market-oriented ethnography: interpretation building and ... 4 market oriented.pdfwendy scheier....

ERIC J. ARNOULD and MELANIE WALLENDORF*

The authors show how ethnography can provide multiple strategicallyimportant perspectives on behaviors of interest to marketing researchers.They first discuss the goals and four essential characteristics ofethnographic interpretation. Then they review the particular contributionsto interpretation of several kinds of ethnographic observation andinterview data. Next they discuss how interpretations are built fromethnographic data. They show how multilayered interpretations of marketphenomena emerge through systematic analysis of complementary anddiscrepant data. Finally, the authors articulate three representationalstrategies that are used to link multilayered interpretations to marketingstrategy formulation. They suggest that ethnographic methods areappropriate for apprehending a wide variety of consumption and usesituations with implications for market segmentation and targeting;product and service positioning; and product, service, and brand

management.

Market-Oriented Ethnography:Interpretation Building and MarketingStrategy Formulation

We address the general question of how ethnographic in-terpretations of the consumption behavior of market seg-ments can be developed and how they can be useful in for-mulating marketing strategy. The term market-orientedethnography refers to an ethnographic focus on the behaviorof people constituting a market for a product or service. Wepropose a systematic process for building market-orientedethnographic interpretation and then describe how these in-terpretations are useful in formulating marketing strategy.Our premise is that studying and interpreting the subjectiveexperiences of the market segments served by specific mar-keting programs is a useful step in establishing enduring, ef-

•Eric J. Amould is an Associate Professor of Marketing, College of Busi-ness Administration, University of South Florida. Melanie Wallendorf is aProfessor of Marketing, College of Business and Public Administration,University of Arizona. Authors' names are listed alphabetically to reflectequal contribution to the article. The authors thank the following people fortheir helpful comments on earlier versions of this article: Merrie Bnicks,Peter Dickson, Deb Heisley, Sue Keaveny, Jakki Mohr, Linda Price, andWendy Scheier. Professor Wallendorf appreciates the helpful comments of-fered by the faculty and doctoral students who attended presentations onthis topic that she made at Ohio State University, the University of Califor-nia at Irvine, Edith Cowan University, and the University of Western Aus-tralia. She also appreciates the support and collegiality she found during hersabbatical at UWA when work on this article was completed.

fective exchange relationships (Denzin 1989) and an organi-zationwide market orientation (Kohli and Jaworski 1990).

In the first section, we emphasize the contributions ofeach of two central types of data collection to building mar-ket-oriented ethnographic interpretation. In the second, weexplain the process of constructing a sound ethnographic in-terpretation from multiple data sources. These sections canbe read from two perspectives: They provide direction tothose conducting market-oriented ethnography, and theyprovide guidance to those evaluating such work. In the lastsection, we link the process of constructing ethnographic in-terpretation to that of marketing strategy formulation by dis-cussing four strategic purposes for which market-orientedethnography is particularly useful.

We differentiate market-oriented ethnography fromethnographies of marketing: Ethnographies of marketingstudy people in organizations carrying out the activities ofmarketing management: planning, product development,and strategy execution (Biggart 1989; Dougherty 1988;Kunda 1992; Mintzberg 1973; Workman 1993); sales activ-ity (Sutton and Rafaeli 1988); and service delivery (Amouldand Price 1993; Hochschild 1983; McGrath 1989; Spradleyand Mann 1975; Van Maanen and Kunda 1989; Whyte1948). Both market-oriented ethnographies and ethnogra-phies of marketing can contribute to theoretical knowledge

Journal of Marketing ResearchVol. XXXI (November 1994), 484-504 484

Market-Oriented Ethnography 485

and marketing practitioner strategies, but in different ways.'Our discussion of the use of ethnography in formulatingmarketing strategy concentrates on the interpretive processand strategic uses for market-oriented ethnographies, leav-ing an opening for a parallel discussion of interpretive tech-niques and strategic implications for ethnographies ofmarketing.

Some caveats are warranted. We do not offer a compre-hensive tutorial that prepares researchers to conduct market-oriented ethnography; as with all complex research meth-ods, this task is beyond the scope of a single source. We donot provide a primer on ethnographic data collection, be-cause extensive materials exist on this topic (Agar 1980;Becker and Geer 1960; Fetterman 1989; Honigman 1970;Jorgensen 1989; McCall and Simmons 1969; McCracken1988; Punch 1986; Snow, Zurcher, and Sjoberg 1982; Ted-lock 1983). Although we suggest issues pertinent to evaluat-ing the effectiveness of data collection and the credibility ofethnographic interpretations, we do not offer an exhaustiveinspection list, nor do we adopt an ethnographic style of ex-position (Clifford and Marcus 1986) given our methodolog-ical rather than interpretive goals.

GOALS OF MARKET-ORIENTED ETHNOGRAPHY

Ethnography is not just a form of data collection; it aimsto clarify the ways culture2 (or microculture) simultaneous-ly constructs and is formulated by people's behaviors andexperiences. Ethnography aims to explicate patterns of ac-tion that are cultural and/or social rather than cognitive (see,e.g., the focus on the sociocultural importance of market an-imation in Sherry 1990a). Ethnography not only establishesthe context and subjective significance (emic) of experiencefor particular groups of persons, but also seeks to convey thecomparative and interpreted (etic) cultural significance ofthis experience (Denzin 1989). To give an account of differ-ences between the world of the group being studied and thatof an audience (scientific, managerial, or popular) that isgrounded in culture, ethnography employs distinctive meth-ods of data collection and interpretation (Tedlock 1983).

Four distinctive features guide ethnographers' researchpractice, help them accomplish these goals, and distinguish

•Frequently in ethnographies of marketing, analysis of the emic inter-pretation of formal rules and the discrepancies between formal rules andbehavior are central to the interpretation. In general, analysis of codifiedprocedures is not a central part of market-oriented ethnographies. Ethno-graphies of marketing may often identify intraorganizational conflicts (e.g..Workman 1993) and interpret whether and how they sustain some aspect ofan organization. Roughly, ethnographies of marketing focus on the supplyside of the exchange transaction, whereas market-oriented ethnographiesinterpret the demand side. Their implications for marketing strategy aretherefore different.

^Culture refers to learned, socially acquired traditions and the lifestyle ofa group of people, including patterned, repetitive ways of thinking, feeling,and acting. Culture can also be thought of as the cumulative total of learnedbeliefs, values, and customs that serve to order, guide, and direct the be-havior of members of a particular society or group. Culture can be thoughtof as that which one needs to know to behave in a manner acceptable to itsmembers. Finally, culture is not a background variable but a dynamic, en-acted, emergent phenomenon. It is a multivocal exchange and an ensembleof behaviors that, rather than being uniform, reflects substantial intragroupvariability. Diversity within particular cultures is enshrined in individualslife experiences embodied by gender, class, age, ethnicity, race, and mate-rial possessions.

ethnographic research for those reading and evaluating suchresearch. First, ethnography gives primacy to systematicdata collection and recording of human action in natural set-tings, in that the deeply embedded sociocultural patterns ofaction it aims to study are resistant to transfer to other re-search settings. In approaching knowledge of behavior inthis way, contemporary ethnography reflects its origins inthe modernist tradition of empirical, interpretive social sci-ence associated with Durkheim, Weber, and Simmel (Gid-dens 1971; Rabinow 1986), rather than the emerging post-modernist perspective (Hirschman and Holbrook 1992).

Second, ethnographic research involves extended, experi-ential participation by the researcher in a specific culturalcontext, referred to as participant observation. Long-termimmersion in context increases the likelihood of sponta-neously encountering important moments in the ordinaryevents of consumers' daily lives and of experiencing revela-tory incidents (Fernandez 1986). Revelatory incidents arenaturally occurring real-time events witnessed by an ethno-grapher that stimulate real-time interpretive insights andlaunch systematic analysis of additional data (e.g., the roleof the opening vignettes in Belk, Wallendorf, and Sherry1989). Because ethnographers follow human action as it oc-curs rather than initiating it, their data collection deploys anevolving sampling plan, takes longer, and is less completelyspecified by a priori design than does much research inmarketing.

Third, ethnography produces interpretations of behaviorsthat the persons studied and the intended audience find cred-ible. In everyday life, culture's mechanisms usually remainunarticulated by participants. People seldom make the sys-tematic connections among their behaviors that are woveninto ethnographies as the experience-distant, analyticallybased, and comparatively informed etic interpretations ofthe researcher. Despite this, ethnographies should be able toconvince the people studied of their credibility (Lincoln andGuba 1984) and the intended reading audience of their trust-worthiness (Wallendorf and Belk 1989). Sanjek (1990) sug-gests three canons for a reading audience to use in assessingan ethnography: theoretical candor, transparent representa-tion of the ethnographer's path through data collection, andaccounting for the relationship between ethnographic inter-pretation and field note evidence. Whereas academic credi-bility can be achieved through narrative virtuosity alone(Crapanzano 1992), ethnographic credibility with the peoplestudied is most likely to be achieved by pluralistic interpre-tations that embrace and explain cultural variation (Joy1991; Tedlock 1983).

A fourth defining characteristic of ethnography involvesincorporating multiple sources of data, a research strategylong advocated in other social science traditions (e.g.,Campbell and Fiske 1959). Rather than use multiple datasources to achieve convergence in interpretations, however,ethnography uses them to generate varying perspectives onthe behaviors and context of interest. Ethnographic interpre-tations are expected to account for (or at least acknowledge)the coexistence of divergent perspectives identified in dataassembled using different methods within a cultural context.

Ethnography does not stipulate a universal sequence ofdata collection methods. The specific sequence of data col-

486 JOURNAL OF MARKETING RESEARCH, NOVEMBER 1994

lection efforts in an ethnographic project is dictated by thenature of the phenomenon (complexity, ubiquity, frequency,and duration), the researchers' prior (undocumented) experi-ence and degree of conceptual understanding of it, and theresearch questions that emerge during the research process.

INTERPRETIVE CONTRIBUTIONS OE OBSERVATIONAND INTERVIEW DATA

Ethnographic interpretation is constructed from twomajor data sources: observation of behavior and verbal re-ports. In this section, we discuss the separate contributionsto interpretation of each, drawing illustrations from severalmarket-oriented ethnographies. We draw more extended il-lustrations from our research on Thanksgiving Day con-sumption (Wallendorf and Amould 1991) than from otherprojects because of (1) the accessibility of the ethnographyto readers, (2) its demonstrated usefulness in formulatingmarketing strategy, and (3) its ability to simplify and providecontinuity to our presentation of interpretation building.However, we refer to other market-oriented ethnographies todemonstrate that the interpretive and strategy-building pro-cesses we illustrate are more broadly applicable.

Naturally Occurring Behavior Observed by a Researcher

Because of its focus on sociocultural patterns of action,ethnography gives primacy to observing behavior andspeech events that naturally occur. Rather than asking peo-ple to comment about what they think they usually do or say,recently did or said, and will do or say, as in phenomeno-logical interviews, ethnographers prefer to observe themdoing it; instead of observing people doing what they mightdo or say if real-world complexities did not impose on them,as in laboratory experiments, ethnographers observe actualpeople's behavior in real time; and rather than asking re-spondents to generalize about their behavior as in survey re-search, ethnographers record the particulars of naturally oc-curring behaviors and conversations.

Ethnographers observe everyday events, settings, interac-tions, conversations, and uses of objects over time andacross specific cases (Jorgensen 1989). These observationsare recorded in field notes and other media and thus becomea data set(s).3 Observation focuses on naturally occurringconstellations of consumption behaviors (Baudrillard 1968;Boyd and Levy 1963; McCracken 1989; Solomon and As-sael 1987) and provides a perspective in action that mani-fests internalized cultural norms, values, and beliefs (Gouldet al. 1974; Snow and Anderson 1987).

Ethnographers assume a variety of roles, ranging fromfull participation to nonparticipation, in response to theoret-ically driven purposes and the dictates of natural events(Adler and Adler 1987). Data from mechanical observationis sometimes used to complement field note data from par-ticipant and/or nonparticipant observation. Next we discuss

'Detailed field notes, as opposed to the hurried scratch notes jotted downas events unfold, are written as soon as possible after immersion in context(Ottenberg 1990; Sanjek 1990). Field notes document action, its setting,participants explanations of their behavior that are given at the time of oc-currence, passages of verbatim conversation, as well as the observers nota-tions about these events (see Lederman 1990 for more discussion).

the contributions of each of these three categories of obser-vation to ethnographic interpretation.

Participant observation. Participant observation is centralto data collection in market-oriented ethnography becauseof the access it provides to the complex behavioral details ofconsumption. It provides details concerning group decision-making heuristics and disagreements, financial negotiations,patterns of product use and substitution, consumers' sponta-neously expressed evaluative judgments (both positive andpejorative), active socialization and indirect leaming, andenactments of culturally pattemed consumption norms andvalues.

As participant observers become insiders over time, theyare granted access to "backstage" areas (Goffman 1959), al-lowing them to leam how consumption behaviors are re-hearsed and performances are scripted. These action settingsmay be protected actively from outsiders; in addition, con-sumers may have limited reporting capability about them.For example, field notes about extended families' prepara-tions for Thanksgiving feasts document complex interweav-ings of behind-the-scenes activity. Their backstage perspec-tives point to an undercurrent of competition for the role ofprovider (e.g., purchaser, server, decision maker, host) thatcoexists uneasily with the publicly performed celebration ofshared abundance (Wallendorf and Amould 1991). A fieldnote excerpt demonstrates this complexity:

It seemed like there was a competition going on be-tween my mom and uncle on whose turkey was better,moister, etc. They even compared stuffing. It wasfunny.... They fought on who stole whose recipe, etc.My mom puts wine in her turkey and my uncle said thatshe stole that idea from him. We put out my mom'sturkey first and he accused us of being traitors. Whichisn't true, but hers was the first to go.... We were all tak-ing votes on whose turkey was the best and it came outthat my mom's turkey won, but my uncle's stuffing wasthe best.

This participant's observations chronicle the complex moti-vational basis for the huge quantity of food purchased andprepared (two whole turkeys in this family). They alsorecord a complex familial adjudication system in whichcompetition is permitted but then resolved by humor and avote that metaphorically names each contestant a winner,thereby restoring family unity. Naturalistic fieldwork helpsilluminate the complex of motivational forces that operatesimultaneously in consumption contexts. Because market-oriented ethnography strives for this complex, textured in-terpretation of culturally constmcted behavior, the accessprovided by participation of the observer is pivotal.

In conducting participant observation, access to differentdomains of meaning is fostered by including research teammembers with varying demographic profiles. Working in bi-gendered teams has compelled us to incorporate the notionof differently gendered worlds of experience in interpretingmarket behaviors ranging from Thanksgiving preparationsto white-water river rafting, the meanings of collections, andpreference formation (Amould and Price 1993; Belk et al.1991; Wallendorf and Amould 1988, 1991; see Bristor andFischer 1993; Wolf 1990).


