young men and their father who also owned a Wrangler. The two parties got acquainted during the event, and afterward they corresponded to share home videos of the Jamboree experience. By prearrange- ment they met again at the Jamboree in French Lick, Ind., where the friendship continued. When we encountered Katie during the second day of activities she told us in animated tones how the young men had just taught her how to remove the doors from her Jeep for a more open feeling. When asked about her intentions for the future of the relationship she quickly stated that there were no romantic considera- tions but that they all intended to keep in touch and meet at other Jamborees.

Hg6: The Brandfest and the Construct of Brand Community

Analysis strongly confirms H6 for the entire sample; brand- fest participation led to significant increases in overall feel- ings of integration in the Jeep brand community (IBC). Not

evident from an analysis of the quantitative data is that the reasons and the processes behind the strengthening of com- munity differ between the two groups. From the ethno- graphic experience, we learned that novice owners at a brandfest begin to feel more a part of the community as they learn to consume the brand in ways that provide greater ben- efits to them, whether those benefits are utilitarian, self- expressive, social, and/or hedonic. For the more experienced owners, the brandfest provides opportunities to demonstrate and reaffirm their community ties while both mentoring and performing for neophyte owners.

Long-Term Impact of Temporary Communities The longevity of relationships, especially those of an interper- sonal nature, that have their genesis in a brandfest or other brand-specific context should be considered. The quantitative data provide only inferential evidence of long-term impact (i.e., attitudinal measures related to future intentions: "I would buy another Jeep"). Additional evidence is provided by pro- prietary tracking research conducted by Chrysler that suggests that its community-building efforts through brandfests result in significantly increased repurchase rates among participants (Duffy 1999). Our ethnographic data strongly indicate a long- term, lasting impact of relationship-building efforts. We have witnessed many participants, like Katie and Kim, who formed interpersonal relationships at a brandfest that were maintained over a distance and reaffirmed at a subsequent event. As researchers, we have maintained long-term (some as long as eight years) connections with Harley owners, many of whom initially built relationships with the community at a brandfest and still value their connections to the brand community.

We have noted that customers who purchase a branded product often do so with the support of other users, which leads to the possibility of brand-focused interpersonal bonds. Social support from such relationships may, in turn, influence increased personal investment in a customer's consumption of the product and the brand. To the extent that the company behind the brand facilitates such interactions, the customer base is likely to reciprocate with increased appreciation for the company and a sense of being an impor- tant part of a larger set of social phenomena.

Part of the reason for the longevity of customer-centered relationships may be their role as exit barriers. Customers value the relationships available to them as a result of brand ownership. For some customers, the expectation of develop- ing these types of relationships motivates initial product acquisition; they are looking for a sense of community. Other customers acquire the brand for more strictly utilitar- ian or self-expressive reasons and discover the benefits of brand community in the course of consumption. Community ties become exit barriers as owners realize that valued inter- personal relationships would be altered or lost if they were to defect to another brand. Positive relationships with mar- keters and bonding with personal possessions also create exit barriers. The same elements that lend longevity to suc- cessful interpersonal and community relationships, such as reciprocity (Gouldner 1960; Sahlins 1972), investment, commitment, interdependence, and integration in social net- works (Lund 1985), exist in the community of brand, prod- uct, company, and customers.

50 / Journal of Marketing, January 2002

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

Marketing Implications In today's marketing environment, sustaining a competitive advantage on the basis of product differentiation often is an exhausting race to a constantly shifting finish line. Any lead in the race is eroded quickly by imitation or even by superior technology from competitors. One way to sidestep this tread- mill is to redefine the terms of competitive advantage. Part of the success of brands like Jeep lies in their focus not merely on the product or its positioning but also on the experience of ownership and consumption. Differentiating on the basis of ownership experience can be achieved through programs strategically designed to enhance customer-centered relation- ships. The events examined here provide opportunities for consumers to experience anticipated but unrealized product benefits, share those experiences with others, meet with the previously faceless and nameless people behind the brand, and learn more about the brand's heritage and values. Our research demonstrates that such marketing programs can have a measurable impact on the full range of customer-centered relationships. By proactively providing the context for rela- tionships to develop, marketers can cultivate community in ways that enhance IBC and thereby increase customer loyalty.

