~ bog nr.: L~ JUBibliotek ex.: /

93 D1~' UDK: o. 001. 13

PARTICIPATORYDESIGN:

Principles and Practices

7 Ethnographic Field Methodsand Their Relation to Design

1993

Edited byD OUGLAS SCHULER

AKI NAMIOKABoeing Computer Services

Computer Professionals f or Social Responsibility

LAWRENCE ERLBAUM ASSOCIATES , PUBLISHERSHillsdale. New Jersev Hove and London

Jeanette BlombergXerox Palo Alto Research Center

Jean GiacomiAndrea MosherPat Swenton-Wa llXerox Corporation '

In this chapter we explore the relationship between developing a descriptiveunderstanding of human behavior and designing artifacts which ostensibly sup­port the activities described. Although there is growing recognition that an under­standing of users' current work practices would be useful in the design of newtechnologies, the debate about what it would mean to acquire such understandingand to link it with design is only beginning. What are the implications ofdeveloping ways of representing the views and activities of communities ofpractice outside one's own such that the knowledge would be useful in design?

The ethnographic approach, with its emphasis on " natives' point-of-view,"holism. and natural settings . provides a unique perspective to bring to bear onunderstanding users' work activities. However, anthropology is mute when itcomes to ways of integrating such an understanding with design. The languagesof design and ofethnography evolved in quite different contexts and in relation todifferent concerns. While the ethnographer is interested in understanding humanbehavior as it is reflected in the lifeways of diverse communities of people , thedesigner is interested in designing artifacts that will support the activities of thesecommunities. The current challenge is to develop ways of linking these twoundertakings.

•We wish to thank members of the Industrial Design/Human Interface Participatory DesignProject who worked with us in our exploration of the relevance and power of ethnographic fieldmethods for design. '

124 BLOMBERG. GIACOMJ. MOSHER. AND SWENTON.WALL 7. ETHNOGRAPHIC FIELD METHODS 125

Aspracticed by most ethnographers, developing an understanding of humanbehavior requi res a period of field work where the ethnographer becomes im­mersed in the activities of the people studied. Typically, field work involvessome,combination of observation, informal interviewing, and participation in theongoing events of the commu nity. Through extensive contact with the peoplestudied ethn ographers develop a descriptive understanding of the observed be­haviors .

. Designe~, on the other hand , are interested in understanding human behaviorInsofar as rt enables them to design artifacts better suited to the needs of the~sers . ~signers J therefore . spend more time testing and evaluating the ir designsIn rel ~tIon to users' needs.and abilities and Jess on understanding the supportedbehavior per se. When designers do attempt to gain a clearerview of the users forwhom they design technologies. they tradi tionally have been limited in the wayssuch a view IS acquired (see section on Traditional Approa ches, pp. 143-147).Ethnography provides an a1 teroative method ology for designers to use , whichgives them access to people's everyday practices as members of social groups.

MOTIVATION FOR USING ANETHNOGRAPHIC APPROACH

In the 1980s there was a refocusing of interest on the part of systems designersaway from the view that technology supported individual tasks and toward theview that human activities were in large part carried out in co operation withothers. Therefore new technologies should be designed to support the cooper­ative nature of most human activities . A new field evolved called CSCW (com­puter support for coopera tive work), which was concerned with the design ofcomputer tools for the support of group work (Greif, 1988). As a consequence ofthis shift In focus, there was a realization that the methods most often used to~alyze users' needs and activities, and to evaluate designs , were not suited tothis changed focus. Looki ng at individual psychological and cognitive processesand evaluating the fit between isolated tasks, users , and technologies would notprovide the perspective needed to d esign and evaluate tech nologies for groupwork . Interest In explonng the possi ble applicab ility of ethnographic methods forunderstanding gro up work practices and for linking this understanding to designCOInCIded With , and may have been motivated by, this refoc using of interest.

THE ETHNOGRAPHIC APPROACH

There is considerable deba te in anthropology about what it means to be " doing"ethnography. At a minimum , most would agree that ethnography requires apenod of field work where the ethnographer becomes involved in the everyday

activities of the people studied . While ethnography often includes a descriptionof the activities and practices of those studied, it is more importantly an attemptto interpret and give meaning to those activities . This interpretation mos t oftentakes its final form as a text written in a somewhat stylized format (see VanMannen 1988, for a discussion of vario us ethnographic forms). Although manybooks have been written about ethnography and ethnographic field methods ,there is no agreed upon set of principles that guide anthropological field work,nor is there a cookbook of methods and techniques applicable in all situations.We offer here some general guidelines for those interested in exploring theusefulness of an ethnographic approach for the design of new technologies andpresent a brief descriptio n of some commonly employed research methods . How­ever, we would falsely portray the field if we left the impression that what wewrite here is either complete or without challenge . We have chosen not to dwellon the con troversies that exist co ncerning what constitutes an adequate ethnogra­phy to avoid complicating the disc ussion. Those interested in delving furth er intothe controvers ies should consult Agar (1980 , 1986), Clifford (1988), Clifford andMarcus ( 1986), Geertz (1973, 1983), Harris (1979), Naroll and Cohen (1970),Pelto (1970), Tyler (1979), and Van Mannen (1988), among others .

Guid ing Princip les of Ethnography

At the risk of overly simplifying the ethnographic endeavor, we begin by describ­ing four main principles that guide much ethnographic work.

Natural Settings

Ethnography is grounded in field work. By this we mean that there is acommitment to study the activities of people in their everyday settings. Thisrequires that the research be conducted in a field setting as opposed to a laborato­ry or experimental setting (see Fig. 7.1). The underlying assumption here is thatto learn about a world you don 't understand you must encounter it firsthand .