Ethnography's reliance on participant observation re-sponds to people's inabilities to report fully on the complexinterweaving of culturally significant behaviors. Even will-ing and articulate consumers do not formulate accuratestatements about many clear-cut sociocultural regularities intheir behavior (Harrison, Rathje, and Hughes 1974; Schiffer,Downing, and McCarthy 1981; Whiting and Whiting 1970).For example, the suppression of evidence of manufacturers'and branded packaging at Thanksgiving feasts and their in-clusion at ordinary U.S. evening meals are consumption be-haviors with marketing strategy implications that consumersdo not typically report. Similarly, guides' orchestration ofsatisfaction on commercial white-water river rafting trips istypically not mentioned in customers' oral reports (Amouldand Price 1993). Participant observers, however, can recorddetails about behaviors that illuminate these pattems duringsubsequent interpretation.

From participant obser\'ation, ethnographers can alsogenerate information about what consumers do when at-tempts to attain valued consumption outcomes are frustrated(e.g., the Thanksgiving meal is spoiled by forgotten ingredi-ents, overcooking, or failure to provide leftovers; the ex-tended family quarrels). Systematic analysis of consumptionfailures across sets of field notes elucidates the effectivenessof subsequent attempts at rectification (e.g., fill-in purchas-es, laughter or reassuring comments, disappointment;changes in consumption group membership). For example,researcher Alec Maclver notes a threat to Thanksgiving con-sumption tradition and its resolution through a seeminglymundane but symbolically rich purchase:

1:45 PM: Suzanne and I were sent on a very importanterrand. Angela had noticed that the munchie tray did notinclude their traditional Polish sausage. We went to thenearby ABCO [a supermarket] to purchase this alongwith another package of pretzels.

2:00 PM: We returned with the fulfillment of a Thanks-giving tradition.

Researcher Tamara Torres reports a similar attempt atrectification:

9:40 AM: I start making the jello salad so that it will gelin time for dinner. We usually have this particularsalad.... The salad is made with Lime Jello, whippedcream, and crushed pineapple. As I start to open thepineapple, I notice that it is chunk pineapple and notcrushed. We need crushed. Dad and I debate whether toput the pineapple in the blender to crush it or whether togo and get some from the Shumway's Storeroom Mar-ket, about % of a mile down the road. Dad wins. 'We'lluse the chunk pineapple for the fruit salad and I'll go tothe Shumway's for crushed pineapple.'

Particijpant observation reveals what fill-in purchases aremade and how they sometimes represent dilemmas solvedand family quarrels averted. Examination of the full set offield notes for these events can also help answer the inter-esting question of why Thanksgiving purchases like thesethat are deemed so central are also so regularly forgotten.

Nonparticipant observation. Depending on context andpurpose, ethnograhic interpretation may employ nonpartici-pant observation data. Participant and nonparticipant obser-vation are differentiated from each other by the membershiprole the researcher adopts (Adler and Adler 1987). Nonin-tmsive observation of trace processes (Webb et al. 1966) isa related form of nonparticipant observation sometimes em-ployed in ethnography. In nonparticipant observation, theresearcher observes and records naturalistic behavior butdoes not become a part of unfolding events. For example, aspart of ethnographic research projects, one of us has ob-served people meeting in parking lots to form car pools tomake recommendations for promoting a govemment agen-cy's Park&Ride program, and the other has observed WestAfrican export marketing transactions to recommend strate-gies for improvements in channel performance (Amouldand Iddal 1992).

Nonparticipant observation is especially useful in record-ing small group behaviors (Whiting and Whiting 1970), per-son-object interactions scripted by taken-for-granted pat-tems or what Langer (1983) terms "mindlessness," andthose requiring high levels of participant expertise (Celsi,Rose, and Leigh 1993). Particularly in one-time observa-tions, researchers may decide that the introduction of anovice participant would dismpt the enactment of focal as-pects of cultural scripts without adding to data quality. Forexample, ethnographers have observed parents changing in-fant diapers to recommend changes in diaper design; theirresearch goals could be met without the ethnographers be-coming adept at diaper changing themselves (Alsop 1986).Fischer-Price's use of a playroom for testing potential newtoys and Vidal Sassoon's use of a salon for testing new hairproducts are other examples of nonparticipant observationbeing sufficient to meet some needs of marketingpractitioners.

In the Thanksgiving Day research, we employed nonpar-ticipant observation to contrast household evening mealswith Thanksgiving dinners. Participation by the observer inthe ordinary evening meals would have changed the con-sumption context to dinner with a guest, entailing more in-tentional displays. Although some acting occurred, ob-servers recorded important contrasts between evening mealbehaviors and those associated with Thanksgiving feasts(e.g., eating on the run or in front of the television, three orfewer menu items, serving foods in their branded packagingand while in the kitchen, using fast foods or take-out foods).Thus, despite the ethnographic preference for participantobservation data, nonparticipant observation is sometimesan appropriate tactic.

Mechanical observation. Ethnographers may also employmechanical devices in observation to obtain complementarydata for interpretation building: for example, making pho-tographs, audiotapes, or videotape/film of action during


consumption events."* Because interview dialogue is a non-naturalistic form of data collection, mechanical recording ofit is discussed as part of verbal report data and is not includ-ed in the category of mechanical observation.

Many of the over 25(X) photographs from the Thanksgiv-ing project illustrate mechanical observation's complemen-tarity with participant observation. For example, a photogrounds the previously mentioned participant observer'sclaim for competition between family members. The pho-tolog excerpt includes the fieldworker's caption in quotes aswell as senior researchers' detailed descriptions of thephoto's contents:

Cherry Morris, Photo # 206-30 Medium Closeup. Inte-rior. "Mom is now carving her turkey as the competitionwatches." Fairly intimate shot. Oblique shot over therange toward the end of the kitchen counter right of therange.... In the foreground right, C.E. (mother's brother)is shown in left profile. He is drinking from a dark glassheld in his upraised left hand. Half hidden beyond himis Cherry's mother. She is shown in left profile as shestands in front of the range. She holds a large carvingknife in her dght hand, poised over a half carved breastof the turkey.

The intimate proxemic distance among people shown in thisphoto (which recurs throughout the photo series of theirfeast) conveys something about overall positive affectamong this family, and the caption written by the familymember who is a participant observer describes competi-tion. This photo series contrasts with the public proxemicdistance portrayed in other photos of some divorced andblended families.

Still and moving pictures taken in natural settings in-crease ethnographers' ability to interpret the following:

1. Recording a photo series permits an analysis of the temporalflow of consumption events (see also Heisley, McGrath, andSherry 1991).

2. Photos taken during a consumption event by participants canhelp researchers identify culturally significant moments thatencode shared emotions, meanings, and transitions (Chalfen1987). Those intimate with the consumption event (whetherresearcher or informant) are likely to feel a pull to take pho-tos at these moments, without necessarily being able to artic-ulate the underlying cultural meanings that are enacted. Nu-merous Thanksgiving photos of turkeys coming out of theoven, table displays, and food ready to be served provide re-curring evidence that is useful in interpreting the importanceof abundance and togetherness.

3. Photos of backstage preparations are useful in interpretingthe taken-for-granted cultural scripts that underlie consump-tion sequences—for example, people shopping for Thanks-giving, people in their bathrobes putting the turkey in the

••For discussions of photography in ethnography, see Collier and Collier(1986); Worth and Adair (1972). Concerning their use in consumer re-search, see Heisley and Levy (1991); Heisley, McGrath, and Sherry (1991).For discussion of audiotape usage, see Moerman (1988); Tedlock (1983).Mechanical observation raises issues about the relationship between ethno-grapher and informant. On the one hand, mechanical recording devices,such as a video camera or Nielsen box, clearly mark the researchers inter-vention in informants lives, an intimacy that some may resent. Displays,acting, and distortion of behavior are likely results. On the other hand, therecording device or camera lens can be a shield, constraining the nonpar-ticipant researcher to objectifying informants with either an idealizing, im-perialist gaze or a voyeuristic one (Sontag 1977).

oven. The interpretive value of such photos stems from focus-ing ethnographic attention on moments when participantsmaintain that "nothing (interesting) is going on," but cultural-ly patterned behaviors are still in evidence, for example, back-stage work.

4. As shown in the previous example, photographs can also beused in interpreting the dramaturgy of proxemic expressionsof social division or integration (Amould and Price 1993; Col-lier and Collier 1986; Heisley, McGrath, and Sherry 1991).

5. Photographs document referents that are not fiiUy explainedby verbal descriptions of human-object interactions, includ-ing the aesthetics of display. Such photos reveal what behav-iors informants are referring to when they speak of "dressingup" for New Year's Eve, or say their Thanksgiving Day in-volves "casual clothes" or "lots of food."

Several Thanksgiving photolog excerpts that contain use-fill material for interpreting these five behavioral phenome-na are displayed in Table 1. These excerpts come from asubset of the data that includes references to the use of but-ter or margarine; these products are of interpretive interestas indicators of the interplay between social class and cul-tural capital (Bourdieu 1984). The excerpts demonstrate notonly that photographic data provide evidence regardingthese five behavioral phenomena, but also that the five phe-nomena may co-occur in data entries.

Limitations of and correctives for ethnographic observa-tion. As with all forms of data, observational data takenalone have limitations. Observational data do not provide di-rect access to the perceptions, values, and beliefs of infor-mants and reveal little about informants' intemal states.Only speech-in-action (Richards 1939), unstmctured inter-views, or questioning during participant observation pro-vides such information. Because market-oriented ethnogra-phy aims to explain emic meanings and accounts of behav-ior (as in ethnomethodology; Garfinkel 1967), as well as et-ically derived regularities, ethnographers prefer to expandtheir data sources beyond what can be obtained through ob-servation alone. Ethnographers may privilege observationaldata in building interpretations but can be expected to com-bine them with verbal report data to account for the phe-nomenon of interest more thoroughly.

Within observational data, membership roles availablealong the nonparticipant-participant continuum all constrainthe interpretive process. If researchers become too muchparticipant insiders, they assume the relative lack of insightinto cultural processes that natives may have and thereforeare less able to attend to actions and explanations that na-tives take for granted (Wirth 1964). The resulting ethno-graphic work is likely to lack critical or insightful conclu-sions. This is a particular challenge for North Americanmarketing researchers studying their native culture (e.g.,Pms 1989). However, certain data collection techniques canminimize subsequent interpretive problems: sampling meth-ods that randomize the selection of times and places for ob-servation, deliberate attention to marginal persons and as-pects of events, regular debriefing by informed colleagues,and use of holistic topical checklists to remind the field-worker to detail events seen, overheard, or experienced dur-ing fieldwork (Denzin 1989; Johnson and Johnson 1990;Lincoln and Guba 1984).


Table 1ILLUSTRATIVE EVIDENCE OF THE INTERPRETIVE VALUE OF PHOTOGRAPHS

Temporal Flow

Photo #139-84 This photo is 84th in a series of 89, Medium closeup. Inte-rior. "Mom's ready to eat." Shot down and over the edge of the table to-ward Mom, Mom is seated in full view, hands in her lap. She looksslightly left towards Dad seated at the adjacent side of round, plastic lam-inate table. Her plate is full. The pitcher of gravy stands on the comer ofthe table to her right. To her left is a big crock of Blue Bonnet margarine,salt and pepper shakers, and a bottle of Old Milwaukee beer.

Culturally Significant Moments

Photo #201-08 Medium shot. Interior, "Grandmother at her specified tablesetting." [Fieldworker notes paper decorative napkins, place cards, litcandles, and fresh flowers. She also notes that pats of butter are stackedup in a neat pile on a saucer in the near left side of the table,] There is abutter knife stuck in the top of the pile. Same shot as 201-05, Shotstraight down the length of the dining table. At each plate there is a watergoblet (filled) and a wine glass. The place cards are held up by littleturkey shaped clips,,,, an empty bread or salad plate at each place. Grand-mother stands in the background left leaning forward over the back of herchair. She stands in right profile and looks right toward the camera,,,,wears dark slacks and a white sweater. There is a picture window in thewall behind her left. The light yellow or beige drapes are opened to theright to the comer where the window wall meets the elephant warriorwall. She has white hair and a plump face. She wears a seriousexpression.

Photo #231-10 Closeup, "Cranberry sauce take your pick," Shot left acrossthe table from Nell's place toward the end wall. On the table can be seena china dish of cranberry sauce and a pressed or cut glass dish of cran-beriy jelly straight from the can, as well as a yellow plastic crock of mar-garine or blended spread and a glass relish tray loaded with black andgreen olives, A glass of milk is perched on the far comer of the table nextto a goblet of blush wine. The plate left foreground has a drumstick onit. The plate center background has a heap of mashed potatoes and someolives on it. Wine and milk glasses here too.

Backstage Preparations

Photo #120-17 Medium shoL Interior, "Aunt Em has helped clear the din-ner table and now she is bringing in the dessert plates and tea," Aunt Emis photographed from the dining room, half tumed toward the camera

reaching for a white coffee cup with her right hand. She has removed hersweater. Her left arm is bent at the elbow. She looks toward the diningroom. Grandma is just visible at the kitchen sink behind her. The cupsand saucers have been stacked on the counter where Aunt Em can easilyget them.

Photo #120-18 Medium shot. Interior. This is from the same photo series asthe previous photo, 'Ted and Grandpa wait for the pumpkin pie," Shotacross the table from comer to comer. Table is nearly cleared, only saltand peppers and butter dish, and two empty glasses remain from the maincourse. Napkins lie on the table. Sugar and two cups and saucers havebeen laid beside the men. Ted, hands in lap, looks at camera. Grandpa,hands together on the table, looks down, almost sadly, at the table.

Dramaturgy of Pmxemics

Photo #140-03 Medium, Interior, "Lee and Jan, the hosts, of our Thanks-giving Day dinner; Jan is drinking wine, while Lee looks on." Noticehow members of families sit together in this picture and the last picture.Model sailboat and small Norfolk pine on top of TV. Lee wears jeans,checked shirt, sneakers. He sits on a hassock while Jan sits in the bluechair. There is tan wall-to-wall carpet. Lee sits forward, Jan sits back.

Human-Object Interactions and Aesthetics of Display

Photo #245-03 Long shot. Interior. "The set dining room table," (Field-worker notes: gold & silver on this table.] No people. Shot at an angleacross the width of a formal dining table towards a large, dark wood andglass china cabinet, A lit crystal chandelier hangs over the table. Thewallpaper is pale stripes of champagne and white. There are lai;ge dou-ble, sashed windows left background at the head of the table ... translu-cent drapes at the windows. The table is set with a simple white cloth foreight. At each setting are a cut crystal goblet and a gold knife, spoon, andfork. The former two utensils are placed right of the place settings. Leftof the fork is a folded cloth napkin. There is a large floral centerpiece ofred camations and baby's breath. Left of the centerpiece are ceramicmale and female Pilgrim figures, and a small covered crock. Right of thecenterpiece is a ceramic pumpkin with a removable top; an omate, cov-ered silver butter dish; and glass and silver salt and pepper shakers. Thechina cabinet has a mirrored back to it and on the shelves can be seenlarge china platters and plates with medallions in the centers of them.Heavy, straight-backed wooden chairs with white upholstered seats arepulled up to the table.