The benefits to a firm of cultivating brand community are many and diverse. Community-integrated customers serve as brand missionaries, carrying the marketing message into other communities. They are more forgiving than others of product failures or lapses of service qtLality (Berry 1995). They are less apt to switch brands, even when confronted with superior performance by competing products. They are motivated to provide feedback to corporate ears. They constitute a strong market for licensed products and brand extensions. In many cases, we even find loyal customers making long-term invest- ments in a company's stock. Customers who are highly inte- grated in the brand community are emotionally invested in the welfare of the company and desire to contribute to its success.

Limitations and Directions for Further Research

Although our research extols the virtues of cultivating cus- tomer relationships, recent research suggests that customers are becoming overwhelmed by attempts of marketers to engage them in relationship marketing (Fournier, Dobscha, and Mick 1998). As marketers consider cultivating relation- ships with customers, it is important to recognize that rela- tionships take on many forms (Fournier 1998). In addition, marketers should recognize that relationships are reciprocal: Both parties give and receive. For example, the brandfests discussed in our work were events in which managers made conscious decisions not to view the events as strictly short- term investments. The marketers provided experiences, entertainment, and education that customers perceived to be in excess of the costs they incurred to participate. The antic- ipation that customers would reciprocate with increased loy- alty and trust (see Morgan and Hunt 1994) was confirmed by this research.

Our study identifies many issues that merit further research. It would be valuable, for example, to examine how Fournier's (1998) work regarding the character of relation- ships applies to the different relationships formed within the brand community. Further research might consider explor-

ing the characteristics of products and services that make them amenable to the types of relationships we have identi- fied. Our experience in the business-to-business context (e.g., De Walt and Mentor Graphics), for example, indicates that our findings may have applicability beyond the con- sumer market. In addition, in the brand communities we have studied, the focal products are rich in expressive, expe- riential, or hedonic qualities. How might IBC be relevant to less flamboyantly experiential brands? The overlapping and interlocking nature of communities suggests that the real importance of some brand communities may lie in strength- ening ties within other communities, such as extended fam- ilies or towns, which may be more important from the con- sumers' point of view. Research has shown that the ownership of brands of more mundane products such as bath soap, tools, and toys provide valued bridges to family, friends, or neighbors (see Fournier 1998; Gainer and Fischer 1991; Olsen 1993), which suggests the potential relevance of our current findings to these types of product categories.

Our largely exploratory research has inherent method- ological limitations. The practical demands of the field study constrained our ability to examine quantitatively many of the relevant issues on which this research touches. We have, how- ever, identified directions for the development of measurement scales. There is a need to further develop and refine instru- ments that measure the relationships of brand community and integration within it. Within any product category, it should also be possible to create an index of IBC to facilitate the diag- nosis of opportunities for building brand community. We posit that any brand community, weak or strong, exists somewhere on the continua of geotemporal concentration, richness of social context, and overlap with other communities and that the brand's position on these continua can be influenced.

Our research also touches on issues that would espe- cially benefit from the types of insights marketers have sought from qualitative research (see Levy 1981; Winick 1961). It would be valuable to examine, for example, what characteristics lead consumers to value brand community and participate in communal activities (cross-cultural research might be of particular interest). Further research might also be directed toward the executional elements of brandfests that are most effective in cultivating a given brand community. For example, our research indicates the importance of designing events with a focus on socializing neophytes while offering special recognition to those who are already the most integrated in the community. It would also be valuable to examine circumstances that may lead to defection from the community or conflict within it and explore the implications of such issues to community vital- ity and potential impacts on brand positioning or brand equity.