Holism

This emphasis on natural settings derives in part from a belief that particularbehaviors can only be understood in the everyday context in which they occ ur. Toremove a behavior from the larger social co ntext is to change it in important,nontrivial ways. This concern with how particular behaviors fit into the largerwhole is often referred to as holi sm (see Fig . 7. 1).

Descriptive

Based on field work ethnographers deve lop a descriptive understanding of thelifeways of the group studied. Ethnographers describe how people actually be­have , not how they ought to behave . This distinction is similar to one made in

126 BLOMBER G, GIACOMI, MOSHER, ANO SWENTON -WAL L

mmmI'-- _Particu lar behaviors understood in relation to howthey are embedded in the soc ial and historical fabri cof everyday life.

Focus on relatio nship between the parts

Resear cher/Designer

7. ETHNOGRAPHIC FIELO METHOOS

1!IImILDL-_ _ - - -----Judgements of the efficacy ofbehaviors observed arew~hh.ld

DESCR IPTIVE ln"eado( PRESCRIPTIVE _

127

Learn about a world you don 't unde rstand b yencou nte ring it first hand. Focus on natu rally

occu rring, everyday ta lk and action.

NATURAL lABORATORY

"Several people handle thedocumen t before it is completed.All Invo lved discove r probl emsand are asked to account fo rchanges to the document,"

"They're still manuallyprocessing these -routine­documents. Passing hardcopyfrom person to person is suchan ineffident way to updatedocuments. An electronic mailsystem linked to an intelligentdatabase could really improvetheir precess."

FIG. 7.1. Two princip les of eth nograpy.

linguistics between descriptive linguists (how people speak) and prescrip­tive linguistics (how people ought to speak). The orientation toward the de­scriptive leads ethnographers to assume a nonjudgemental stance with respectto the behaviors they study. Maintaining such a nonjudgemental stance issometimes referred to as cultural relativism , the notion that other people'sbehaviors should not be judged by the standards of some other group (see Fig.7.2 for an illustration of how descriptive and prescriptive characterizations con­tras!.)

FIG. 7.2. Contrast between descriptive and pr escripti ve charact eriza­ti on s of act ivity.

Members' Point-oF- View

Ethnograph y involves understanding the world from the point-of-view ofthose studied. Anthropologists attempt to understand how peo ple organize theirbehavior and make sense of the world around them. With the reali zation that onecan never truly get inside the head of another or see the world exactly as anotherdoes , researc h methods are aimed at getting as close to an insider's view of thesituation as possible. With such an orientation, ethnographers are concerned with

128 BLOM BERG, GIACOMI, MO SHER, AND SWENTON·WALL 7. ETHNOGRAPHIC FIELD METHODS 129

JrRese archer!Designer

describing behavior in terms relevant and meaningful to study participants . Th isco ntrasts with the requirements of survey researc h where relevant categoriesmust be known before the study begins and must not vary across participantcommuni ties . As such, the terms in which behavior is characteri zed often arethose of the researcher, not the study part icipants (see Figs . 7 .3a and 7 .3b for anillustration of this principle).

In general then , ethnography is concerned with understan ding other people 's

IMEMBERS ' POINT OF VIEW

Understand other peoples' behaviorfrom their point of view

Descriptive catego ries a retbose of the com munity of practice

Top "",,,,,,, ~ctKrent cases ---...Midd le drawer: ~ <l Place 'Of shareddead storage ., c::::I ~ reference books;Bot1omdlOl'M!r: pla ce to postpersonal bclongings K::I reminders, notes

RW!- ~ lom\,

PC;nh~;I"" .~

I(romMarw .Don 't knowhow to use

.... mostof the .... IRoom wher~ we sohware Typewriter still my

• can meet pnvately old standby. Used

\

fo<envelopes ,~. • • forms, quick notes'-J to bos s

Researcher! Only copi er that /

Designer th rough the will handle my ~eyes of participants \K over;ized origina ls

~---- .~ Place to meetrriends and find outabou t one anothers'activities

FIG.7.3a. Descri ptive categories of th e stu dy part ici pants as con­tra sted with those of the researcher (Fig. 7.3b l.

Contrasled With _

Descriptive categories arethose of the resear cher

Book shelv es

______~4 only partial ly

File ~ usedCab inet " c:o i'\.

/

c:l <l Desk with no

~ IOC\~Y

Nrf EJ ··6I . . IMeet ing room to Table with IBMsupport _ 286, Xerox

\~ ~7VE:J -:::_;.,..

FIG. 7.3b. Descrip tive catego ries of the researcher as cont rasted w ithth ose of the study part ici pants (Fig . 7.3a).

behavio r in the context in which it occ urs and from the point-of-view of the

peo ple studied .

ETHNOGRAPHIC FIELD METHODS

Ethnographic field work requ ires the personal involvement of the investigator, awillingness to be in situations out of one's contro l and as such an abandonme nt ofstric t " scientific contro l." It also involves an iterative , impro vizational approach

130 BLOMB ERG, GIACOMI, MOSHER, AND SWENTON-WALL7. ETHNOGRAPHIC FIELD METHODS 131

to understanding, wherein partial and tentative formulations are revised as newobservations challenge the old, and where adjustments in research strategy aremade as more is leamed about the particular situation at hand.

Although ethnography cannot be reduced to a set of methods or techniques,but must be unders tood in relation to the abo ve principles, we discuss a fewcommonly employed field techniques that have proven useful to anthropo logistsin the field .

Observation

Because ethnographers are interested in understanding human activity in theeveryday settings in which it occurs . most ethnographic investigations invo lvesome period of observation. The ahility to observe and record ongoi ng activitiesbecomes critical to the succes s of the endeavor.

Why Observe?