"The fieldworker's caption is in quotes and the senior researcher's detailed descriptions follow.

By contrast, ethnographers who restrict themselves tononparticipant observation risk interpretive problems asso-ciated with this membership role. They may not access na-tive experiences that are actively or politely hidden fromoutsiders and may be drawn to describing surface similari-ties rather than accounting for cultural variation; interpreta-tions may tend toward ethnocentrism and stereotyping. Thisis a particular concem in studies of consumers who are de-mographically or culturally distinct from researchers. Cor-rectives include long-term fieldwork, procedures to ensureextensive field note detail, and member checks to access di-vergent emic perspectives (Lincoln and Guba 1984; e.g.,contrasting perspectives of buyers and sellers on the whole-someness versus illegal activity of a swap meet in Belk,Sherry, and Wallendorf 1988).

Participant observation is time-consuming; its costs de-rive from the requirement of long-term researcher immer-sion in cultural context. As with other research approachesthat interpret qualitative data (Stem 1993; Thompson, Lo-cander, and PoUio 1990), effective ethnography cannot beconducted rapidly by part-time interviewers, nor can it be

conducted solely by student research assistants who "mn in-formants" for a principal investigator. Although students, re-search assistants, or hired interviewers may assist an ethno-grapher or participate in data collection, ethnographic re-search cannot be conducted without time-consuming, first-hand immersion in data collection by the principal investi-gator. The process by which a researcher gains access to andunderstanding of the cultural context improves participant-observation insights but limits the number of settings a re-searcher can study. In commercial settings, these constraintsare likely to result in reliance on firms or consultants withextensive prior experience and expertise in participant-observation research. Nonetheless, with careful samplingand interpretation building, what is lost in breadth of cover-age is compensated for by interpretive depth regarding so-ciocultural pattems of action.

Verbal Reports Elicited in Ethnographic Interviews

Verbal reports are the second major form of data em-ployed by ethnographers. Verbal report data elicited inethnographic interviews constitute selective memories and


prospections, based on informants' emotions, experiences,and expectations. In building an interpretation, ethnogra-phers do not necessarily use people's words about their be-haviors as accurate accounts of behavior. Instead, verbal re-ports from interviews are relied on to provide emic perspec-tives of action: people's value-laden stories and accounts oftheir own and others' behaviors (Agar 1986; Gould et al.1974; Moerman 1988; Snow and Anderson 1987; Tedlock1983). Through their stories and accounts of their behavior,participants recall, interpret, script, and give meaning toconsumption events.

Verbal reports elicited in response to researcher-initiatedquestions and probes do not chronicle naturally occurringbehavior. Whether recorded mechanically or not, ethno-graphic interviews are not naturalistic (unless, of course, thebehavior being studied is how people behave in interviews).Verbal report data serve a different purpose in constructingethnographic interpretation than do observational data. Thetwo types are not interchangeable or roughly approximatemeasures of the same phenomena, because each measuresdifferent phenomena. Verbal reports typically supplementobservational data in ethnographic interpretation, providingemic, culturally particular understandings to interpretation.Variations in degree of structure produce two frequentlyused types of ethnographic interviews: unstructured inter-views and structured surveys.

Unstructured ethnographic interviews. Unstructured in-terviews have been the object of considerable renewed in-terest in marketing research, as both a part of ethnographicwork (Hill 1991; McCracken 1989) and a stand-alone datacollection strategy to accomplish goals other than those ofethnography (Levy 1981; Rook 1985; Thompson, Locander,and PoUio 1990). Unstructured interviews typically involveindividual informants (although group interviews may beconducted) participating in a conversation with a researcher,guided by general rather than highly specific a priori topicalstructure, controlled in greater measure by the researcherthan is naturalistic observation. Some ethnographic inter-views are formally designated by both the researcher and in-formant as an arranged interview, in which an interviewerelicits detailed emic accounts. Others arise spontaneouslyout of informal conversations between participants and re-searchers, or researchers' questions about participants' be-havior during observation.

Typically ethnographers design and define topics butallow informants to provide interview content (Denzin1989; Fetterman 1989; McCracken 1988). Because unstruc-tured interviews are used to elicit emic meanings, open-ended probes are used to increase the likelihood that the re-searcher discovers how informants construct their world(McCracken 1988). The researcher's task in an unstructuredinterview is to use various probes in a way that builds a con-versation-like dialogue (Bitner, Booms, and Tetrault 1990;McCracken 1988; Snow, Zurcher, and Sjoberg 1982), ratherthan asking questions that impose categorical frameworkson informants' understanding and experiences.

Structured surveys in ethnography. At the other end of thespectrum in terms of structure, ethnography may incorporatedata collected through structured surveys. In appearance andconduct, these resemble surveys used in much market re-

search. They contain structured questions and may includequantitative response scales that are statistically analyzed.However, the role of structured surveys in ethnography dif-fers from their use in nonethnographic market research intwo ways: First, the topics, themes, and scale items forethnographic surveys are usually generated from emic cate-gories derived from previous participant observation and/orunstructured interviews with the population of interest,rather than from literature-based theoretical propositions orpreviously developed scales (see Bearden, Netemeyer, andMobley 1993); and second, ethnographic surveys are used tosupplement observational methods rather than as a stand-alone data collection technique. Ethnography relies onstructured surveys to provide a perspective of action, but nota necessarily veridical indicator of behavior. Ethnographymay supplement participant observation data with structuredsurveys of larger, carefully selected samples (Honigman1970) concerning variation in readily quantified phenomenasuch as demographics, household budgets, gift exchanges,resource endowments, and input-output measures (Bennettand Thaiss 1970).

Contribution of Verbal Report Data to EthnographicInterpretation

Verbal report data contribute perspectives of action toethnographic interpretation. In addition to recollections ofspecific behaviors, verbal reports can include (1) uses of in-stances of behavior to support overgeneralizations, (2) re-ports of behavior that are metaphorically glossed by themeanings of that behavior, and (3) claims of idiosyncracybased on ostensible contrasts between one's own experienceand that of others. Rather than dismissing verbal data aboutbehavior as unreliable, ethnographic interpretation accountsfor emic meanings evident in informants' stories, overgener-alizations, metaphoric glosses, and claims of idiosyncracy.

Overgeneralizations. Verbal reports from ethnographicinterviews often contain overgeneralizations—verbal ac-counts that understate the amount of variation in actions re-ported, despite informants' having witnessed such variation.Even interview material that purportedly describes actualexperience is potentially distorted by what informantswould like to experience or recall. Many potential overgen-eralizations occur in the unstructured interview data con-cerning Thanksgiving, as in the exchange between re-searcher Karen Celestino and her informant, PeggyMorrison:

Karen Celestino: I would just like you to tell me aboutyour Thanksgiving holiday—anything you would liketo talk about; please feel free to elaborate.

Peggy Morrison: Well. My daughter and I always gettogether; we are both single mothers. We take turns asfar as houses go ... and we always have a turkey. That isthe tradition. And the guys always watch the game. Wedon't mind because they stop for dinner.... Well, theturkey is a tradition but we always experiment with thedressing. Depending on whose house it is at. We havejello salad, mashed potatoes, creamed carrots, and al-ways pecan pie for dessert. I always clean up—mydaughter helps if it is her house. The men never do—oh,I take that back, Steve helps clear. But give me abreak—the men just eat. My two grandchildren come....

I^ari<et-Oriented Ethnography 491

Sometimes I wish they didn't because the seven-year-old is a pill. She never eats. She always complains thatshe doesn't like the food.... Well, I like candles and anice table setting with table cloth and a sit-down dinner.My daughter usually has a buffet.... I think everyonelikes mine better. I do. It is more formal, I like that. Mydaughter always serves finger foods and cocktails be-fore; I think that spoils the dinner.... Food plays a bigpart in both types of celebrations. Grace is started byone of the kids and we all take tums saying something,then dig in. Dinner is long—two hours and we wait tohave dessert; everyone sits around and talks and catch-es up. I really enjoy it. We all have coffee—not the kidsthough, they hate it. That is it... Well, when it is at myhouse we always eat on time—2:00 or 3:00, but every-one knows they had better be there if they want to eat—I don't mess around, and I don't wait!!... I like knowingwhen I am going to eat. [italics added]

The description takes the form of a conventionalized nar-rative describing actual behavior. Yet the declarative toneand the words we have highlighted cue the ethnographer thatPeggy is probably overgeneralizing the incidence or preva-lence of certain behaviors. Her statement uses overgeneral-izations to reflect her idealized goals of inclusion and sta-bility over time by masking these goals as accurate reportsof actual behavior.

Rather than view the informant as deceptive or inaccurate,ethnographers use overgeneralizations to point to the infor-mant's cultural purposes: Overgeneralizations can smoothinstability, repair domestic cleavages, and clarify relations,thereby making future action seem more predictable andobvious.

Glosses. Rather than describing what an observer mightsee, verbal report data include glosses, which are infor-mants' metaphors for depicting events or descriptions of ac-tion entangled with their perspective of what the eventsmean. Glosses encode the metaphoric meaning or sociocul-tural significance of behavior in the guise of an actual report(e.g., "he was putting pressure on me," "we got dressedup"). Glosses do not detail what actually happened in theway a film script would; they describe the meaning of thebehavior to the person providing the explanation. As such,they help the ethnographer understand the speaker as muchas, or sometimes even more than, they help describe actualevents.

Glosses emerge when meanings are so taken for grantedthat their behavioral referents remain unarticulated and un-examined. For example, unstructured interviews aboutThanksgiving Day include frequent reference to foods thatare "made from scratch" or "homemade" without detailingthe actions indicated by these emic categories. This gloss isexemplified in an excerpt fi-om researcher Colette Gruen-ert's interview with informant Lydia Rector:

Colette: What tyjjes of food do you eat at Thanksgivingdinner?

Lydia: Well, everything was homemade, because mygrandmother made everything from scratch. She madestuffing, mashed potatoes, cranberry sauce, rolls, andturkey. My uncle would carve the turkey every year nomatter what, [italics added]

Informants' recurrent mentions of making "homemade"foods or foods "made from scratch" convey a consequentialvalue connected with consumer satisfaction with the holi-day. Informants did not detail the specific food preparationtasks they employ to transform purchased bulk commoditiesand branded products into dishes they call "homemade."Such verbal data may contain identical reports from manyinformants that mask variations in the details of actual per-formance. What is transformed and how it is transformeddiffers in these numerous emic reports; nonetheless, there isa common cultural meaning operating across cases.

In other glosses, informants refer to the prevalence of cer-tain behavioral patterns, despite lacking access to informa-tion permitting them to draw such conclusions or seeminglyignoring information that challenges their judgments. Be-cause they have not done systematic research, comparativestatements by informants can only articulate what they be-lieve to be true. Rather than providing evidence of the exis-tence of these comparative patterns, material containingthese glosses provides evidence about informant beliefs.

In the Thanksgiving Day research, unstructured inter-views contain substantial material in which informants re-port or imply that they are similar to others in the way theycelebrate die holiday. As shown in Table 2, analysis of theverbal report data indicates that informants gloss the varia-tions that exist between their own celebration and that ofothers, preferring to focus on similarities by saying they dothe same thing as "everyone," serving "the traditional meal."Yet by rummaging across interview field notes (McCracken1988, p. 33), we discover the many varieties of householdconsumption encompassed by these glosses. What is impor-tant in interpreting verbal report data is the emic importanceand meaning of the glossed versions of behaviors informantsreport, not the level of accuracy of the reports.

Glosses may also be present in survey data. Survey re-spondents are likely to report those behaviors that they per-ceive as routine. For example, one Thanksgiving researchsurvey question asked respondents to check a list of Thanks-giving activities in which they participated, with an "other,please specify" category included at the end. In the "other"category, respondents reported shopping, watching televi-sion, and selecting a Christmas tree; however, they did notreport taking walks, taking and showing family pho-tographs, telling stories of past Thanksgiving Day mishaps,or giving Thanksgiving food to pets. Nonetheless, these un-reported behaviors were noted with regularity in participantobservation data. Survey informants were reporting aglossed version of their Thanksgiving activities that only in-cluded what they perceived as general and routine.

Claims of idiosyncracy. Verbal reports may include infor-mants' comments about what they regard as idiosyncratic orunique about their experiences. Claims of uniqueness are es-pecially common with U.S. informants, who pride them-selves on individuality and a strong sense of self (Triandis1989). However, lacking a researcher's access to a cross-section of accounts, informants may describe an experienceas idiosyncratic and surprising, despite its consonance withrecurrent behavioral patterns of which they are unaware(e.g., the strength of weak ties phenomenon identified byGranovetter 1973). As a result, ethnographers consider in-


Table 2ILLUSTRATIVE VERBAL REPORT DATA INDICATING GLOSSES

Verbal Report Comments About Presumed Similarity Across Households

Dave Hawthorn: Why do you think your family celebrates Thanksgiving?Hester Duncan: Same reason everyone else does, I guess.Dave: Which is....

Hester: Its a chance for everyone to get together because of the holiday.Nel Raleigh: So what you're saying is that it's an enjoyable time to look at

your blessings and spend time with each other. Are there any other strongfeelings about Thanksgiving?

Mrs. Jones: I feel everybody sort of feels the same about Thanksgiving. It'sa real special holiday, because we are celebrating our founding fathers. Itwas how they celebrated what God had done for them, and that's why wekeep the tradition up. It's a religious holiday....

Nel: Is there any specific memory about Thanksgiving that you can thinkof?

Mrs. Jones: No, we're very common. I really enjoy planning for it. ThisThanksgiving we're going to Anne's in California. It'll be the first timefor a Thanksgiving like this. I'll probably be in charge of planning thewhole dinner. There will be quite a few of' us over there. It will be a nicechange.... We always have a traditional dinner. I remember one timewhen we had a little program before dinner. Our daughter Quinn recitedthe 100th Psalm, and everyone did a little thing.

Derek Scheulen: One parent? Are your parents divorced?Ken Moore: Yes they are. I generally spend Thanksgiving day with my

mother, and a couple of days later do it all over again at my father andstepmother's house.

Dereic: Did one of those two occasions have more meaning than the other?Ken: No, not really.Derek: Were the meals the same, or different between the two?Ken: No, both were the traditional turkey and dressing. Neither cook was

better than the other.