Appendix Scale Development

We used LISREL 8 (Joreskog and Sorbom 1993a) to per- form a CFA of the survey items based on the dimensions of community integration observed in the qualitative data col- lection. The fit of the data to this model was marginal (x2

Building Brand Community / 51

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

with 98 degrees of freedom [d.f.] 520.79, root mean square residual [RMSR] = .0647, comparative fit index [CFI] = .829), and we refined the set of items using the approach suggested by Gerbing and Anderson (1988). The process concluded after we dropped three items: two from the owner-brand construct and one from the owner-product construct. Although we deleted two items from the owner- brand construct, we judged the substantive content to have changed minimally: The final set of items contained two brand association items and three brand loyalty/quality items. At this point, we judged the fit of the data to the model to be acceptable (X2 with 59 d.f. = 197.485, RMSR = .0503, CFI = .917).

The reliabilities of three of the hypothesized constructs (Gerbing and Anderson 1988) were strong (see Table 2). In addition, we calculated the AVE for each construct. The AVE is a more conservative measure than reliability, which examines the amount of variance captured by the construct (Fornell and Larcker 1981). The minimum acceptable value for AVE is .50; all four of these constructs easily exceed the .50 value. By using Gerbing and Anderson's (1988) approach, we created unidimensional measures of each con- struct based on the internal consistency and external consis- tency criteria implied by the multiple-indicator measure- ment model.

Second-Order Model To assess whether these constructs were an adequate reflec- tion of a higher-order construct, the analysis proceeded to test a second-order factor structure (see Figure 3). This model fits the data reasonably well (X2 with 61 d.f. = 223.61, RMSR = .060, CFI = .903). Compared with the CFA, the second-order solution has two more degrees of freedom and a chi-square value 26 points higher. This chi- square difference is statistically significant, but the marginal decrease in the CFI (from .917 to .903) indicates that the change has little practical impact. The second-order load-

ings (Yik) are significant and reasonably uniform across the four first-order constructs (Table 3). In addition, the reliabil- ity of the community-integration construct is strong (.89), and the AVE is good (.68).

The fit of the model and the consistency of the second- order loadings provide solid evidence that these constructs are a reflection of the second-order factor, community integration.

Associations Between Pre- and Post-event Measures

After the structural equation analyses, we directed our atten- tion to determining whether a significant connection existed between the pre- and post-event scores for each component of community integration, as well as the composite measure of community integration.

We performed two tests to better understand the rela- tionship between the pre- and post-event measures. The first test involved an assessment of the importance of a structural parameter linking pre- and post-event community integra- tion in a structural equations model. The second test used a repeated measures analysis to determine if significant dif- ferences in level existed between the pre- and post-event constructs.

Structural equation test. We modeled both pre- and post- event community integration as unobservables (Ti). We assessed the importance of estimating a structural connection between the constructs in the structural model by observing the difference in the fit between the model that estimated P (2,1) and the model in which P (2,1) was fixed at zero (see Figure 4).

Given the longitudinal nature of the data collection, which used the same items as in the pre- and post-event question- naires, it is not unexpected that measurement errors (0E) for these items would be correlated (Jireskog and Sorbom 1993b). Anticipating that correlated measurement error would be added to .the model, we deemed it advantageous to have fewer input variables in the analysis. As we determined the final set of items for each of the four dimensions to be unidimensional on the basis of the criteria for internal consistency and external consistency (Gerbing and Anderson 1988), we summed the individual items to create a single score for each dimension. The input to LISREL was then eight variables: a pre-event and post-event score for each of the four dimensions of IBC.

Correlated error terms were added one at a time, until all pre-item error variances were correlated with the corre- sponding post-item error variances. Adding each correlated error term resulted in a statistically significant drop in the chi-square value and a corresponding increase in the fit indices. It was noted in the process of adding correlated error terms that the modification indices indicated that estimating the structural coefficient between the pre- and post-event community integration would produce a greater decrease in the chi-square value than allowing correlated error between any of the pre- and post-event community integration factors.