There is a well-known axiom in anthropology that what people say and whatthey do are not the same . This is one of the principal motivations for includingohservations of ongo ing acti vity in any study of human behavior. The distinctionbetwee n what people say and what they do is related to the distinction betweenideal and manifest behavior. Ideal behavior is what every " good " member of thecommunity should do , whereas manifest behavior is what people actually do.Sometimes asking people abou t their behavior will yield respo nses closer to theideal than the manifest. People may distort , either knowingly or unknowingly,accounts of the ir own behavio r, often simply providing an approximation con­structed either for the questioner 's benefit Ot to match cultural expectations. Withregard to this latter point, Whiting and Whiting (l 970) give examples fromstudies of child rearing where women 's descriptions of their child reari ng prac ­tices we re at odds with their observed practices. Women in the United Statesrepo rted that their children could dress and toilet themselves earlier than theyactual ly could. Whit ing and Whiting assert that this supported the value thesemothers placed on independence.

The distinction betwee n what people do and what they say is also related tothe fact that people often don 't have access to the inarticulated , tacit know ledgeassoc iated with ce rtain activities . There are many activities that are so much apart of our every day lives that we are unable to provide accurate accounts eve nwhen asked to reflec t upon these activiti es . In some cases . we may not have thevocabulary to talk abou t them.

Observational Role

There are many ways to conduct observations . At one extreme one mayatte mpt to beco me the proverhial " fly on the wall," in which case the fieldworker tries to be as unobtrusive as poss ible . Maintaining the strictly observer

role is difficult and frequently requi res being given some culturally appropria tetole that allows the observer to " hang around" and observe .? Some have charac­terized this role as that of the observer pa rticipant, where the participant co mpo ­nent is simply the culturally appropri ate status given to the ethnographer.

At the other extreme is the participant observe r . In this case the fiel~ workerbeco mes a full participant in the activities studied] and has the opportunity to bean ohserver as well . There are both advantages and liab ilities in atte mptmg to beboth ohserver and participant. One major adva ntage is that the field worker h asfirsthand experience of the events under study due to the fact of aCl1v~ parncip a­tion . In some settings the only way of gaining access to the actrvmes of thecommunity is by assuming an active role; s imply being an observer to events IS

not acceptable (K1uckhohn, 1940). . .On the other hand , the liahil ities of being both observe r and parncipant are

many. Logistics alone can be overwhelming if one is trying to participate andtake field notes at the same time . It often hecomes necessary to wnte-up fieldnotes as after the fact recollections of events participated in . Becau se the eth­nographer is ofte n new to the activities participated in, a gr~t deal. of energy canbe expe nded simply figuring out what a next appropri ate action might he . Und erthese conditions participating at the same time you are rrymg to make sense ofthe total ity of events as they unfold may be impo ssibl e . . .

As a participant one has hoth the advantage and disa~vantage of having a viewof the act ivities from a part icular van tage po int , one s partICIpatory role . Par­ticipating in a different capacity would provide a somewhat different perspectiveon the events . Whil e this may seem like a problem , every account pro vides aview from somewhere , if not from a participant's standpoint, th~n from theoutside and some particular outside at that. There are always multiple vantagepoints from which to gain an understanding of the activities of a community and

each will enli ghten in different ways.The unobtrusive and participant obse rver role s are two extremes along some

continuum. Most often one moves back and forth betw een partic ipation andobservation, variously wearing the hat of the insider and outsider. Differentstudies and observational episodes within the same study provide different oppor­tunities to e ither take an unobtrusive observer role or to particip ate in the ongo ingactivities of the community. One should not feel bound to choose among these

obse rvational extrem es .Observations of whateve r varie ty requ ire that the field worker maintain good

2While unobtrusive observation may be desirable, there areethical issues that surround observingor videotaping people for study without their awareness. To the great~st e~t~~t possible , studyparticipants should be informed about observations or videotaping of their acnvtues.

3Full participation in theactivities of the communityunder study ~as some advan.tages. However,anthropologists have cautioned fieldworkers that "going. nativ~ " might .h~ve negauve consequencebothfor theresearch project andfor the well beingof the mvesttgator. TIns IS a greater problem whenfield work is conducted far from home, among markedly differentcultural groups.

\I

132 BLOMBERG, GIACOMI, MOSHER, AND SWENTON-WALL 7. ETHNOGRAPHIC FIELD METHODS 133

relations with the peop le ohserved. The field worker must be socially skilled andaware of the sensitivities of those whose activities arebeing observed. Jeopardiz­ing one's position in the community by insisting on observing or participating inan activity, when doing so is dee med inappropria te by the co mmunity, couldeasily spe ll disaster for the project. It is more advantageous to proceed slowly,ga imng trust before insisting on access to certain events, people , and activities .Often patience and greate r familiarity with local custom will pay off. What wasoff limits initially may become open in time.

Focus of Observation

Once the deci sion to observe is made . there are still many questions toanswer. One must decide what to observe, when to observe , where to observe,and when you've ohserved enoug h (Whiting & Whiting, 1970). Answers to theseand other questions will depe nd on research direct ion. One might decide toobserve meetings (event focus), or individuals as they go through their dailyrout ine (person focus), or the activities at and around a receptionist's desk ( placefocus), or the life history of a document as it moves from office to office andperson to person (ohject focus) (see Suchman & Trigg, 199 1, for a slight varia ­tion). Choices like these become necessary because of the impossibility of takingm all that is going on in a particular sett ing and beca use different researchquestions require different observational strategies.

Although there is no fixed rule co ncern ing when one has observed eno ugh, ageneral principle is that when you're no longer surprised by what you' re observ­ing, you've probably seen enough. In other words, when you can predict whatwiJI occ ur during some period and these predi ctions are consistently born out byrepeated observation. you can be secure in believing that a range of behaviorsand activi ties have been adeq uately sampled. However, this confidence depen dson appropria te sampling strategies for the observat ional periods. For example , ifone's interest is in the activities at or around a particular location, it is importantto observe those activities at various times of the day. The morning activitiesaround a receptionist 's desk may differ significantly from those at the close of theday. The issue of sampling strategy is important no matter what the focus ofobservation (person, event, place , or object).