Verbal Report Evidence of Differences Across Households

Nancy Spring: Tell me what you will be cooking.Suzi Stone: Barbie's [her daughter] favorite is the cranberry dessert that 1

make with the graham cracker crust. I usually make turkey with dress-ing, different gravies, mashed potatoes and yams, cut up vegetables, rel-ish tray, pumpkin pie, and olives.

Walter Scott: What are some of the typical things they prepare?Leticia Johnson: We will usually have turkey, dressing, macaroni and

cheese, and many vegetables, pies, cakes, etc. Just the works!!!Walter: What's your favorite, Leticia?Leticia: Oh, I like just about everything. I really look forward to Thanks-

giving dinner.Walter: You mean your mom and aunts don't have any special recipes?Leticia: I just like it all. I usually make a lemon pie that everyone really en-

joys. It goes so quick that I usually make 2 of them. One of my auntsmakes a delicious peach cobbler and my mom is known for her sweetpotato souffli.

Hetty Perry: What food is served for Thanksgiving dinner?Alice White: For dessert we have pumpkin and mincemeat pie. For meat we

have both turkey and duck. We also have relishes, sweet potatoes, stuff-ing, and creamed cauliflower. Its a big production.... We drink wine andkir. Kir is creme de cassis and white wine.

Alexandra Yeltsin: What did you eat?Taylor O'Brian: Reejeenies [stuffed cheese balls that are breaded] that

grandma makes that are the best. Turkey, which Dad and Uncle slice.Mostly all that traditional stuff.

Alexandra: Like what?Taylor: TXirkey, stuffing, artichokes, marshmallow salad stuff. But the best

eatings are what grandma makes, like the pies, which she starts cookingtwo weeks before.

Mary Hecht: What would you serve at your Thanksgiving feast in your newhome?

Susie: Oh the usual. Turkey of course, with all the trimmings.Mary: What do you consider all the trimmings to be?Susie: Mmmm let's see. Pumpkin pie, sweet potatoes, mashed potatoes,

combread stuffing, what else?... oh yeah, good old cranberry-orange rel-ish. That's really about it.

Mary: Sounds pretty traditional to me.Susie: Yeah. But I guess I forgot my additional traditional element...frozen

margaritasW.

Rosemary Briggs: Your two favorite dishes are green jello and green peas?Do you always bring the same dish each year?

Angela Hogan: (Laughing) Yes, every year I bring this green jello with wal-nuts, cream cheese or whipped cream, and other stuff mixed in. I alwaysput it in this huge dish that is shaped liked a Christmas tree, and my momhates the dish because it takes up so much room...

formant claims of idiosyncracy as indicators of infonnantbeliefs rather than as veridical statements about rarity. Inbuilding an interpretation, ethnographers may determinethat some emic claims of idiosyncracy actually representcultural patterns or regularities.

For example, in the Thanksgiving Day research, numer-ous informants said or implied that elements of their mealsare "special" or different from that served by other people orat other times of the year, as illustrated. Referring to partic-ular inclusions or behaviors this way is not accidental; it fitswith a more general experience of holiday consumption as"special," and consumers' expenditure of imagistic-emo-tional energy to create that experience (Hirschman and Hol-brook 1982). People use a variety of consumption objects(e.g., spaetzla, cookies, yams, fruit salad, a special glass,turkey plates) to embody this meaning. Through excerptssuch as those in Table 3, we learn not only what is con-sumed, but what it means.

As the ethnographer moves across a large number of in-terviews, claimed idiosyncracies may be interpreted etically

as reflecting common, recurrent values or consumption pat-terns. Consider two more informant tales that initially ap-pear to involve idiosyncracy but later are interpreted as re-flecting common values. The first involves a recipe:

Priscilla: ...and Mom would bring the mashed potatoes.Made the same way. Take the potatoes, peel them, oilthem, mash them and add butter and junk and a littlesalt and pepper. And that's it, and mash them and leavereal little lumps in them, real little lumps. And we callthem homemade lumpy mashed potatoes. And grandmastarted making them and my brother and I fell in lovewith them when we were real little kids, and so it is tra-dition to make homemade lumpy potatoes and home-made gravy.... I watch my grandmother every year andI still don't know how to make it. [italics added]

The informant reinforces the idiosyncratic qualities of thesepotatoes by identifying their origin (" 'grandmother") andthe mysteries of their making (" 'I still don't know how"). Inanother story, an informant described what was regarded as


Table 3ILLUSTRATIVE VERBAL REPORT DATA CONCERNING

SPECIAL OUALITIES

Morton Parker: What else happened?Eleanor Matsushita: My dad would always watch the games on TV and the

kids would help clean up. My grandma would always try to serve thefood while everyone else was eating. We would all tell her to sit down,but she never seemed to listen. Our Thanksgivings were not very con-ventional. My grandma is Polish and she always made spaetzla, which isnoodles and gravy and it is really good. We never had stuffing or pies oranything like that.

Morton Parker: Do you eat turkey at least?Eleanor Matsushita: Yes, we have turkey, but we also have wast beef and

all kinds of other foods like cream of spinach, which I love. It's like a bigfeast.

Morton Parker: What is your personal favorite?Eleanor Matsushita: The spaetzla my grandmother used to make was the

best. My mother makes it now, but it's not the same. I even make it some-times, but it never seems to be as good as when my grandma made it.

Morton Parker: So she had that special touch?Eleanor Matsushita: Yes, she also made sauerkraut which I also loved. We

would always take home so much food that there would be enough to lastour whole family for at least a couple of days. I remember I would havepeas and sauerkraut for breal^ast and it would remind me of the goodtime I had the day before.

Morton Parker: What are some other foods that you like a lot?Eleanor Matsushita: Well, my parents liked the more traditional foods like

cranberry sauce and things like that, but I liked things like pickles andthese cookies my grandfather used to make. He owned a bakery and heused to have these special cookies made up for Thanksgiving. They werein the shape of a turkey and they had a really good frosting on them thatI really can't describe. But since my grandma died, he is really sad anddoesn't make them anymore.

Hector Ybarra: Do you always have turkey?Luther Chavez: Always! There is always left over for a couple of days.

They only make special dishes that day that they don't really eat duringthe rest of the year (e.g, yams, fruit salad).

Barbara Herdon: Tell me about Thanksgiving with your family.Dutch Ackerman: When I got older, I got to partake in the family tradition

of drinking the wine. Cold Duck. Before I was old enough, I had to drinkthe Koot Aid in a special glass.

Karen Hunt: We eat dinner with turkey plates.Dave Befu: What is the turkey plate?Karen Hunt: Those are china with turkey's picture in the middle. My

grandma has a set. My mom and my aunt have each set, also.

an unusual form ("special occasion") of consumer behaviorassociated with the turkey:

Janice: Yeah, it's [the food] really good. My mother isan excellent cook and we really have a lot of fun. (Gig-gle) I remember when I was a little kid I always want-ed to shift gears. This sounds really funny, but it is true.When I was little my mom would get up in the morningand start making dinner. And I would always get upwith her; my mom would usually wake me up, becausethis has been going on for about six years now, and aftera few years my mom would wake me up for this specialoccasion. She would clean the turkey and then wake meup. I would run out into the kitchen, I was so excited! Iwould stand in front of the turkey, put both of my handson the legs and start shifting the gears as if I'm drivinga car, and it was just so funny. My family would get atotal kick out of it. Everyone would come in and startlaughing.

Dana: Do you still do that now?

Janice: To tell you the truth, last Thanksgiving when Iwent home, my mom came in and woke me up and said,"Janice, just for old time's sake, come out and shift thegears on the turkey." I jumped out of hed, ran to thekitchen just like I used to and I shifted those gears. Mymom and dad laughed so hard there were tears in oureyes. It was funny, it was really neat, you know, doingthat after all these years.

Dana: So you just had to do it one more time.

Janice: Yeah, I did. And it was really fiin. It was worthit, seeing a smile on my parents' faces, [italics added]

Informants interpret such occasions and traditions literallyas idiosyncratic. Yet by rummaging across unstructured in-terview data, we interpret their occurrence across casesmore figuratively. Etically, we interpret these as times whenimperfect (e.g., lumpy mashed potatoes) or innovative (e.g.,special cookies, shifting gears) consumption behaviors aregiven the not-so-unusual meanings of celebrating long-standing family traditions ("for old time's sake"), affect as-sociated with temporary family togetherness ("tears in oureyes"; "really fun"), and situational consensus about famil-ial authority ("my mom came in and woke me up" [youngadult female speaking]). Although certainly not everyonepretends to "shift gears" on a turkey or takes pride in "lumpymashed potatoes" etically we interpret these behaviors ascommon, culturally constructed rites of integration (Triceand Bayer 1984).

Other claims of idiosyncracy pervade each ethnographicproject we have conducted among Americans. For example,white-water river rafting customers often credit the trip withbringing their particular family closer together. Etically, theresearcher is challenged to account for the features of com-mercial river trips that recurrently produce this unexpectedsubjective experience (Amould and Price 1993). In everycase, a strategic concem for marketers is how to deliver thevalues and meanings represented in such verbal reports.

Limitations of and correctives for ethnographic verbal re-port data. Although verbal reports richly detail emic per-spectives, their key methodological limitation in meeting thegoals of ethnography is their inability to register actual be-havior in context. As ethnographer Margery Wolf (1990, p.351) says, "Rich as I believe these interviews are, they arefrozen in time, individual statements only vaguely anchoredin the social and historical context that created them." Un-like context-rich, historically informed, and relationally sit-uated observational data, verbal report data often overlookphenomena that are politely or purposefully hidden or thatthe researcher does not know about or think to ask about.What respondents report about their behavior is often, for avariety of reasons, not entirely consistent with what is ob-served by the researcher. Rather than regarding informantsas intentionally deceptive—though this is always a possibil-ity (Dean and Whyte 1958)—ethnographers regard and in-terpret informants' verbal reports about behavior as situated,particularistic, and motivated.

Because of the inherent inconsistencies and ellipses inoral reports, verbal data alone are not regarded as sufficientfor developing ethnographic interpretation. The preferredcorrective for these limitations is combining verbal reportdata with data from long-term participant observation in cul-


tural context. We now discuss the challenge of combiningmultiple sources of data into a credible ethnographicinterpretation.

USING MULTIPLE METHODS TO BUILDETHNOGRAPHIC INTERPRETATION

Moving from Data Collection to Interpretation Building

During data analysis and interpretation building, ethnog-raphers move to iterative engagement with data now in-scribed in field notes and recordings (Lederman 1990). In-terpretation aims to make sense out of records of observedbehaviors and verbal expressions of subjective experience; itilluminates the cultural meanings and conceptual structuresthat direct and are constructed from particular individual ex-periences (Geertz 1973). Wolf's (1990, p. 347) descriptionof her iterative interpretive analysis of ethnographic data istypical:

The presence of unfocused, wide-ranging, all-inclusivefield notes was essential to the success of this un-planned [writing] project, hut so were the purposefullysubjective "data" recorded in my journal, and the so-called objective data recorded under the stopwatch inthe child observations. From parts of each of them Ipieced the puzzle together.... [This is] the value of usinga variety of methods to record details and conversationsthat may or may not seem to make sense at the time.

Rather than a truly separable stage in a sequence, the pro-cess of interpretation building typically begins in the fieldwith a set of field note entries, including records of revela-tory incidents and other working hypotheses or ideas aboutrecurrent pattems. Data analysis and interpretation buildingbecomes more intense and focused when the ethnographerhalts fieldwork and continues to code, sort, and index data.Analysis may sometimes necessitate a return to the field togather additional data to address conceptual gaps noted dur-ing initial coding and indexing.

One of the biggest challenges of ethnographic interpreta-tion is combining data obtained through multiple methodsinto a credible account. Ethnographic interpretation buildingis sometimes described as demanding sufficient data to iden-tify the themes that summarize recurrent emic understand-ings and behaviors (Denzin 1989; Fetterman 1989); howev-er, it demands much more than that. Ethnography demandsthat its data provide the empirical grounding for an etic rep-resentation that accounts for the cultural significance of dis-junctures between informants' understanding and their be-haviors. Because ethnographic interpretation goes beyondmerely cataloging repetitive emic themes, the decision toend fieldwork is more complex than merely verifying repe-tition in emic understandings. Both redundancy and varietyin the data are needed to permit a dynamic emic representa-tion that marks pattems of similarity and difference. In ad-dition, redundancy and variety in the data are needed for de-veloping an etic interpretation that adds critical commentaryand builds theoretical insight from emic perspectives.

In positivist research practice, multiple methods are usedto establish convergent validity (Anderson and Shugan1991, p. 24, fn. 2; Campbell and Fiske 1959). Typically, theyare deployed, along with multiple measures obtained using

one method, with the goals of providing internal consisten-cy, test-retest reliability, reliability across cases, and refiningresearch instruments (Bearden, Netemeyer, and Mobley1993).

In contrast, multiple methods of data collection are usedin ethnographic research to access different realms of expe-rience that may diverge from each other. We use the termdisjunctures to refer to differences between the perspectivesprovided by various ethnographic data sources. Rather thanasking which type of data provides greater validity, ethno-graphic research uses each data type to give voice to a par-ticular perspective on behavior (see Sanjek 1990). Assessingthe credibility or trustworthiness of an interpretation (Wal-lendorf and Belk 1989) hinges on whether the data providessufficient conceptual depth, repetition, and variation in bothperspectives in action and perspectives of action to accountfor convergence and disjuncture between the two.

In considering data adequacy, ethnographers strive forcredible rather than exhaustive interpretations because cul-tural meaning eludes exhaustive appraisal. The challenge instudying culture and language is that they are generative andinfinite, rather than conventional and finite, in their mani-festations. In addition, interpretation is rendered partial bythe fact that social science is historically contingent and isitself infused with cultural meaning (de Certeau 1984; Fou-cault 1972; Merieau-Ponty 1962). Assessing credibilityrather than exhaustiveness of interpretation acknowledgesthat ethnographers, like all social scientists, are limited bytime and the data collection devices deployed. As such, theprocess of building credible ethnographic interpretation canbe understood as an iterative effort to articulate and accountfor disjunctures.

Accounting for Disjunctures

Ethnographic research develops an interpretation by com-bining observational and verbal data in a way that accom-modates or accounts for variation between them. Interpreta-tion can begin with verbal reports and then turn to recordedbehavioral observations to check for disjunctures and con-vergences. Alternatively, interpretation can use verbal re-ports to account for the emic meanings of recurrent pattemsof behavior and deviations from expectations noted in ob-served behavior. Interpretation may also begin with a reve-latory incident, using both observational and verbal data tocontextualize and explicate its cultural significance. Al-though each source of data has its limitations, together theyprovide a potent basis for penetrating the cultural meaningsenacted in particular consumption contexts.