The standardized estimate for the structural parameter (P 2,1) and the fit statistics for the final models, which both included four correlated error terms, are presented in Table 4. As correlated error terms were added to the model, the structural parameter connecting the pre- and post-event measures was consistently positive and significant, and fix- ing this parameter at zero resulted in a significant decrease in the fit of the model (Table 4). The comparison of these models indicates a strong positive association between pre- and post-event community integration.

Repeated measures analysis. On the basis of an exami- nation of the pre-event frequencies, it was clear that a subset of the product owners came to the brandfest with high lev- els of enthusiasm for the product, brand, sponsoring organi- zation, and other owners. As a result of these owners' strong positive feelings, it would be difficult for the brandfest to have added significantly to these owners' enthusiasm levels.

Taking this into account, we sorted each of the four com- ponents and overall community integration by the magni- tude of their pre-event scores. Each set was split at the median of the pre-event score, and we performed a repeated- measures ANOVA on the upper and lower half of each com- ponent and overall community integration.

When we examined the ANOVA results, eight of the ten data sets exhibited significant differences in variance between the pre-event and post-event scores on the basis of Levene's (1960) test. When a more variable group is com- pared with a less variable group, GLS regression can be used to conduct a repeated measures analysis (Judge et al. 1985).

52 / Journal of Marketing, January 2002

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

An additional benefit of the regression approach is an easily interpretable measure of effect size, R2. We adapted all data sets to a regression format and estimated those that had sig- nificant pre-event/post-event differences in variance using the GLS procedure in the SHAZAM econometric package (White 1997). We estimated the data sets that satisfied the homogeneity of variance assumption in a regression format using OLS estimation.

Using the approach suggested by Cohen and Cohen (1983), we estimated the percentage of between-subjects

variance and calculated the amount of within-subjects vari- ance. We began this process by regressing the attitude scores on a dummy variable for time (pre-event, post-event) using OLS for the data sets that satisfied the homogeneity of vari- ance assumption and GLS for the remainder. The R2 from this regression is the proportion of the total variance in event scores accounted for by the difference in time (Cohen and Cohen 1983). The proportion of total variance was used to calculate the percentage of within-subjects variance explained by time (Table 5).

REFERENCES Aaker, David A. (1996), Building Strong Brands. New York: The

Free Press. Aaker, Jennifer L. (1997), "Dimensions of Brand Personality,"

Journal of Marketing Research, 35 (August), 347-56. Arnould, Eric J. and Linda L. Price (1993), "River Magic: Extra-

ordinary Experience and the Extended Service Encounter," Journal of Consumer Research, 20 (June), 24-45.

Belk, Russell W. (1988), "Possessions and the Extended Self," Journal of Consumer Research, 15 (September), 139-68.

Berry, Leonard L. (1983), "Relationship Marketing," in Emerging Perspectives on Services Marketing, Leonard L. Berry, G. Lynn Shostack, and Gregory Upah, eds. Chicago: American Market- ing Association, 25-28.

------- (1995), "Relationship Marketing of Services: Growing Interest, Emerging Perspectives," Journal of the Academy of Marketing Science, 23 (4), 236-45.

Bhattacharya, C.B., Hayagreeva Rao, and Mary Ann Glynn (1995), "Understanding the Bond of Identification: An Investigation of Its Correlates Among Art Museum Members," Journal of Mar- keting, 59 (October), 46-57.

Boorstin, Daniel J. (1974), The Americans: The Democratic Expe- rience. New York: Vintage.

Brown, Tom J. and Peter A. Dacin (1997), "'The Company and the Product: Corporate Associations and Consumer Product Responses," Journal of Marketing, 61 (January), 68-85.