Note Taking

Note taking is a very individual activity, but it is one of the important linksbetween the field experience and how one later interpre ts that expe rience (Jack­son, 1990). Al though field notes are never a complete record of the experiencesand observations of the field worke r, they are often used to evoke memories ofexperienced eve nts . As such, field notes are most useful to the original fieldworker.

There are many strategies that have been developed for takin g field notes .

Different situations and different observational roles will influence the pos­sibilities for taki ng field notes . As mentioned ear lier, the full participant role maymake it impossible to take notes while events are unfold ing. Some situations maydictate that only sketc hy, temporary notes can be take n at the time of observa tion,the intent being to expand the notes soon after the event is over. Other situationsmay allow taking very extensive notes which might include verbatim transcrip ts ,paraphrases of things said, interpre tations of activities observed, as well asdescrip tions , maps , or diagrams of the setting or the movements of people andartifacts within the setting . When field notes include different types of informa­tion, it is often important to indicate the status of the information. Is it ~ v~rbatim

transcript or simply a rough paraphrase ; is it a member's stated motivanon forparticipating in the event or a conjecture on the part of the ethnographer? Wh ilethese distinctions may ·seem clear enough at the time the notes are taken, onemust remember that field notes may be referen ced months or years after theywere taken . In this light it is critical tha t field notes include at a minimum a date ,time, place , and listing of the persons present.

Videotaped Records as Notes

Video cameras have come to play an increasingly impo rtant role in eth­nograph ic studies (see section on Video Anal ysis). They are sometimes used as asupplement or even substitute to field notes . When memory fails or field notesare inadequ ate , the videotape may be able to provide some of the missinginformation. However, caution needs to be exercised when relying on the videorecord for a co mplete record of observed events . Viewing a videotape at a laterdate and experiencing an event firsthand provide the researcher with differentkinds of access to the activities in question. While the camera records thoseactivities within its field of view, part icipating in the activity allows one to abso rbthe " taste, smell, and feel " of the activity, and to refocus attention in response tounfoldin g events . A video tape can not capture the ways the eve nt is experiencedby the observer/partici pant.

Suppl ementing videotaped records with field notes also can help with lateranalysis of the video tapes . Anyone who has recorded hours of video tape know show laborious it is log the tapes for content. Careful, painstaking analysis oftenis conducted on only a very small portion of the available videotape . Field notescan be a great help in se lecting segments of the tape for more focused anal ysis.

Observations Coupled With Interviews

Observations seldom stand alone and are frequen tly coupled with interviewsand informal discussions. Because ethnographers make no assumption that ap­propriate questions or ways of asking them are known in advance. observationsprovide ethnographers with one way to learn how to ask appropriate questionsfrom the point-of-view of the memb ers of the community under study. Observa-

134 BLOMBERG, GIACOMI, MOSH ER, AND SWENTON -WALL 7. ETHNOGRAPHIC FIELD METHODS 135

tions and informal discussions also follow interviews providing the opportunityto observe behaviors previously described by interview respondents .

Int e rv iewin g

Informal interviewing is a part of most ethnographic research. Early in field workthese interviews are intentionall y unstructured and open-ended to allow the par­ncipants to help shape the discourse, the topics to be discussed , and the relevantways of talking about them . There are few assumptions going in about whatquestions to ask or how best to ask them ' The ethnographer may have generalareas of interest to explore, but if the discussion moves away from these areas toissues more "relevant" to the respondent. this is likely to be viewed as anopportunity to learn about the commu oity rather than a situation to be avoidedlest field work proceed too slowly. Throu gh participation in community activitiesand through informal interviews the ethnographer begins to learn enough aboutthe community to conduct more structured, systematic interviews. The moti­vation to begin with informal interviews, as opposed to structured ones. derivesfrom a belief that asking highly constrained questions before enough is under­stood about the situation likely will produce poor quality answers. In such asituation questions may be understood in unknown ways or may be irrelevant tothe respondent, in which case the researcher might get a mistaken sense of therelationship between the ques tions asked and the responses given.

It also should be noted that interv iewing, whether formal or informal, is not asimple way of recording objective fact (Mishler, 1986). As mentioned in thepreceding section on observation, what people say and what they do are not thesame thing. Asking people to describe some activity in which they engage willnot produce the same insights as one would gain by observing people engaged in .the activity. In the same way asking people about their beliefs will not tell youabout how those beliefs are manifest in their everyday activi ties. If we simplyrelied on the interview as our window to some objective reality, we might cometo some very erroneous conclusion about the lifeways of the communities westudy.

If instead we view the interview (following Briggs, 1986) as a communicativeevent, we must know something of the interactional dynamics that shaped theinterview to make sense of the discourse (the questions and responses). As Briggs(1986) states , " Like speech events in general, it (the soc ial situation created bythe interview) shapes the form and content of what is said" ( p. 22). The implica-

"When field work is conducted among people who speak a different language from rbe eth­

nographer. a point is madeto learn their language so that questions can be askedin the language andvernacular of the study participants.

tions of this are twofold. First, the context' in which the interview takes place isnot neutral with respect 10 the discourse that ensues . Second, to make sense ofwhat is said during an interview one needs to exami~e th~ ~ial construction ofthe questions and responses in relation to the SOCIal Situalton created by the

interview itself (Suchman & Jordan , 1990).

Interview Location

Because ethnography is field-work based , interviews most often occur i? thelocal setting. There are advantages to interviewing in the respondent ~s envl~~n­ment. ot only are the respondents more likely 10 feel comfortable In familiarsurroundings , but they have access to people and objects that may figure into thetalk as it unfolds. If a respondent is trying to describe an activity ID which heparticipates , having available the artifacts , physical surroundings, and peoplethat typically belp shape the activity can be a resource for the talk . Th is said onemust also be aware of situations where the respondent 's environment does notprovide the privacy needed 10 talk about'some subjects . If other members of thecommu niry can ove rhear, respo ndents may restrict what they are willing to talkabout. The point here is that the setting for Ihe intervie w is never n.eutral withrespect to how the interview unfolds and this must be taken into consideration m

any subsequent analysis of the responses .