In this section, we illustrate how interpretation is builtwhen a disjuncture exists between observed behavior andthe overgeneralizations, glosses, and claims of idiosyncracyin verbal reports. Although disjunctures are not limited tothese three types, they illustrate the way identified disjunc-tures are a starting point for interpretation building. We thensummarize the sequential interpretive processes of codingand building layers of meaning through troping (as depictedin Figure 1) in more general terms.

We emphasize at the outset that, in practice, the interpre-tive process is not the overly simplified linear process por-trayed by written language and a summary figure. However,


Figure 1CODING, TROPING, REPRESENTATION, AND APPLICATION

IN MARKET-ORIENTED ETHNOGRAPHY

QUALITY SPACE (•mio)BEHAVIORAL CONSTELLATION (•tlo)

B

JCOOEil CONfTiniCT,

"̂ t1

1

11

1

1 ''1 '' '••C0NSTmCT2

//

//

C0DB31 cowmucT,

'4i| ^ hl|11(11

f

(

CODEf I CONtTKUCT 4

f REPRESENTATION 1 §

REPRESENTATION 2 %

STRATEC3Y

THEORY

neither is the interpretive process chaotic, random, andwhimsical. For our purposes, we endeavor to illuminate theprogression of interpretation building using words, a figure,and a mathematical notation; yet interpretation is not a de-terministic, prespecified process. As in all data analysis,ethnographic interpretation also requires researcher creativ-ity and insight rather than rote adherence to a sequence ofsteps.

Overgeneralizations. Any type of verbal ethnographicdata may contain statements that can be identified as over-generalizations. At first, field notes containing informants'verbal generalizations (not yet known to be overgeneraliza-tions) are coded to note their referents and their recurrenceacross informants. In the Thanksgiving research, this meantthat uses of several types of absolute descriptors regardingwhat happens (e.g., "always," "never," "everyone") weremarked for subsequent computer searches using the codeALWAYS. (We use capital letters to differentiate codes fromverbatim wordings of field notes, shown in quotationmarks.) This first step in interpretation building uses codesto note phenomena (words, phrases, complex behavioral se-quences, or meanings) that recur across informants, eventhough the cultural importance of the recurrence may not yetbe fully understood. Because the recurrence may not yet befully understood, the codes are used merely to mark similarpassages and are not intended to be conceptual representa-tions of their meaning. Initial codes represent different lev-els of abstraction; some represent emic categories, and someemerging etic understandings.

Next in the interpretive process, multiple data sources areused to identify disjunctures to begin to build the empiricalgrounding for an etic understanding. For example, examin-ing multiple sources of data about Thanksgiving reveals a

disjuncture between informants' recurring verbal reportsthat they "always" do the same thing each year, and obser-vation of them negotiating about what should be done andwho should do it as the ritual unfolds (see, e.g., the previousdiscussion of competition between mother and uncle). Dis-junctures between observational and verbal report data pro-vide an opening for etic interpretation building that goes be-yond mere sunmiaries of emic perspectives.

Even if comparisons between different sources of data in-dicate that generalizations accurately report behavioral reg-ularities, the ethnographer questions whether this is the onlyreason for their inclusion in verbal reports. In addition to theaccuracy of the words, are there otlier reasons why infor-mants keep using them? If unstructured interview infor-mants do not report everything that actually happens, whydo so many of them choose to mention tliese absolutes?Why are these absolutes important to them? What are themore abstract cultural values or meanings that are containedin their references? Why are certain observed behaviorsovergeneralized and not others?

Answers to such interpretive questions are developed notby speculation or projection but by looking at the full sets offield notes for these informants to see what other behaviorsare linked to the presence or absence of overgeneralizations.The assumption guiding ethnographic interpretation build-ing is that important cultural values are expressed throughmore than one behavior in a setting; constructs that explaincultural meaning in a setting resonate across behaviors thatinitially appear dissimilar (McQuarrie and Mick 1992; Wal-lendorf and Amould 1991).

In the Thanksgiving Day research, we note informants'tendency to use absolute descriptors when family stabilityhas been or is challenged, when there are cleavages to be re-paired. We further note numerous behaviors that threaten thepossibility of demonstrating stability of the domestic group,even for only one day. Changes in product availability, lifecycle positions, and household composition and resourcesinfuse instability into celebrations. The disjuncture betweenovergeneralizations in verbal reports and observations ofvariation in behavior is explained by a cultural ideal thatguides, but does not perfectly predict, behavioral enact-ments. We conclude that asserting that the same thing isdone each year proves ideals of stability, while negotiationof arrangements for the feast accommodates family mem-bers' evolving preferences, resource endowments, andchanging roles. Both are necessary in this consumption con-text to accommodate an ideal of familial stability.

Etic interpretation of overgeneralizations indicates that insome cases a behavioral ideal is held so strongly that it pre-vents informants from critically considering their reports inlight of their own behavior (Amould 1989). In identifyingsuch disjunctures, ethnographers attempt to account both forstated ideals and behaviors that differ from ideals. Etic in-terpretation articulates the potential conflicts averted byconsumer practices that initially appear to contradict over-generalizations, as well as the cultural functions served byideals. The overgeneralization revealed in claims of "al-ways" is a strong force guiding future expectations as wellas present action. As we demonstrate susbequently, etic in-terpretation of the behavioral associations, meanings, and


functions of ideals serve as potent initial premises in devel-oping marketing strategy.

Glosses. Glosses are useful in interpretation building bydirecting the ethnographer to account for disjunctures be-tween how people interpret their own behavior and the ac-tion observed by the ethnographer. Glosses may be codedfirst according to the explicit claims made by informants,claims that are easily recognized because they are often ex-pressed through conventionalized vocabulary. Disjuncturesbetween verbal glosses and actual behavior are most readilyascertained when the verbal reports accompany behaviorrecorded during participant observation.

For example, we observed numerous households consum-ing foods on Thanksgiving Day that they referred to as"homemade" or "made from scratch." Conventionalized vo-cabulary makes coding these verbal reports using the termHOMEMADE straightforward. But ethnographic interpreta-tion must go beyond marking recurrences in emic under-standing. Comparing verbal reports with observation fieldnotes about their accompanying actions reveals that the spe-cific behaviors people reference consist in assemblingbranded rather than homegrown products and that the prod-ucts and procedures differ by household. The task for ethno-graphic interpretation at this point is to account for variationin the actions performed as well as the functions served byemic glossing.

To determine why the gloss recurs, the ethnographer ex-amines sets of field notes containing glosses for evidence ofother behaviors that may express similar cultural meanings.For example, participant observation field notes contain nu-merous discussions about whose house should be the settingfor the feast this year as well as who should be included, andnumerous interview comments express negative sentimentsabout eating at a restaurant on "Thanksgiving Day. Thesecomments all assert the importance of a domestic and fa-milial locus and are therefore noted as being symbolicallysimilar to references to homemade food in the constructsthat explain their cultural importance. Moreover, matchingthese behaviors with more intensive labor in meal prepara-tion and more frequent removal of branded packaging thanduring everyday dinners point to interpretive constructs thatarticulate underlying cultural values, namely "decommodifi-cation"5 and family continuity (Wallendorf and Amould1991). These behaviors and glosses are culturally interpret-ed as proving (semiotically) the productive potential of thedomestic group.

In another type of gloss, people assert that their behaviorsor concerns are similar to those of a larger set of people.Though sometimes accurate, such claims may also be con-

'The term decommodification is adapted from Kopytoffs (1986) notionof singulaHzation and the work of Appadurai. It refers to a special kind oftransvaluation, in which objects are placed beyond the culturally demarcat-ed zone of commoditization. This type of transvaluation can take differentforms in different societies, but it is typical that objects which representaesthetic elaboration and objects that serve as sacra are, in many societies,not permitted to occupy the commodity state (either temporally, socially ordefmitionally) for very long (Appadurai 1986, p. 23). An altemative mightbe Miller's (1987) concept of appropriation, referring to creative recontex-tualization or restructuration and reinvention of market resources accordingto domestic situations and expectations about uniqueness as part of moregeneral strategies of social positioning.

sidered glosses when the informant makes them without ac-cess to a sufficient body of evidence or range of experienceto justify them. Then the informant is using such glosses toexpress an emic meaning rather than the behavior's statisti-cal frequency.

In addition to checking whether claims are statisticallyaccurate through observation and surveys, market-orientedethnographers question why informants make such claims.Why is it important to them that their consumption behaviorbe regarded as typical? Particularly in a culture known for itsfocus on individualism, why are such claims of uniformitymade?

For example, in their verbal reports, many informantsmaintain or imply that they do what "everyone" does forThanksgiving. However, the accuracy of these assertions isnot readily available to participants, because they typicallycelebrate within a limited network of kin and close friends.They disattend any evidence of differences when they saytheirs is "a classic Thanksgiving" with "typical Iowa food"and "the traditional turkey and dressing" served in a "typi-cal Thanksgiving scene." Field note references to idiosyn-cratic menu items such as "smoked turkey done on the grill,""sweet potatoes cooked in pineapple juice," and "pistachiofluff' are included at feasts described by these same infor-mants as representing "traditional" Thanksgiving fare.

In developing interpretations, ethnographers analyze in-formants' comparative statements about themselves, askingwhat sample of experiences these inferences are based onand what function they serve. Ethnographic interpretationgoes beyond accepting emic assertions of typicality at facevalue. An ethnographer's etic constructs interpret mentionsof "traditional" as a recurring belief in commonality thatsuppresses perceptions of temporal, regional, racial, histori-cal, and class differences that otherwise permeate U.S. con-sumption behavior. The recurring belief that wbat one doesis "typical" is a form of universalism central to what we in-terpret as the reparative, homogenizing function of Thanks-giving as national holiday. Both glossed understanding andbehavior as observed are elements of a market's understand-ing of the products used; an account of both must be includ-ed in ethnographic interpretation.

Informants'claims of idiosyncracy. When compared withbehavioral observation data, informants' verbal claims of id-iosyncracy may also provide a useful basis for building in-terpretation. For example. Thanksgiving Day participantsdescribe some dishes as if they are unique to their family.These mentions, typified by the entries in Table 3, werenoted for their recurrence and coded as SPECIAL. As wecompared across multiple data sets, we saw the inclusion ofsomething coded as SPECIAL as typical of many Thanks-giving celebrations. Etically, the fact that so many familiesinclude something that they regard as idiosyncratic repre-sents a paradigmatic similarity. Although the particular dish-es vary somewhat, we interpret repeated claims of idiosyn-cracy as meaningful and culturally significant.

Ethnographic interpretation must then attempt to ascer-tain constructs that explain why the recurrent feature is soimportant that it is repeatedly mentioned by informants ininterviews and enacted in behavior. Data analysis proceedsto examine other behaviors that are linked to the recurrent


phenomenon. For example, mentions of special recipes anddishes are frequently associated with intergenerationaltransfers of knowledge. These in turn are associated with theprofound sadness informants express about the dissolutionof the family and/or the death or prospective death of oldermembers. The connection between all these behaviors isneeded to grasp the constructs that articulate the underlyingcultural meaning of dishes informants consider special. Byadding preferred brands of particular fruits to preferred fla-vors of gelatin or specific ingredients to preferred brands ofbread cubes, families represent the resultant dishes to them-selves as a decommodified symbol of intergenerational con-tinuity and enduring family tradition. In this way, claims ofidiosyncracy are interpreted as having an emic basis (e.g., infact, different dishes were reported as "special" to differentfamilies) while also illustrating a commonality across infor-mants that points the way to etic interpretations of cultural-ly salient features.

Model of Interpretive Process

In this section, we summarize the process of building in-terpretation in market-oriented ethnographies. We show howthis process links ethnographic data in a multilayered andmultivocal interpretation of behavioral constellations. Wediscuss in general terms the interpretive processes describedwith more specific referents in the preceding three sectionson interpreting disjunctures. Figure 1 provides a visualmodel to summarize (but not fully represent the complexi-ties of) the interpretive processes of coding, troping, andrepresenting market-oriented ethnographic data.

As shown in the upper part of Figure 1, market-orientedethnography gathers data about behaviors (B) that occur in acontext. Rather than studying the purchase or use of onlyone product, market-oriented ethnography studies the fullset of behaviors that naturally co-occur with purchase or useof that product (Bj,). Thus, market-oriented ethnographystudies the actions associated with breakfast or getting readyto go to work or school in the morning rather than con-sumption of a particular brand of breakfast cereal. Emic un-derstanding of the "goes together" nature or coconstitutingmeanings of this set of behaviors is termed the quality space(Fernandez 1986); etic understanding of co-occurring be-haviors (whether or not their co-occurrence is recognized byparticipants) and their emergent meanings is called a behav-ioral constellation (Baudrillard 1968; Fernandez 1986; Mc-Cracken 1989; Solomon and Assael 1987).6

Market-oriented ethnography provides a contextualizedunderstanding of the meanings of product use that goes be-yond the individual brand or product attributes that man-agers may be accustomed to considering; it studies the

meaningful behavioral constellation in which product or ser-vice use is embedded.

It is in this respect that ethnography differs most clearlyfrom research in marketing on brand or product attributes orthe comparative attributes of competing brands, includingthe motivation research that gained prominence in the 1950sand is still in use in the private sector. Motivation researcharticulates the psychodynamics surrounding use of individ-ual products (e.g., Dichter 1964) by pairing them with dis-embedded psychoanalytic meanings (Durgee 1991).Ethnography differs from motivation research and succes-sive brand-attribute or competing-brand research in four re-spects: First, both motivation research and brand-attributeresearch study individual products or brands in isolation(e.g., Dichter [1964] interpreted use of deodorant separatelyfrom use of toothpaste). Ethnography, on the other hand,studies product use as part of culturally contextualized con-stellations of behaviors, such as grooming rituals (Rook1985) that include both toothpaste and deodorant. Second,ethnography differs from motivation research and brand-at-tribute research in the types of meanings studied. Motivationresearch attempts to isolate and freeze psychoanalytic prod-uct meanings; brand-attribute research attempts to isolateand freeze cognitive product meanings; and market-orientedethnography attempts to convey multiple, often conflictingperspectives and articulate dynamic cultural meanings ofproducts. Third, as is appropriate to their focus on psycho-dynamic and cognitive meanings, motivation research andbrand-attribute research both rely primarily on verbal reportdata; this contrasts with the primacy ethnography gives toobservation of behavior. Fourth, ethnography differs frommotivation research and brand-attribute research in its ca-pacity to cross-check interpretation of the behaviors relatedto one product by examining the behaviors related to otherproducts within the same constellation.