Celsi, Richard L., Randall L. Rose, and Thomas W. Leigh (1993), "An Exploration of High-Risk Leisure Consumption Through Skydiving," Journal of Consumer Research, 20 (June), 1-23.

Cohen, Jacob and Patricia Cohen (1983), Applied Multiple Regres- sion/Correlation Analysis for the Behavioral Sciences, 2d ed. Hillsdale, NJ: Lawrence Erlbaum Associates.

Cook, Thomas D. and Donald T. Campbell (1979), Quasi- Experimentation. Boston: Houghton Mifflin.

Cornwell, T Bettina (1990), '"T-Shirts as Wearable Diary: An Examination of Artifact Consumption and Garnering Related to Life Events," in Advances in Consumer Research, Vol. 17, M.E. Golberg, G. Gorn, and R.W. Pollay, eds. Provo, UT: Association for Consumer Research, 375-79.

Deighton, John (1996), '"The Future of Interactive Marketing," Harvard Business Review, 74 (November/December), 151-66.

Denzin, Norman K. (1997), Interpretive Ethnography: Ethno- graphic Practices for the 21st Century. Thousand Oaks, CA: Sage Publications.

- and Yvonna S. Lincoln (1994), Handbook of Qualitative Research. Thousand Oaks, CA: Sage Publications.

Doney, Patricia M. and Joseph P. Cannon (1997), "An Examination of the Nature of Trust in Buyer-Seller Relationships," Journal of Marketing, 61 (April), 35-51.

Dowling, Grahame R. and Mark Uncles (1997), "Do Customer Loyalty Programs Really Work?" Sloan Management Review, 38 (Summer), 71-82.

Duffy, Daintry (1999), "Drive Customer Loyalty," CIO, (August 15), 76-80.

Dwyer, F Robert, Paul H. Schurr, and Sejo Oh (1987), "Develop- ing Buyer-Seller Relationships," Journal of Marketing, 51 (April), 11-27.

Ehrenberg, Andrew (1988), Repeat-Buying, 2d ed. New York: Oxford University Press.

Fischer, Eileen, Julie Bristor, and Brenda Gainer (1996), "Creating or Escaping Community? An Exploratory Study of Internet Consumers' Behavior," in Advances in Consumer Research, Vol. 23, Kim P. Corfman and John G. Lynch, eds. Provo, UT: Association for Consumer Research, 178-82.

Fornell, Claes and David F Larcker (1981), "Evaluating Structural Equation Models with Unobservable Variables and Measure- ment Error," Journal of Marketing Research, 18 (February), 39-50.

Fournier, Susan (1998), "Consumers and Their Brands: Develop- ing Relationship Theory in Consumer Research," Journal of Consumer Research, 24 (March), 343-73.

-------, Susan Dobscha, and David Glen Mick (1998), "Prevent- ing the Death of Relationship Marketing," Harvard Business Review, 76 (January/February), 42-51.

Gainer, Brenda and Eileen Fischer (1991), "''To Buy or Not to Buy? That Is Not the Question: Female Ritual in Home Shopping Parties," in Advances in Consumer Research, Vol. 18, Rebecca Holman and Michael Solomon, eds. Provo, UT: Association for Consumer Research, 597-602.

Garbarino, Ellen and Mark S. Johnson (1999), "The Different Roles of Satisfaction, Trust, and Commitment in Customer Relationships," Journal of Marketing, 63 (April), 70-87.

Gardner, Burleigh B. and Sidney J. Levy (1955), 'The Product and the Brand," Harvard Business Review, 69 (March/April), 33-39.

Gerbing, David W. and James C. Anderson (1988), "An Updated Paradigm for Scale Development Incorporating Unidimension- ality and Its Assessment," Journal of Marketing Research, 25 (May), 186-92.

Gouldner, Alvin W. (1960), "'The Norm of Reciprocity: A Prelimi- nary Statement," American Sociological Review, 25 (April), 161-78.