Contextual Interviewing

Interviewing can be combined with observation w~e.re the re.sea:che~ inter­views respondents while they are engaged in some aCI'v,~y, In thts situauon theresearcher " interrupts" the observed activity to ask quesnons, motivated by theobservations being made . However, there is a trade off here. By asking a ques­tion about some activ ity, in the context in which the activity is taking place , onewill influence the course of the activity. Bearin g this in mind , the contextualinterview can provide access to information that observation alone might fail to

uncover.

Who to Interview?

Because it often is impossible to interview all members of a community,decisions must be made concerning who to interview. These choices are influ­enced by the research questions being posed , the availability of respondents , the

.5'The context is not simply thesum total of lhe physical andsocial characteristicsof lhe situa~n _Thecontext is conlinu.ally being jointly produced by the participants and as such cannot be easilyseparated from me activities and talk in question. See Cicourel ( 1982) and Mehan (1979) for a

discussion of these issues.

136 BLOMBERG , GIACOMI , MOSH ER, AND SWENTON-WALL 7. ETHNOGRAP HIC FIELD METHODS 137

project time frame, and a co ncern with sampling the views and activities of across-section of the community. This issue of representative ness was not alwaysa concern for ethn ographers . Early ethn ographers were not particularly con­cerned to find themselves in situations where they were relying on informationelicited from one individual because they held the view tha t any member of acommunity could provide information about the beliefs, values , and customs ofthe entire community. However, as ethnographers began to question assumptionsabout cultural shar ing , the practice of relying on a single informant (eth­nographers' traditional term for respondent) all but disappeared . Ethn ograph ersquestioned whether the chosen informant (often a "westernized" or marginalmember of the community or one who shared characteristics with the researcher)could adequately represent the beliefs and practices of all community members .Rely ing on a single informant became known in anthropological circles as the" well-informed informant problem."

Rules of Thumb in Interviewing

The ethnographic interv iew is not bound by explic it rules . To the contrary, agreat deal of latitud e is given to the intervi ewer to exploit the particulars of anygiven interview situation (see Suchman and Jordan , 1990 for a discussion of thepitfalls of rigid interviewer guidelines). Nevertheless, there are still a few basicrules , or what might be better described as interactional style considerations , thatshould be menti oned . First, it is important to allow respondents to help shape thecontent and character of the interaction. In keep ing witb this precept, it isgenerally not advisable to interrupt unnecessari ly, to complete the respondents'utterances , or to answer your own questions . While the interviewer may be ableto anticipate what a respondent is going to say, much more is learned by allowingthe respondent to answer the question unaided by the interviewers' presupposi­tions . If the interviewer is mistaken about the answer contemplated by the re­spondent, there is the risk that the respondent will accept the interviewer 'srespon se, thinking it "sounds better " than the one contemplated or will not wantto contradict the interviewer. In any event, an opportunity to learnmore about therespondent's world will be lost.

A second guideline in ethnographic interviewing is that rapport with therespo ndent should not be sacrificed to obtain a response. The respondent maychoose not to answer a question for any number of reasons (e. g., too personallyrevealin g, co ncern that the information will get to other parties, inability tounderstand why the interviewer wants to know such information, etc .). Eth­nographers rely on being able to find new, more appropriate ways of askingquestions as they learn more about the setting and as the respondents becomemore famili ar with them .

Third , it is important that tbe interv iewer be willing to acknowledge knowingless than the respondent. After all, the point of the interview is to learn some-

thing about what the respondent knows or how the respondent sees the ~orld .Although the interv iewer may be more knowledgeable about some topic , theobjective of the interview is not to produce the most accurate or co mpleteunderstanding of the topic, but to gain a better unde rstanding of what the respon­dent knows and thinks about the topic . The purpose of the intervi ew may bedefeated if the interviewer is more concerned with self-aggrandizement thanlistening to what the respondent has to say. A related problem when ethnographyis part of a design project is the tend ency of some designers , whose Job after all isto solve problems, to come up with solutions to a respo?dent's expressed prob­lem befo re takin g the time to get an adequate understanding of the problem. Thepush to solve design problems may conflic t with taking the time to fully appreci­

ate the nature of the problem .

Video Ana lys isThe use of video cameras in ethnographic research is on the increase with thegrowing availability of inexpe nsive , small , portable equipment '. There are avariety of ways ethnographers make use of video records of activity. For somethey are a supplement to field notes , for others they are used in teaching andrepo rting situations, and for still others they are the primary data for analysis ."Space limitations preclude a detailed discussion of the techmques used 10 theanalysis of video records , but we offer a few general arguments for bringing

video cameras to the field .

Why Videotape?One of the strongest arguments in favor of videotaping is tha t human activities

unfold so fast that it is impossible to capture their complexity by observationalone (Jordan , Henderson, & Tatar, in preparation). Field notes are onl y a partialrecord of activiti es observed or participated in, and words are often inadequate todescribe what is observed , including bodily movem ents relevant to the analysis .The videotape preserv es these actions for careful viewing and analysis . Vid­eotaped records also allow one to look at an activity from different perspectivesfrom the one held at the time the video record was made.

As the field work progresses and the researcher develops new understandingsof the activities , new perspectives can be brought to bear on the activities pre­viously recorded . The abi lity to review videotapes also allows the researcher theopportunity to correct erroneous characterizations and interpretations (Suchman& Trigg , 1990). In this sense the researcher can corroborate the field record

(notes) with the video record .