In market-oriented ethnography, data about the severalbehaviors in a behavioral constellation (Bjj, h = 1 ... m,where m = the number of behaviors the ethnographer iden-tifies as part of the constellation) are gathered from purchaseor consuming units in the sample (i = 1 ... n), using variousethnographic data collection methods (j = 1 ... x, where x =the number of ethnographic data collection methods em-ployed). Effective ethnographic interpretation begins duringdata collection by paying attention simultaneously to threeaspects of sampling: across purchase or consuming units(denoted by subscript i), across several behaviors of thoseunits or individuals that naturally co-occur (subscript h), andacross forms of data collection (subscript j). The data set Bjjyspans the several behaviors in a constellation, for numerouspurchasing or consuming units, collected using multipledata collection methods.

notion of behavioral constellations is both theoretically appealingand operationally challenging. A product can be used as part of several be-havioral constellations: Eating ready-to-eat cereal can be placed in the con-stellation of behaviors that composes breakfast eating as well as those thatconstitute after-school snacking. Using toothpaste can be placed in the con-stellation of behaviors that composes morning grooming and those thatconstitute preparing for romance. Further complexity is introduced by thefact that a geographic site or temporal event may contain several kinds ofconsumers enacting different behavioral constellations: Different kinds ofbaseball spectators enact differing behavioral constellations despite being

present at the same game (Holt 1993). Swap meet consumers at the samesite include those engaging in family recreation (which includes buyingsnacks), as well as those enacting their version of economic rationalitysearching for deals (and thus resisting buying snacks). A commercial white-water river rafting trip may include people searching for a predictable,guide-controlled ride resembling those at a water sports theme park as wellas people intent on athletic adventure and unexpected challenges. Becauseof the complexity of this concept, ethnography looks to systematic, empir-ical observation of individuals in context to determine the composition ofbehavioral constellations.


Interpretation building commences as an ethnographer or-ganizes the data using brief codes to represent behaviors thatrecur across the sample of individuals. Such data organiza-tion may be accomplished by hand as it has for decades inits originating disciplines using margin notes (Miles andHuberman 1988) or may be aided by any of the many textcoding and retrieval computer programs (e.g., AskSam,Ethnograph, GoFer, Notebook, Tally, WordCruncher, Zy-Index) available in the social science and literary analysisfields (Felding and Lee 1991). Although either approach canproduce sound, credible, and insightful interpretations, com-puter programs are of enormous benefit in overcominghuman information processing deficits by drawing system-atically and rapidly from the full data corpus to reach inter-pretive conclusions.^

The initial process of organizing data about particular be-haviors (Bjjy) using codes (e.g., Codejj) is shown in Figure 1with solid arrows. Codings are developed initially by notingfield note recurrences of a particular behavior across thesample (i = 1 ... n) for one data collection method. At thisstage the ethnographer examines the entire sample from onedata collection method with respect to a particular behaviorin the constellation. Examples from the previous discussioninclude recurrent verbal reports coded as ALWAYS, HOME-MADE, and SPECIAL.

Initially, each behavior is coded with as many emergentcodes as are needed to record the ethnographer's developingunderstanding. Rather than being developed a priori, as isconventional in content analysis (Kassarjian 1977; Wind,Rao, and Green 1991), ethnographic codes are developed asthe ethnographer reads the data and notices recurring wordusage, phrases, complex behavioral sequences, or meanings.Codes are nothing more than a way to mark recurrences, butthey typically go beyond merely marking recurrences inemic language (e.g., SPECIAL is a code used for more thanjust emic mentions of the word "special"; it includes allfoods, service objects, or behaviors regarded as noteworthyby the informant). Some behaviors may be marked withonly one code, whereas others may be pertinent to severalcodes (e.g., the emic statement "Every year, we have mom'sfamous, homemade lemon pie" might be marked with threecodes: ALWAYS, HOMEMADE, and SPECIAL). Somecodes may apply only to one behavior, and others may befound to mark similarity in several different behaviors (e.g.,HOMEMADE codes descriptions of the preparation of a va-riety of foods including rolls, pies, potatoes, and cranberrysauce, as well as some table decorations).

But ethnographic interpretation must go beyond codesthat mark recurrences in emic understandings and practice.Interpretation of the meanings of codes is elaborated by ex-ploring additional connections to the data. The market-ori-ented ethnographer next iterates between the codes andother data; this effort attempts to identify convergences anddisjunctures by comparing data about a behavior gatheredusing different data collection methods (moving across B|,j = 1 ... x). Codes for data about a behavior from one datacollection method are compared with data about the samebehavior collected using a different data collection method

7We appreciate the helpful input of a JMR reviewer who pointed out thisbenefit during the review process.

(e.g., the code for verbal reports describing "homemadepies" is compared with data from photos and observations ofpeople preparing and serving pies). Comparisons indicatesome convergences as well as some disjunctures. Ratherthan just noting disjunctures, or (worse) assimilating them toone convergent interpretation, market-oriented ethnogra-phers question why convergences and disjunctures betweenmethods occur. The interpretive meanings of codes are"thickened" into constructs in light of these additional data.

Figure 1 shows further refinement of the codes into con-structs when comparisons are made across the behaviorsthemselves (B|j). Because ethnography presumes that im-portant cultural values or meanings are expressed by sever-al behaviors in a quality space, data interpretation at thispoint attempts to identify these resonances. Data about otherbehaviors that initially appeared only loosely connected arecompared with the codes, using iterative tacking betweendata sets.

These comparisons are used to construct tropes, whichare meaningful symbolic links between various behaviors orverbal statements (Fernandez 1986, 1991). Comparing dif-ferent behaviors to identify symbolic links is known as trop-ing. Rather than merely assessing whether codes can be di-rectly applied to additional behaviors, the ethnographer con-structs tropic relations by assessing the symbolic applicabil-ity across behaviors of the meanings represented by codes.Systematic relations of contrast and association among be-haviors are considered (Durham and Fernandez 1991, p.192), stemming from their shared cultural foundation. Thesemay include the following relations:

1. Paradigmatic—A set of behavioral variants of a single type,such as the various kinds of HOMEMADE dishes, confirmingthe value placed on domesticity;

2. Syntagmatic—Part of a culturally prescribed temporal or nar-rative sequence, such as interrupting the viewing of televisedfootball games to eat the meal or the telling of ThanksgivingDay stories about bad times overcome, to confirm both theprimacy and resiliency of familial life;

3. Metaphoric—A figurative relationship of similarity, such assimilarity between "stuffed" Thanksgiving turkeys, pies,houses, and people and the cultural value of abundance beingembodied; or

4. Metonymic—A relation in which part is taken as an emblem-atic representation of the whole domain, such as use of astuffed turkey to imply the entirety of the Thanksgivingholiday.

The process of constructing tropes is shown in Figure 1 asupward dashed lines going from initial codes to data aboutother behaviors and then downward dashed lines returningto refine the codes into constructs. The process of troping ormaking these comparisons refines the codings so that theyare now referred to as constructs that form identifiablethemes (represented as solid, horizontal lines in Figure 1) inthe interpretation. Interpretive layers or themes that are con-nected to numerous behaviors in the quality space representimportant cultural constructs and, therefore, are accordedbroader etic significance in the final ethnographic represen-tation of the constellation.

Through the interpretive activities of coding and troping,layers of meaning comprising an interpretation are devel-oped and checked across informants and across multiple


sources of data about the numerous behaviors that form aconstellation (across Bjjjj). Ethnographic interpretation artic-ulates constructs that are shared, resonant, or discrepantacross the constellation. Simultaneously operating layers ofmeaning identified in the interpretation are complementaryrather than mutually exclusive. It is this multilayered andcultural nature of ethnographic interpretation that is refer-enced in Geertz's (1973) often-cited term thick description.,it is the desired outcome that ethnographers refer to as arichly textured interpretation.

FROM ETHNOGRAPHIC REPRESENTATION TOMARKETING STRATEGY

Once a multilayered interpretation has been developed ina market-oriented ethnography, it can be put to a variety ofuses. The goal dependence of written representations of anethnographic interpretation is depicted in Figure 1 throughthe use of a dotted frame around each of two representa-tions; this emphasizes the possibility of conveying an ethno-graphic interpretation through several different kinds of rep-resentations, each of which can highlight different featuresor layers in the interpretation to accomplish different goals.Perhaps most central to academic pursuits, the interpretationof a behavioral constellation can be transformed into a writ-ten representation that extends theory involving consump-tion or market processes (see further discussion in Sherry1990b). Such representations are typically presented asethnographic articles in academic journals (e.g., Amould1989 on diffusion), books and monographs (e.g., McCrack-en 1989 on "homeyness").

However, as shown in Figure 1, extending theoreticalknowledge is not the only use for interpretations from mar-ket-oriented ethnographies. In addition to its long-standingtheory-building role in social science, ethnography has along history of strategic use in a variety of applied contexts,including agriculture, education, health care, and public pol-icy (e.g., Wulff and Fiske 1987). In this section, we point tothe ways market-oHented ethnographies can be used to de-velop strategic recommendations for marketing practition-ers. We focus on the use of ethnography in identifyingstrategic recommendations rather than its use in theorybuilding because of the absence of such a discussion in themarketing literature.

We discuss four types of managerially useful representa-tions that can be prepared from market-oriented ethnogra-phies, each reflecting the strategic goals of its particular au-dience. One basic managerial goal that can be accomplishedwith market-oriented ethnography is providing a thoroughunderstanding of the many unarticulated layers of consumermeaning embedded in behavioral constellations. Suchethnographic understanding gains strategic usefulness whenthe consumption experience hinges on significant taken-for-granted sociocultural understandings, and customer satisfac-tion is not based solely on cognitively-available product at-tributes. A representational style that articulates a thick de-scription of a behavioral constellation can belp marketingpractitioners understand sociocultural layers of meaning thatconsumers have difficulty articulating but nonetheless acton.

This was the case in the managerially sponsored researchon commercial white-water river rafting conducted byAmould and Price (1993). In addition to preparing the theo-ry-building representation cited, the ethnographers prepareda representation for river rafting outfitters that described thecomplex, unarticulated, cultural sources of customer satis-faction (Price and Amould 1991). In standard surveys cus-tomers reported concems with attributes such as safety,planned food service, pleasant company, and logistics. How-ever, market-oriented ethnography showed that customers'satisfaction is primarily affected by experiencing personalgrowth and renewal, harmony with nature, and liminal com-munitas (Tumer 1969) with others on the trip, including theguide. Using this interpretation, the ethnographers prepareda representation to provide strategic recommendations formanagers, suggesting specific modifications of employeetraining regarding these implicit benefits (Price and Amould1991). This use of market-oriented ethnography urges man-agers to accommodate and address a wider range of mean-ings than they have previously recognized.

A second related managerial goal involves assessing gapsin a specific product or service's ability to deliver resonantmeanings across behaviors in a constellation. This assess-ment of gaps follows the interpretive troping shown in Fig-ure 1, in which behaviors are linked to more than one layerof meaning. This managerial goal is accomplished by iden-tifying strategies that would link a product to additional lay-ers of meaning. If a specific product or service participatesonly partially in the cross-behavior resonances identified introping, suggestions for its modification or repositioning canbe offered.

This was the case in a corporate-sponsored research pro-ject that gathered unstructured interview and focus groupdata on use of a particular branded food product at Thanks-giving feasts, building on the previous theory-building rep-resentation (Wallendorf and Amould 1991). Although thereis a constellation of foods (e.g., turkey, dressing or stuffing,mashed potatoes and gravy, cranberry sauce, pumpkin pie)that are so metonymically connected to Thanksgiving thatthey are included in almost all feasts, it is possible to capi-talize on cross-behavior resonances to expand sales for anyof these food products during the holiday season (Wallen-dorf 1992). We illustrate this by pointing out how ethno-graphic research could guide the development of marketingstrategies to enhance sales of cranberry products atThanksgiving.

Such strategies are developed by examining layers ofmeaning that are connected (by arrows in Figure 1) to otherproducts in the constellation, but for which no tropic rela-tions with cranberry products are apparent. Identified linksbetween another product in the constellation, namely pies,and several layers of meaning provide clues to strategic op-tions for enhancing holiday sales of cranberry products.Typically several whole pies (pumpkin plus others) are pre-pared for the feast, even though one might be sufficient. Thisis similar to other behaviors in the constellation tbat involvepreparing much more food, sufficient for leftovers, than isneeded to satiate ("stuff") everyone. This resonant focus onquantity suggests that proving abundance is an importantpart of the ritual task, but does not fully explain why varied


types of pies are prepared. Feast preparers interviewed aboutttiis observation express pleasure at providing each person'sfavorite. Underlying this pleasure is the challenge, faced bymiddle-aged men or women who often prepare the feast fortheir families, of making the family "whole," even if just forone day. A behavior that enacts all these layers of meaninginvolves preparing several types of pies. Through consump-tion, abundance is proven; simultaneously, men and womenare confirmed in tiieir pivotal care-giving role, individualpreferences are met, and all are drawn together (making thefamily whole).

This interpretation concerning pies suggests a promotion-al strategy to enhance Thanksgiving sales of cranberry prod-ucts. Promotions could encourage service of more than onetype of cranberry sauce: soft-texture jellied sauce as well ascrunchy-texture whole berry sauce. As is also true of pies,preparing several types of cranberry sauce could be posi-tioned as a way to satisfy more individual preferences, de-liver a message about abundance, and enact the importanceof bringing diverse family members together.

Another strategy for cranberries is suggested by thebranded assemblages used to construct "homemade" and"special" pies. Promotional materials could feature foodpreparations in which cranberries respond to the desire fordecommodification that permeates the feast. For example,promotional materials could include recipes indicating how"homemade" sauce can be prepared from fresh wholeberries; they could indicate how the cooked whole cranber-ries can become "special" by adding ingredients, such as or-ange rind, nuts, or ginger. Use of fresh, whole cranberries ina special recipe for Thanksgiving allows for a symbolic,paradigmatic marking of the dish that does not encroach onconsumers' use of other canned cranberry products for ordi-nary dinners. "Homemade" pies and cranberry sauce both fit(paradigmatically and metaphorically) with consumers' at-tempts at decommodification to demonstrate the productivepotential of the household symbolically.

These types of strategic possibilities are developed bynoting layers of convergent meaning and gaps in the focalproduct's delivery of meanings connected to other productsin the constellation. Unlike brand- or product-focused mar-ket research that formulates strategy by developing contrastswith substitutes or competitors, market-oriented ethnogra-phy turns to complementary products in the constellation,and tropic relations of similarity and contrast in formulatingstrategic suggestions. In this way, market-oriented ethnogra-phy can be used to suggest strategies to foster product usesthat resonate with existing meanings of other products in thebehavioral constellation.

A third strategic use of ethnographic interpretation is toprovide managers with vivid exemplars of the layers ofmeaning that organize behavioral constellations. Althoughthick description is a necessary goal for ethnographic inter-pretation, when carried through in representation, it mayconvey an impersonal, univocal nonparticipant stance thatdoes not always meet managers' strategic goals. Managers,especially in new product development or creative advertis-ing work, sometimes need vivid exemplars of meaningfulconstructs rather than textured verbal explanations. Thismanagerial need is best met by a representational style

known as thick transcription (Clifford 1990), originated byethnographic photographers (e.g., Stoller 1992).