Granitz, Neil A. and James C. Ward (1996), "Virtual Community: A Sociocognitive Analysis," in Advances in Consumer Research, Vol. 23, Kim P. Corfman and John G. Lynch, eds. Provo, UT: Association for Consumer Research, 161-66.

Grubb, Edward L. and Harrison L. Grathwohl (1967), "Consumer Self Concept, Symbolism, and Market Behavior: A Theoretical Approach," Journal of Marketing, 31 (October), 22-27.

Gruen, Thomas W., John O. Summers, and Frank Acito (2000), "Relationship Marketing Activities, Commitment, and Mem- bership Behaviors in Professional Associations," Journal of Marketing, 64 (July), 34-49.

Gundlach, Gregory T., Ravi S. Achrol, and John T Mentzer (1995), "The Structure of Commitment in Exchange," Journal of Mar- keting, 59 (January), 78-92.

Holbrook, Morris B. and Elizabeth C. Hirschman (1982), "The Experiential Aspects of Consumption: Consumer Fantasies,

Building Brand Community / 53

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions

Feelings, and Fun," Journal of Consumer Research, 9 (Septem- ber), 142-54.

Holt, Douglas B. (1995), "How Consumers Consume: A Typology of Consumption Practices," Journal of Consumer Research, 22 (June), 1-16.

(1998), "Does Cultural Capital Structure American Con- sumption?" Journal of Consumer Research, 25 (June), 1-25.

Jacoby, Jacob and Robert W. Chestnut (1978), Brand Loyalty Mea- surement and Management. New York: John Wiley & Sons.

Joreskog, Karl G. and Dag Sorbom (1993a), LISREL 8. Chicago: Scientific Software International.

and - (1993b), LISREL 8 User's Reference Guide. Chicago: Scientific Software International.

Judge, George G., William Griffiths, R. Carter Hill, Helmut Ltitke- pohl, and Tsoung-Chao Lee (1985), The Theory and Practice of Econometrics, 2d ed. New York: John Wiley & Sons.

Kalwani, Manohar U. and Narakesari Narayandas (1995), "Long- Term Manufacturer-Supplier Relationships: Do They Pay Off for Supplier Firms?" Journal of Marketing, 59 (January), 1-16.

Kozinets, Robert (1997), "'1 Want To Believe': A Netography of the X-Philes Subculture of Consumption," in Advances in Con- sumer Research, Vol. 24, Merrie Brucks and Deborah Macln- nis, eds. Provo, UT: Association for Consumer Research, 475.

Levene, H. (1960), "Robust Tests for Equality of Variances," in Contributions to Probability and Statistics, I. Olkins, ed. Stan- ford, CA: Stanford University Press, 278-92.

Levy, Sidney J. (1981), "Interpreting Consumer Mythology: A Structural Approach to Consumer Behavior," Journal of Mar- keting, 45 (Summer), 49-61.

Lund, Mary (1985), "The Development of Investment and Com- mitment Scales for Predicting Continuity of Personal Relation- ships," Journal of Social and Personal Relationships, 2 (Febru- ary), 3-23.

Markus, Hazel and Elissa Wurf (1987), "The Dynamic Self- Concept: A Social Psychological Perspective," Annual Review of Psychology, 38, 299-337.

McAlexander, James H. and John W. Schouten (1998), "Brand- fests: Servicescapes for the Cultivation of Brand Equity," in Servicescapes: The Concept of Place in Contemporary Mar- kets, John F. Sherry Jr., ed. Chicago: American Marketing Asso- ciation, 377-402.

McGrath, Mary Ann, John F Sherry Jr., and Deborah D. Heisley (1993), "An Ethnographic Study of an Urban Periodic Market- place: Lessons from the Midville Farmer's Market," Journal of Retailing, 69 (3), 280-319.

Moore, Eric G., Sanal K. Mazvancheryl, and Lopo L. Rego (1996), "The Bolo Game: Exploration of a High-Tech Virtual Commu- nity," in Advances in Consumer Research, Vol. 23, Kim P. Corf- man and John G. Lynch, eds. Provo, UT: Association for Con- sumer Research, 161-71.