6Many ethnographers employ analytic techniques developed in the~ of conversational .andinteractionalanalysis. Those interested in learning more about these techniques andthe assumptionsthat underlie them should refer to Atkinson and Heritage (1984) andGoodwin and Heritage (1990).

138 BLOMBER G, GIACOMI, MO SHER, AND SWENTON -WALL 7. ETHNOGRAPHIC FIELD METHODS 139

Another advantage of having a video record is that it can be made available topeople not present at the time the recording was made. Field notes absent anaccompany ing elaboration by the person who took them, are of limited use toanyone but the original note taker. On the other hand, videotapes of activity canbe viewed and analyzed by a wide range of people (other researchers, designers ,product developers, etc.). The participation in the analysis by people involved inthe original videotap ing often can facilitate the analysis and contribute to whatcan be learned from the tape .

Videotaping ongoing activi ty also provides another example of unobtrusiveobservation since a camera can be se t up and the researchers can leave thescene." Despite the opportunity to observe without influencing the course ofevents, in some circumstances it may not be desirable for the researcher to leavethe ~cene ' .There are times when the investigator's presence is appropriate; when~he In: estIgator wants to be a participant in the recorded activity or when theinvestigator wants to be present to ask for clarification and elaboration . Theinvestig~tor's presence in the scene should not determine whether videotaping isappropriate , but an awareness of the possib le influence of the investigator 'sp.resence on the activities should be taken into consideration in subsequent analy­SIS.

Problems With Videotaping

Videotaping is not problem free . Because one can quickly generate largequantrnes of tape all of which cannot be analyzed in detail, the researcher facesthe. problem.of identifying sections of the tapes for care ful, painstaking analysis.This IS facilitated by logging the tapes soon after they are made, annotating themWIth general descriptions of activities and highlighting places where particularresearch questions are addressed. Field notes can be useful in producing thesecontent Jogs, but if one was not present when the tape was made, the contentlogging activity will require viewing the entire tape at least once to produce anadequate content log. This can amount to hours of work. Once the content log iscompleted, particular sections of the tape can be selected for later carefu l analy­SIS.

Analyzing videotapes is a time-consuming activity that cannot be delegated toothers. Unlike survey researc h where interviewing participants, coding re­sponses , and conducting statistical analys is can be done by others, the job ofanalyzing VIdeotapes by repeated viewing can be done only by those who will beengaged in interpretation. Insights come only through partici pating in carefulanalysis (Heritage & Atkinson, 1984; Suchman & Trigg, 1990).

1While. (he influence of (he camera's presence on behavior cannot be ruled out, in most cases (hecameraquickly becomes part of the backgroundand only occasionally surfaces in the participants'awareness.

Another problem with the use of video cameras is that some human activitiesare difficult to record on videotape . There are times when a space may be toosmall to get an acceptable view with the camera, or when noi se level s are notadequate for suitable audio recording. The spatial distribution of an activity mayrequire multiple cameras to adequately record the activity and the changinglocation of an activity may require mobile cameras . Multiple (mobile) camerasare not possible in all settings and one may have to settle for recording only partof the activity. If it is poss ible to use more than one camera, new analyticchallenges emerge. While the use of split screen images , time code to syn­chronize multiple recordin g, and computer controlled editing equipment makessuch analys is possib levi t is difficult nonetheless.

UNDERSTANDING HUMAN BEHAVIORS AS AMECHANISM FOR CHANGE

As mentioned, ethnography is a way of developing a descriptive understandingof human activities . Insofar as such an understanding can be brought to bear ondesigning new technologies, its role as a mechani sm for change must be consid­ered. To greater and lesser degrees new technologies always resul t in change forthe communities into which they are introduce d. As Ehn (1988) puts it, " Whatwe design is not just artifacts but by intervention a changed or reformed practice "( p. 128). As such those involved in linking ethnog raphy and design must beaware of their role as "change agents." This raises the question . as it does foranthropologists who act as change agents in more traditional settings ; " In whoseinterest does one operate?" Does one serve the people for whom new technolo­gies are designed (those whose activities are the subjec t of the inquiry) or doesone serve the sponsors of the work? Arensberg and Niehoff (1971) contend themain concern of the anthropologist involved in promoting change" . .. must bewith the people who he hopes will accept the new ideas" ( p. 7). In many casesthese are the end-users of the new tecbnologies .

Because ethnography typically involves extensive contact with the peoplestudied and an attempt to "see" the world through their eyes, ethnographersfrequently identify with the interes ts of those studied regard less of researc hsponsorship. Van Mannen (1988) writes, " . .. the fieldworker not only repre­sents but takes the side of the studied" ( p. 42). This orientati on toward theconcerns of the people studied is a central characteristic of anthropo logy and hasbeen codified in the " Principles of Professional Responsibility" adopted by theAmerican Anthropological Association in 1971. The first principle states, " Inresearch, an anthropologist's paramount responsibility is to those he studies.When there is a conflic t of interest, these individuals must co me first." (Ethicsand Anthropology: Dilemmas in Field work , p. 183). As anthropologists becomemore involved in systems design and development, and as ethnographi c field

140BLOMBERG. GIACOMI. MOSHER. AN D SWENTON.WALL

methods becom e more widely used in such ffi ".involved in this work to reflect upo th the orts, It IS unportant for thosewhose work is su d n ese e leal concerns. In particular, those" both user advoc~:"Jrte d

bytechnology companies must ask ifi t is possible to be

'. es an purveyors of technology?" (Kathleen Carte Icomm unIcatIOn). r, persona

PARTICIPATION IN FORMULATING DESCRIPTIONS OFNATIVE PRACTICES

Little has bee n written abo ut the role of those studi d i .of their own practices Wh 'j th h em formulatmg descriptions

. lee et nographer may ttexperiences of those studied from th " . , . a emp t to represent the. . e natives pomt-of-vlew " act i . .