In thick transcription, the market for a product or servicespeaks of and for itself. The ethnographer's role is one ofcompilation and juxtaposition of texts based on his or her in-terpretation of relevant layers of meaning. The presentationis polyphonic; the voices of native participants comethrough clearly as they enact and recount their consumptionexperiences. After developing an interpretation, the ethnog-rapher selects stellar exemplars of identified layers of mean-ing and presents them using such devices as videos of cus-tomer stories, transcriptions from observations and unstruc-tured interviews, or excerpts (often on video) from focusgroups. In a sense, this is often what managers hope to getfrom watching focus group proceedings firsthand; instead,what they experience is a lengthy string of unedited and un-interpreted consumer commentary, only selected portions ofwhich effectively capture in a vivid way the underlyingmolar constructs that organize the behavioral constellation.Nonetheless, when properly interpreted and then edited,such material can be quite useful in meeting managerialgoals. Young & Rubicam has produced thick transcriptionsto generate insights their creative staff used in developingadvertising executions (Alsop 1986).

Colgate-Palmolive has used thick transcription to im-prove targeting to the diverse Mexican market for cleaningproducts with repositioned as well as new product offerings.Managers claim to have learned much about the role ofcleaning chores in daily life, lifestyles and women's self-image, product substitutes, low levels of brand loyalty, andthe subjective meaning of time spent cleaning. To accom-plish this, Colgate-Palmolive researchers made extensivevideo recordings of customers segmented by class, age,place of residence, and cleaning task. Women informantswere videotaped in the act of cleaning while explaining whatthey were doing (e.g., applying talcum powder to greasestains) and the particular problems they encounter (rangingfrom a scarcity of water to careless housemaids to uncer-tainty about their social security). After many hours of tapewere recorded and interpreted, the researchers edited theminto five series of poignant, telling vignettes for presentationto and discussion by marketing and advertising managers(Evans and Berman 1992, Video 3, No. 21).

A fourth managerial goal involves sustaining an under-standing of a market over time. Conventionally, new marketresearch is launched when a need for new informationemerges. However, the comprehensive, muUilayered inter-pretations of market-oriented ethnography offer manage-ment a generative altemative. Particularly when the ethnog-rapher becomes an intimate insider in a consumption con-text or subculture, a return to the interpretive layers of amarket-oriented ethnography can provide answers to ongo-ing strategic questions that develop after the ethnography isconducted. The researcher can use his or her ethnographicunderstanding to translate likely concerns between theworld of customers/informants and that of managers andpredict likely responses with little or no new data collectionrequired.

The mode of representation that is appropriate here isthick inscription (Clifford 1990; e.g., Stoller and Olkes


1987), which privileges the experiences of the researcherwho becomes immersed in an acquisition or consumptionuse context. This representational mode, more than others,acknowledges the ethnographer's presence in and experi-ence of the consumption context. However, the privilege af-forded the ethnographer is not exclusive, because the re-searcher's involvement includes extensive interaction withother participants. This nonexclusive privilege differentiatesmarket-oriented ethnographies using thick inscription fromresearcher introspection (Wallendorf and Brucks 1993). Theethnographer is a full, if not naive, participant, attending toobservations of other participants' actions as well as to hisor her own embodied, subjective experience.

The central strategic value of thick inscription is the abil-ity to check potential strategies against the researcher's em-bodied understanding of emergent meanings at any point intime. In this respect, the relative costs of ethnographic re-search are reduced because of the longevity of its useful-ness. This use of market-oriented ethnography becomes par-ticularly appropriate when the needs of micro market seg-ments must be identified, such as the equipment concems ofthose who engage in high-risk commercial recreations(Celsi, Rose, and Leigh 1993) or when the market mixes forlifestyle or immigrant market segments must be refined(Pefialoza 1994).

Illustrative of the use of market-oriented ethnography toaddress evolving strategic concerns is Schouten andMcAlexander's enduring relationship with the Harley-Davidson Company based on their ongoing ethnography ofHarley-Davidson motorcycle owners. Although the re-searchers have prepared theoretical representations of thiswork for academic audiences (Schouten and McAlexander1992a, 1993a), they have also prepared representations forstrategic use within the firm (Schouten and McAlexander1992b, 1993b).

The four uses presented here are not the only managerialuses of market-oriented ethnography. However, they point tothe close connection between ethnographic interpretationand managers' strategy formulation needs. Marketing strate-gies are suggested by resonances and disjunctures identifiedin the multilayered ethnographic interpretation of behavioralconstellations associated with product purchase and use.

CONCLUSION

We describe the process of building interpretations inmarket-oriented ethnographic research and demonstratetheir value in developing marketing strategies. We discussthe strengths and limitations of two ethnographic datasources (observation of behavior in context and verbal re-ports in interviews) in building interpretations of behaviorand meaning in a cultural context. In general, ethnographygives primacy to observation of behavior in context to pro-vide a perspective in action and relies on verbal reports in in-terviews for a perspective of action. The two data collectionmethods play different roles in the process of building an in-terpretation. Comparisons between the two sources of dataprovide systematic access to recurrent disjunctures betweenperspectives of and perspectives in action.

Through coding and troping, interpretation builds to anunderstanding that incorporates both divergent and conver-

gent elements in a thick, multilayered representation of mar-ket behavior. Interpretation is a multiphase process, drivenby data and refined by cross-behavior resonances and dis-junctures that point to the sociocultural basis of structuredpattems of action. Disjunctures between observations of ac-tual consumption behavior in context and verbal reportscontaining overgeneralizations, glosses, and claims of id-iosyncracy are shown to have multiple strategic implica-tions. Stated most generally, accounting for disjunctures im-proves managerial prospects for providing markets with theproducts and services needed to narrow undesired gaps be-tween behavior and goals, values, norms, and beliefs.

Finally, we point out that ethnography has a long historyof practical applications. We illustrate how market-orientedethnographies have produced actionable marketing strate-gies with respect to four types of managerial goals. Thesemanagerial goals are served by different representationalmodes, including thick description, thick transcription, andthick inscription.

Without being overly prescriptive or restrictive, we cansuggest that these modes of ethnographic representationmay be especially useful to the formulation of marketingstrategy in the following ways. Thick description has beenuseful to marketing managers wishing to know more aboutthe meaning of particular consumption constellations to par-ticular market segments; it can also help to explain the in-fluence of marketing uncontroUables on the behavior ofmarket segments; and it can be useful in improving productpositioning and promotional strategies based on cross-prod-uct complementarities. Thick transcription has been particu-larly useful to marketing managers wishing to generateevocative representations of consumption behaviors to de-velop line extensions, product reformulations, and more ef-fective promotional strategies. Thick inscription has beenuseful to marketers seeking to develop strong niche market-ing strategies, especially with segments characterized bystrong insider-outsider boundary mechanisms, such as em-bedded subcultures, collectors and other enthusiasts, andyouth cultures.

Our explication of interpretation building in market-ori-ented ethnography can be read in two ways. First, it suggestsguidelines for those attempting to build interpretation inmarket-oriented ethnographies. Although this article is not acomprehensive tutorial, we strive to make this interpreta-tion-building process more accessible and transparent (San-jek 1990) for a broad audience. However, our explication ofinterpretation building can also be read as suggesting evalu-ative criteria for those reading market-oriented ethnogra-phies. Our account aims to make it easier for researchersopen to methodological pluralism (Lutz 1989) to assess therigor and conduct of individual pieces of ethnographic work.We do not urge the universal application of context-inde-pendent rules to the conduct of ethnography, but neither dowe support an "anything goes" approach. Implicitly, wemaintain that ethnographies may vary in quality or rigor. Insummary, our discussion of market-oriented ethnographyaspires to elevate sophistication in the field of marketing be-yond the equally unacceptable positions of unwavering ad-miration for and total dismissal of research employing a


methodology with a long tradition of fruitful theory-build-ing and application.

REFERENCESAdler, Patricia A. and Peter Adler (1987), Membership Roles in the

Field. Newbury Park, CA: Sage Qualitative Research Methods,Series 6.

Agar, Michael H. (1980), The Professional Stranger: An InformalIntroduction to Ethnography. New York: Academic Press.

(1986), Independents Declared: The Dilemmas of Inde-pendent Trucking. Washington, E)C: Smithsonian InstitutionPress.

Alsop, Ronald (1986), " 'People Watchers' Seek Clues to Con-sumers' True Behavior," Wall Street Joumai (September 4), 29.

Anderson, Eugene W. and Steven M. Shugan (1991), "Reposition-ing for Changing Preferences: The Case of Beef Versus Poultry,"Joumai of Consumer Research, 18 (September), 219-32.

Appadurai, Arjun (1986), "Introduction: Commodities and the Pol-itics of Value," in The Social Life of Things, A. Appadurai, ed.Cambridge: Cambridge University Press, 3-63.

Amould, Eric J. (1989), "Towards a Broadened Theory of Prefer-ence Formation and the Diffusion of Innovations: Cases fromZinder Province, Niger Republic," Joumai of Consumer Re-search, 16 (September), 239-67.

and Sidi Mohammed Iddal (1992), Niger: Export Mar-keting ofNigerien Onions. Washington, DC: U.S.A.I.D.; Post-Harvest Institute for Perishables, University of Idaho; and Abt &Associates.

- and Linda L. Price (1993), "River Magic: Hedonic Con-sumption and the Extended Service Encounter," Joumai of Con-sumer Research, 20 (June), 24-45.

Baudrillard, Jean (1968), Le Systime des Objets. Paris: Gallimard.Bearden, William O., Richard G. Netemeyer, and Mary F. Mobley

(1993), Handbook of Marketing Scales. Newbury Park, CA:Sage Publications.

Becker, Howard S. and Blanche Geer (1960), "Participant Obser-vation: The Analysis of Qualitative Field Data," in Human Or-ganization Research, Richard N. Adams, and J. Preiss, eds.Homewood, IL: Dorsey Press, 267-89.

Belk, Russell W, John F. Sherry, Jr., and Melanie Wallendorf(1988), "A Naturalistic Inquiry into Buyer and Seller Behaviorat a Swap Meet," Joumai of Consumer Research, 14 (March),449-70.

, Melanie Wallendorf, and John F Sherry, Jr. (1989),"The Sacred and the Profane in Consumer Behavior: Theodicyon the Odyssey," Joumai of Consumer Research, 16 (June),1-38.

-, and Morris B. Holbrook (1991),"Collecting in Consumer Culture," in Highways and Buyways:Naturalistic Research from the Consumer Behavior Odyssey.Provo, UT: Association for Consumer Research, 178-215.

Bennett, John W. and Gustav Thaiss (1970), "Survey Research inAnthropological Field Work," in A Handbook of Method in Cul-tural Anthropology, RaouU Naroll and Ronald Cohen, eds. Gar-den City, NY: The Natural History Press, 213-337.

Biggart, Nicole Woolsy (1989), Charismatic Capitalism: DirectSelling Organizations in America. Chicago: University of Chica-go Press.

Bitner, Mary Jo, Bernard H. Booms, and Mary Stanfield Tetrault(1990), "The Service Encounter: Diagnosing Favorable and Un-favorable Incidents," Joumai of Marketing, 54 (January), 71-84.

Bourdieu, Pierre (1984), Distinction: A Social Critique of theJudgement of Taste, trans, by Richard Nice. Cambridge, MA:Harvard University Press.

Boyd, Harper W and Sidney J. Levy (1963), "New Dimension inConsumer Analysis," Harvard Business Review (November/De-

cember), 129-40.Bristor, Julia M. and Eileen Fischer (1993), "Feminist Thought:

Implications for Consumer Research," Joumai of Consumer Re-search, 19 (March), 518-36.

Campbell, Donald T. and D.W. Fiske (1959), "Convergent and Dis-criminant Validation by the Multitrait-Multimethod Matrix,"Psychological Bulletin, 30, 81-105.

Celsi, Richard, Randall L. Rose, and Thomas W. Leigh (1993), "AnExploration of High-Risk Leisure Consumption Through Sky-diving," Joumai of Consumer Research, 20 (June), 1-23.

de Certeau, Michel (1984), The Practice of Everyday Life, trans, bySteven Rendall. Berkeley, CA: University of California Press.

Chalfen, Richard (1987), Snapshot Versions of Life. BowlingGreen, OH: Bowling Green State University Popular Press.

Clifford, James (1990), "Notes on (Field)notes," in Fieldnotes,Roger Sanjek, ed. Ithaca, NY: Cornell University Press, 47-70.

and George E. Marcus (eds.) (1986), Writing Culture:The Poetics and Politics of Ethnography. Berkeley, CA: Univer-sity of California Press.

Collier, John, Jr. and Malcolm Collier (1986), Visual Anthropology(revised and expanded). Albuquerque, NM: University of NewMexico Press.

Crapanzano, Vincent (1992), Hermes' Dilemma and Hamlet's De-sire: On the Epistemology of Interpretation. Cambridge, MA:Harvard University Press.

Dean, John P and William Foote Whyte (1958), "How Do YouKnow if the Informant is Telling the Truth?" Human Organiza-tion 17 (2), 34-38.

Denzin, Norman K. (1989), Interpretive Interactionism, Vol. 16,Applied Social Research Method Series. Newbury Park, CA:Sage Publications.

Dichter, Ernest (1964), The Handbook of Consumer Motivations.New York: McGraw-Hill.

Dougherty, Deborah Jane (1988), "Understanding New Marketsfor New Products," Strategic Management Joumai, 11, 59-78.

Durgee, Jeffrey (1991), "Interpreting Dichter's Interpretations: AnAnalysis of Consumption Symbolism in The Handbook of Con-sumer Motivations" in Marketing and Semiotics, Hanne HartvigLarsen, David Glen Mick, and Christian Alsted, eds. Copen-hagen: Handelshojskolens Forlag, 52-74.

Durham, Deborah and James W. Fernandez (1991), "Tropical Do-minions: The Figurative Struggle over Domains of Belongingand Apartness in Africa," Beyond Metaphor: The Theory ofTropes in Anthropology, James W. Fernandez, ed. Stanford, CA:Stanford University Press, 190-212.

Evans, Joel R. and Barry Berman (1992), Marketing, 5th ed. NewYork: Macmillan.

Felding, Nigel and Raymond M. Lee (1991), Using Computers inQualitative Research, Qualitative Research Methods Series.Newbury Park, CA: Sage Publications.

Fernandez, James W. (1986), Persuasions and Performances: ThePlay of Tropes in Culture. Bloomington, IN: University ofIndiana.

(ed.) (1991), Beyond Metaphor: The Theory of Tropes inAnthropology. Stanford, CA: Stanford University Press.