Morgan, Robert M. and Shelby D. Hunt (1994), "The Commitment-Trust Theory of Relationship Marketing," Jour- nal of Marketing, 58 (July), 20-38.

Muniz, Albert and Thomas O'Guinn (1996), "Brand Community and the Sociology of Brands," in the Special Session, "Com-

munities of Consumption: A Central Metaphor for Diverse Research," in Advances in Consumer Research, Vol. 23, Kim P. Corfman and John G. Lynch, eds. Provo, UT: Association for Consumer Research, 265-66.

and - (2001), "Brand Community," Journal of Con- sumer Research, 27 (March), 412-32.

O'Guinn, Thomas (1991), "Touching Greatness: The Central Mid- west Barry Manilow Fan Club," in Highways and Buyways: Naturalistic Research from the Consumer Behavior Odyssey, Russell W. Belk, ed. Provo, UT: Association for Consumer Research, 102-11.

Olsen, Barbara (1993), "Brand Loyalty and Lineage: Exploring New Dimensions for Research," in Advances in Consumer Research, Vol. 20, Leigh McAlister and Michael Rothschild, eds. Provo, UT: Association for Consumer Research, 575-79.

(1995), "Brand Loyalty and Consumption Patterns: The Lineage Factor," in Contemporary Marketing and Consumer Behavior: An Anthropological Sourcebook, John F. Sherry Jr., ed. London: Sage Publications, 245-81.

Peppers, Don and Martha Rogers (1993), The One-to-One Future: Building Relationships One Customer at a Time. New York: Currency/Doubleday.

Price, Linda L. and Eric J. Arnould (1998), "Conducting the Choir: Representing Multimethod Consumer Research," in Represent- ing Consumers: Voices, Views and Visions, Barbara B. Stern, ed. London: Routledge, 339-64.

and (1999), "Commercial Friendships: Service Provider-Client Relationships in Context," Journal of Market- ing, 63 (October), 38-56.

Reicheld, Frederick F and W. Earl Sasser Jr. (1990), "Zero Defec- tions: Quality Comes to Services," Harvard Business Review, 68 (September/October), 105-11.

Sahlins, Marshall D. (1972), Stone Age Economics. Chicago: Aldine-Atherton.

Schouten, John W. and James H. McAlexander (1995), "Subcul- tures of Consumption: An Ethnography of the New Bikers," Journal of Consumer Research, 22 (June), 43-61.

Sherry, John F., Jr. (1983), "Gift Giving in Anthropological Perspec- tive," Journal of Consumer Research, 10 (September), 157-68.

Stewart, Alex (1998), The Ethnographer's Method. Thousand Oaks, CA: Sage Publications.

Tambyah, Siok Kuan (1996), "Life on the Net: The Reconstruction of Self and Community," in Advances in Consumer Research, Vol. 23, Kim P. Corfman and John G. Lynch, eds. Provo, UT: Association for Consumer Research, 172-77.

Wallendorf, Melanie and Eric J. Arnould (1988), "'My Favorite Things': A Cross-Cultural Inquiry into Object Attachment, Pos- sessiveness, and Social Linkage," Journal of Consumer Research, 14 (March), 531-47.

Webster, Frederick E. (1992), "The Changing Role of Marketing in the Corporation," Journal of Marketing, 56 (October), 1-17.

White, Kenneth J. (1997), SHAZAMfor Windows Version 8.0. Van- couver, BC: University of British Columbia.

Winick, Charles (1961), "Anthropology's Contributions to Market- ing," Journal of Marketing, 4 (July), 53-60.

54 / Journal of Marketing, January 2002

This content downloaded from 129.94.107.249 on Mon, 27 Jul 2015 04:47:30 UTC All use subject to JSTOR Terms and Conditions