non of the studied in reviewing and J ' . • ive part tcipa ,rare . At most , the studied may be ana yzmg the e thnographer's formulations isaspects of the ethnogra hers field come mvolved 10 check ing the "accuracy " of

sentations of native p~tjces ofte~O::'d~e;I~th~ographer's accounts and repre­ticipat ion of those studied is rarel . pe .far from the field where par­practice of not involving the stud i~l:slbl~. eve~ if desirable In defense of thestudied are not in a good posit' t 0; irect y, Some have argued that thosemost likel sim I II: Ion 0 re .ect on their Own behaviors and wouldiors if as: ed t/pYart° er POt st hoc ratIOnalizations or j ustifications for their behav-

rcrpa e 10 analysis ,Ethnograp hers beco me concerned with issu f " .

come involved directly in maki es 0 partIcIpatIOn when they be-tices . Whe n they take the role OI;;h';,commendatJOns for changes to native prac- .material effects on those studied ge agents and their formulations could haveNiehoff ( 1966) " ' partrcipatinn becomes an important issue.

• writes In a casebook of soc ial change ,

No project which will effect soc ioeconomic cha .recipi ents do not panicipate It . " nge can posslbty succeed if theneglected to make Sure th~y' it · IS Surp~smg ho~ frequently action agents have

ave co mmitted partICipation on th - .appears [hat the principal reaso n Why this ' ' . err projects. Itso often is that many technical advi hPnm":Y mgredlent has been overlooked

. . . rsers ave viewed their task as ' )providi ng some kind of tec hnical solution ( p. 18). Simp y one of

Anthropologists ofte n are brought into action- '. .reali zation that the change agents need a be oriented projects when there IS thebeliefs of the " beneficiaries" of th . tter understanding of the practices and

i~volved in the specification and i~t~r~~~~~na~~ : at these jndi.viduals need to begres . The anthropologist beco me a li~k to th e dnew pract ices and technolo-

In attempting to link eth h ~se In rgenous comm unities .nograp y to design it i .

involving those studied in the specification of the ~~w I:e::~rtant that ways ofbe developed. Thi s is needed fo al . no ogles and practices

r sever reasons: First , by involVing those

7. ETHNOGRAPHIC FIELD METHODS 141

studied it is possible to gain new understandings of the studied activi ties inrelation to the evolving design . Second , insofar as researchers and designersoperate as "change agents," they need to respect the interests of those studied inwhatever technology solutions are developed. Third , in those case s where thosestudied will become the actual users of the new technology, their early involve ­ment may aid in adoption later on.

EXPECTATIONS OF THOSE PARTICIPATING IN ANETHNOGRAPHIC STUDY

Ethnographers typically promise little in return for the opportunity to study acommunity of people other than to avoid major disruptions in the community'sactivi ties and to look after the community's interests in subsequent interactio nswith outs iders (e.g. , public ation s , presentations, discussions). The argument hasbee n made that ethnog raphy is exp loitive to the degree that it is not concernedwith the use of the knowledge gained to better the condi tions of the studyparticipants. Strathem (1987) write s that people may experience exploitatio n".. . when [they] perceive that others have the power to tum data into materialswhose value cannot be shared or yielded back to them in return ( p. 20) . Somehave argued that one reaso n ethnogra phers often study the less advantaged isbecause these communiti es are not in a position to demand something in returnfor participation in the study. Nader (1974) states, "Anthropologists might in­deed ask themselves whether the entirety of field work does not depend upon acertain power relationship in favor of the anthropologist , and whether indeedsuch dominant- subordinate relationships may not be affecting the kinds of theo­ries we are weaving" ( p. 289). Issues of access beco me salient as ethnographersattempt to study franchised and more powerful co mmunities .

When ethnography is a part of a technology development effort issues ofaccess and reciprocity must be confronted. In some situations members o f fran­chised communities must be co nvinced to allow acces s to the settings in whichthey work , without the promise of providing them with a technology solution.This may be the case beca use the technology under deve lopme nt may neverbecome commercially available or, if it does , it might be years before it is on themarket. The abili ty to gain access to the communities of study and the promi sesthat can be made about materially bettering the lot of those studied is directlylinked to the type of technology development effort undertaken.

WHY IS ETHNOGRAPHY RELEVANT TO DESIGN?

Ethnography is relevant to design for several reasons. First , since designers oftencreate artifacts for work settings they know little about, some understandin g ofthose settings is needed so that the technologies suit the situations of their use .

142

TRA DITION AL APPROACHES

To put the preceding discussion of ethnographic field methods in perspective , itis useful to describe some traditional approaches used to provide designers withsome understanding of user needs and behaviors. Following these descriptionswe discuss differences between these approaches and ethnography.

Because of differences betwee n the languages and perspectives of ethnographyand design and because the ethnographer is likely to have little knowledge orappreciation for the immediate concerns of the designers , this is not a simpletask . Making the findings of an ethnographic study useful for day-to-day designconcerns becomes a major undertaking .

Second, an ethnographic study might be undertake n by a team of investigatorsconsisting of ethnographers and designers. In this case the insights and under­standings, in part, would be embodied in the experiences of the desig ners whowere firsthand participants in the study. As Penniman (1974) ohserved, experi­ence underlies all understanding of social life. Active involvement by designersin the field work and in cons tructing interpre tations of the work activities at thestudy site also would help focus the ethnographic study on issues more central tothe design task and would make the interpretat ions more relevant to the design .

Third , a project could be undertaken by a team of ethnographers. desig ners .and users. The understandings and insights derived from the study would notnecessari ly be represented in a wrinen report , but instead would be reflected in acodesigned art ifact. User partnership in developing and evaluating the tech­nology in relation to current and imagined work activities would be aided bydesigner participation since designers would bring knowledge of technologyconstraints and opportunities to the collaboration. The success of the projectwould be evaluated on the basis of how well the technology supported the workactivities " In this last situation the ethnographer would adopt . in part, thedesigner's orientation of seeking to understand human behavior insofar at itenabled the design of artifacts better suited to the needs of the users.