Fetterman, David M. (1989), Ethnography: Step by Step, Vol. 17,Applied Research Methods Series. Newbury Park, CA: SagePublications.

Foucault, Michel (1972), The Archeology of Knowledge. NewYork: Pantheon Books.

Garfmkel, Harold (1967), Studies in Ethnomethodology. Engle-wood Cliffs, NJ: Prentice-Hall, Inc.

Geertz, Clifford (1973), The Interpretation of Cultures. New York:Basic Books.

Giddens, Anthony (1971), Capitalism and Modem Social Theory:An Analysis of the Writings of Marx, Durkheim, and Max Weber.Cambridge: Cambridge University Press.


Goffman, Erving (1959), The Presentation ofSeifin Everyday Life.Garden City, NY: Doubleday.

Gould, Leroy C , Andrew L. Walker, Lansing E. Crane, andCharles W. Lidz (1974), Connections: Notes from the HeroinWorld. New Haven, CT: Yale University Press.

Granovetter, Mark S. (1973), "The Strength of Weak Ties," Amer-ican Journal of Sociology 78 (May), 1360-80.

Harrison, Gail, William L. Rathje, and W. Wilson Hughes (1974),"Socioeconomic Correlates of Food Consumption and WasteBehavior: The Garbage Project," Paper presented at the Ameri-can Public Health Association Meetings.

Heisley, Deborah and Sidney J. Levy (1991), "Autodriving: A Pho-toelicitation Technique," Journal of Consumer Research, 18(December), 257-73.

, Mary Ann McGrath, and John J. Sherry, Jr. (1991)," 'To Everything There is A Season': A Photoessay of a Farm-ers' Market," in Highways and Buyways: Naturalistic Researchfrom the Consumer Behavior Odyssey. Provo, UT: Associationfor Consumer Research, 141-66.

Hill, Ronald Paul (1991), "Homeless Women, Special Possessions,and the Meaning of 'Home': An Ethnographic Case Study,"Journal of Consumer Research., 18 (December), 298-310.

Hirschman, Elizabeth C. and Morris B. Holbrook (1982), "Hedo-nic Consumption: Emerging Concepts, Methods, and Proposi-tions," Journal of Marketing, 46 (Summer), 92-101.

and (1992), Postmodern Consumer Research.Newbury Park, CA: Sage Publications.

Hochschild, Arlie (1983), The Managed Heart. Berkeley, CA: Uni-versity of California Press.

Holt, Douglas B. (1993), "Forms of Consuming," doctoral disser-tation. Northwestern University.

Honigman, John J. (1970), "Sampling in Ethnographic Fieldwork,"in A Handbook of Method in Cultural Anthropology, RaoullNaroll and Ronald Cohen, eds. Garden City, NY: The NaturalHistory Press, 266-81.

Johnson, Alan and Oma R. Johnson (1990), "Quality Into Quanti-ty: On the Measurement Potential of Ethnographic Fieldnotes,"in Fieldnotes, Roger Sanjek, ed. Ithaca, NY: Cornell UniversityPress, 161-86.

Jorgensen, Danny L. (1989), Participant Observation: A Method-ology for Human Studies, Vol. 15, Applied Research MethodsSeries. Newbury Park, CA: Sage Publications.

Joy, Annamma (1991), "Beyond the Odyssey: Interpretations ofEthnographic Writing in Consumer Behavior," in Highways andBuyways: Naturalistic Research from the Consumer BehaviorOdyssey. Provo, UT: Association for Consumer Research,216-33.

Kassarjian, Harold H. (1977), "Content Analysis in Consumer Re-search," Journal of Consumer Research, 4 (June), 8-18.

Kohli, Ajay and Bernard Jaworski (1990), "Market Orientation:The Construct, Research Propositions, and Managerial Implica-tions," Journal of Marketing, 54 (April), 1-18.

Kopytoff, Igor (1986), "The Cultural Biography of Things: Com-moditization as Process," in The Social Life of Things, A. Ap-padurai, ed. Cambridge: Cambridge University Press, 64-94.

Kunda, Gideon (1992), Engineering Cultures: Control and Com-mitment in a High-Tech Corporation. Philadelphia: Temple Uni-versity Press.

Langer, E. (1983), The Psychology of Control. Beverly Hills, CA:Sage Publications.

Lederman, Rena (1990), "Pretexts for Ethnography: On ReadingFieldnotes," in Fieldnotes, Roger Sanjek, ed. Ithaca, NY: CornellUniversity Press, 71-91.

Levy, Sidney (1981), "Interpreting Consumer Mythology: A Struc-tural Approach to Consumer Behavior," Journal of Marketing,45 (Summer), 49-^1.

Lincoln, Yvonna S. and Egon G. Guba (1984), Naturalistic In-

quiry. Newbury Park, CA: Sage Publications.Lutz, Richard (1989), "Positivism, Naturalism, and Pluralism in

Consumer Research: Paradigm in Paradise," Advances in Con-sumer Research, Vol. 16, Thomas K. SruU, ed. Provo, UT: Asso-ciation for Consumer Research, 1-8.

McCall, G.J. and J.L. Simmons (eds.) (1969), Issues in ParticipantObservation. Reading, MA: Addison-Wesley.

McCracken, Grant (1988), The Long Interview. Newbury Park,CA: Sage Publications.

(1989), " 'Homeyness': A Cultural Account of OneConstellation of Consumer Goods and Meanings," in Interpre-tive Consumer Research, Elizabeth C. Hirschman, ed. Provo,UT: Association for Consumer Research, 168-85.

McGrath, Mary Ann (1989), "An Ethnography of a Gift Store:Wrappings, Trappings and Rapture," Journal of Retailing, 65(4), 421-24.

McQuarrie, Edward F. and David Glen Mick (1992), "On Reso-nance: A Critical Pluralistic Inquiry in Advertising Rhetoric,"Journal of Consumer Research, 19 (September), 180-97.

Merleau-Ponty, Maurice (1962), The Phenomenology of Percep-tion. London: Routledge & Kegan Paul.

Miles, Matthew B. and A. Michael Huberman (1984), QualitativeData Analysis: A Sourcebook of New Methods. Newbury Park,CA: Sage Publications.

Miller, Daniel (1987), Material Culture and Mass Consumption.Oxford: Basil Blackwell.

Mintzberg, Henry (1973), The Nature of Managerial Work. NewYork: Harper and Row.

Moerman, Michael (1988), Talking Culture: Ethnography andConversation Analysis. Philadelphia: University of Pennsylva-nia.

Ottenberg, Simon (1990), "Thirty Years of Fieldnotes: ChangingRelationships to the Text," in Fieldnotes, Roger Sanjek, ed. Itha-ca, NY: Cornell University Press, 139-60.

Pefialoza, Lisa (1994), "Atravasando Fronteras/Border Crossings:Mexican Immigrant Consumer Acculturation," Journal of Con-sumer Research, 21 (June), 32-54.

Price, Linda L. and Eric J. Amould (1991), A Report on DvorakRiver Rafting for Summer 1990: Customer Satisfaction and In-tentions. Boulder, CO: School of Business Administration, Uni-versity of Colorado.

Prus, Robert (1989), Making Sales. Newbury Park, CA: Sage Pub-lications.

Punch, Maurice (1986), The Politics and Ethics of Fieldwork,Qualitative Research Methods Series No. 3. Newbury Park, CA:Sage Publications.

Rabinow, Paul (1986), "Representations are Social Facts: Moder-nity and Post-Modernity in Anthropology," in Writing Culture:The Poetics and Politics of Ethnography, James Clifford andGeorge E. Marcus, eds. Berkeley, CA: University of CaliforniaPress, 234-61.

Richards, Audrey I. (1939), "The Development of Field WorkMethods in Social Anthropology," The Study of Society, F.C.Bartlett et al., eds. London: Routledge & Kegan Paul, 272-316.

Rook, Dennis (1985), "The Ritual Dimension of Consumer Be-havior," Journal of Consumer Research, 12 (December),251-64.

Sanjek, Roger (1990), "On Ethnographic Validity," in Fieldnotes,Roger Sanjek, ed. Ithaca, NY: Cornell University Press,385-418.

Schiffer, Michael B., Theodore E. Downing, and Michael Mc-Carthy (1981), "Waste Not, Want Not: An EthnoarchaeologicalStudy of Reuse in Tiicson, Arizona," in Modem Material Cul-ture: The Archaeology of Us, Michael B. Schiffer, ed. New York:Academic Press, 67-86.

Schouten, John W. and James H. McAlexander (1992a), "MarketImpact of a Consumption Subculture: The Harley-Davidson


Impact of a Consumption Subculture: The Harley-DavidsonMystique," in European Advances in Consumer Research, GaryBamossey and Fred van Raaij, eds. Provo, UT: Association forConsumer Research, 389-93.

and (1992b), "The Harley-Davidson Con-sumption Subculture," presentation made to Marketing Opera-tions Group, Harley-Davidson Co., Milwaukee, August.

and (1993a), "Subcultures of Consumption:An Ethnography of the New Bikers," working paper. Universityof Portland College of Business.

and (1993b), "Harley-Davidson ResearchSummit," presentation made to personnel from Marketing Oper-ations Group of Harley-Davidson Co. and from contract re-search firm, Chicago, August.

Sherry, John R, Jr. (1990a), "A Sociocultural Analysis of a Mid-western American Flea Market," Journal of Consumer Research,17 (June), 13-30.

f 1990b), "Post-modem Alternatives: The InterpretiveTUm in Consumer Research," in Handbook of Consumer Behav-ior, Thomas Robertson and Harold H. Kassarjian, eds. Engle-wood Cliffs, NJ: Prentice-Hall, Inc., 548-91.

Snow, David A. and Leon Anderson (1987), "Identity Work Amongthe Homeless: The Verbal Construction and Avowal of PersonalIdentities," American Journal of Sociology 92 (May), 1336-71.

, Louis Zurcher, and Gideon Sjoberg (1982), "Interview-ing by Comment: An Adjunct to the Direct Question," Qualita-tive Sociology, 5, 285-311.

Solomon, Michael R. and Henry Assael (1987), "The Forest or theTrees? A Gestalt Approach to Symbolic Consumption," in Mar-keting and Semiotics: New Directions in the Study of Signs forSale, Jean Umiker-Sebeok, ed. Berlin: Mouton de Gruyer,189-218.

Sontag, Susan (1977), On Photography. New York: Farrar, Strauss,Giroux.

Spradley, James P. and Brenda A. Mann (1975), The Cocktail Wait-ress: Women's Work in a Man's World. New York: Alfred A.Knopf.

Stem, Barbara (1993), "Feminist Literary Criticism and the De-construction of Ads: A Post-Modem View of Advertising andConsumer Responses," Journal of Consumer Research, 19(March), 556-67.

StoUer, Paul (1992), The Cinematic Griot. Chicago: University ofChicago Press.

and Cheryl Olkes (1987), In Sorcery's Shadow. Chica-go: University of Chicago.

Sutton, Robert and Anat Rafaeli (1988), "Untangling the Relation-ship Between Displayed Emotions and Organizational Sales:The Case of Convenience Stores," Academy of Management/ourmi/, 31(3),461-87.

Tedlock, Dennis (1983), The Spoken Word and the Work of Inter-pretation. Philadelphia: University of Pennsylvania Press.

Thompson, Craig J., William B. Locander, and Howard R. Pollio(1990), "The Lived Meaning of Free Choice: An Existential-Phenomenological Description of Everyday Consumer Experi-ences of Contemporary Married Women," Journal of ConsumerResearch, 17 (December), 346-62.

Triandis, Harry C. (1989), 'The Self and Social Behavior in Dif-fering Cultural Contexts," Psychological Review, 96 (3),506-20.

Trice, Harrison M. and Janice M. Bayer (1984), "Studying Organi-zational Cultures Through Rites and Ceremonials," Academy ofManagement Review, 9 (4), 653-69.

1\imer, Victor W. (1969), The Ritual Process. Chicago: Aldine.Van Maanen, John and Gideon Kunda (1989), " 'Real Feelings':

Emotional Expression and Organizational Culture," Research inOrganizational Behavior, 11, 43-203.

Wallendorf, Melanie (1992), "Holiday Behavior Seminar," presen-tation to senior marketing and brand management at OceanSpray Cranberries, Inc., Lakeville-Middleboro, MA, January.

and Eric J. Amould (1988), " 'My Favorite Things': ACross-Cultural Inquiry into Object Attachment, Possessiveness,and Social Linkage," Journal of Consumer Research, 14(March), 531-47.

and (1991), " 'We Gather Together': The Con-sumption Rituals of Thanksgiving Day," Journal of ConsumerResearch, 19 (June), 13-31.

and Russell Belk (1989), "Assessing Trustworthiness inNaturalistic Consumer Research," in Interpretive Consumer Re-search, Elizabeth C. Hirschman, ed. Provo: UT: Association forConsumer Research, 69-84.

and Merrie Brucks (1993), "Introspection in ConsumerResearch: Implementation and Implications," Journal of Con-sumer Research, 20 (December), 339-59.

Webb, Eugene J., Donald T. Campbell, Richard D. Schwartz, andLee Sechrest (eds.) (1966), Unobtrusive Measures: Non-Reac-tive Research in the Social Sciences. Chicago: Rand McNally.

Whiting, Beatrice and John Whiting (1970), "Methods for Observ-ing and Recording Behavior, in A Handbook of Method in Cul-tural Anthropology, Raoull Naroll and Ronald Cohen, eds. Gar-den City, NY: The Natural History Press, 282-315.

Whyte, William Foote (1948), Human Relations in the RestaurantIndustry. New York: MCGraw-Hill.

Wind, Jerry, Vithala R. Rao, and Paul E. Green (1991), "BehavioralMethods," in Handbook of Consumer Behavior, Thomas S.Robertson and Harold H. Kassarjian, eds. Englewood Cliffs, NJ:Prentice-Hall, Inc., 507-32.

Wirth, Louis (1964), On Cities and Social Life. Chicago: Universi-ty of Chicago Press.

Wolf, Margery, (1990), "Chinanotes: Engendering Anthropology,"in Fieldnotes, Roger Sanjek, ed. Ithaca, NY: Comell UniversityPress, 343-55.

Workman, John (1993), "Marketing's Limited Role in New Prod-uct Development in One Computer Systems Firm," Journal ofMarketing Research, 30 (November), 405-21.

Worth, Sol and John Adair (1972), Through Navajo Eyes: An Ex-ploration in Film Communication and Anthropology. Blooming-ton, IN: Indiana University Press.

Wulff, Robert M. and Shirley J. Fiske (eds.) (1987), Anthropologi-cal Praxis: Translating Knowledge into Action. Boulder, CO:Westview Press.

market-oriented ethnography: interpretation building and ... 4 market oriented.pdfwendy scheier....

Documents