1437. ETHNOGRAPHIC FIELD METHODSBLOM BERG, GIACOMI, MOSHER, AN D SWENTON-WALL

Second, because technolog ' h I h(Ehn, 1988' Blomberg 1987 '~~88e )p s ape the work practices of their users

, be', " , a , It IS Important that the designers' w Id

;:;:;i~~ th~m:::t~~::~~r:~~a~e::C~::I~;~~I:~;::;;~rs~av':l in~ informa~~nrely on their ow n experiences and im . . ~ e s.t ey can do IS

technologies better suited to their ne:~~n~~a:n::~~~~~ng the nlsk of designingThird there - . . e actua users.

uses are,~nkno= ~~~~lJ~~~a;;:::e~~~7:~~;~~t~:~~~~~~~gie: whose possiblean applica tion. Some understanding of the work in which nOtogy~n search of

:~~aged can help identify possible uses aod refine the Origi: 1~:hno~~~ :~

Fourth, since the user's experience of a techn I "context of its use (Blomberg 1987 1988 ) " a ogy IS Influenced by thetechnology use than that arriv '. , a • ~~mmg a broader perspective onfocus on the human-machin:

ddat ~ough tradit IOnal operabihry testing (with its

Approaches p 144 Tor a d ya. IS Important (see secnon on Traditional. , . ,' . escn pnon of operability tests).

giv=,~~~:::e~~~:gm~g radical ly new technologies , users often are unahle to

gies. They n~ed to ::P::i~Juenthes about how they might use such techno10-

h WI a way of envtstoning and expe . . thtee nology in the context of their ow k uracti nencing en war practices before they ibto such a discussion. To create the context for such a discus . cdan cbe°ntn utepartners ' th . . sron an to useful

. In e JOint exploration of the relation between work d h

:::~~~~r~:;;thave some understand ing of the user's work (BI::~~; ~o~:~:

theF~~~ir; ~fet~~nhg~:;:~i~:';;;~: ~~;;,":: :: c:nology design efforts is ill-suited tosingle task or the tasks of the single use . s lnte: ratlOn. SImply focusing on aarticulates with that of many others Ii r Ignores

low the work of one individual

may rely on the work of documen . or examp e, a pn nt shop operator's work

artists , sales representatives and ~a~~~~~~W;rdtroc~Slng specialisu , g raphicoperator 's work should be desi . . ys ems at support the pont shopwork of these oth . . esigned WIth Some larger understanding of how the

ers Impinges upon the work of the print shop operator.

LINKING ETHN OGRA PHY A ND DESIGN

There are various ways one mi ht'. . .ring the knOWledge gained fro~ .,:magme acq~mng , representing. and transfer-in the context of technology design ~eno~PhIc analysIs of user work practicestrained ethnographer mi ht be as ' W I m enti on only a few of them . First, a

The insights from this ~tudY mi~:r t~h:~%~e w~r~actices of Some gro up.written reports a d I ' ans e to deSIgners throughidentifying the r~le~:t~:~:~~~~ . The deSi!ners would then have the task of

e reports .or their particular design efforts .

8Wolcott (1990) questions whether such an undertaking should be co nsidered ethnography at allsince , ". . . the research process deserves the label ethnography only when the intended product isethnography (e. g ., some written account or cultural interpretation)" ( p. 47). Van Mannen (1988)makes a similar point in distinguishing between doing ethnographic field work and producing anethnography: ..Ethnography as a written product, then has a degree of independence (how culture isportrayed) from the field work on which it is based (how culture is known )" ( p. 4). In situati onswhere there never was the intention of developing a written account of the practices studied other thanas required by the design effort, can we truly consider such activity ethnography? If not , how is thefield work that accompanies such a technology development effort different from more traditionalethnographic studies?

144 BLOMBERG. GIACOMI. MOSHER. AND SWENTON-WALL

Customer Surveys

Customer.surveys i.n~oJve administering a standard questionnaire to customersabofillt potential, or e~l stlng . products with the intention of determining customepre erences for certa in technologies and technology features Th rtempt to provide information about the nature of the custome~' eseks.urve

YI s.at­

to the tech nology i . C' . wor In re atronh hi Y In question. r or example, customers may be questioned about

~~~ t~:;gSt~: ~yepete~~n~I';;ie~ currently in use, the frequency with which they, as s t ey perform on them, and how the view

features or technologies. Customer surveys typically are conducte: by m;~::~:::;v~ ~Ither : t the very early or late stages of product developmen t. The

n age 0 customer surveys IS that they provide responses from a lar e; prese ntatlve sample of customers in a short amount of time. Typically ~ '. esl~ comm

lumty receives a tally of responses and a summary statement o~tlin~

ing e resu ts of the survey.

Operability Assessments

task: approach i nvolves asking potential users of a product to perform severalas s or operations USing a simulator or worki ng prototype. Users are asked

perform tasks designed to test features of the user interface . Measurements aret~en to . reco rd the amount of time required to complete a task the f uenceWIth which tasks were successfully completed , etc. Problem are':' are i':e:,tifi';;~n~ highlighted as the test proceed s. Designers not involved in the operabilityes are.given use.f perfo~ance ~tatist ics , short descriptions of problemsencoun­tered , and potential solutions to Identified prob lems Operabilitusually d d . . I Yassessments are

con ucte In a laboratory setting by members of the design communitmost frequently at very late stages of product development .? y,

Focus Groups

b Foeus groups bring together individuals from a cross-section of the custase to evaluate products and product concepts in a discussion grou format .

r:tarket researchers work WIth designers to obtain a description or ch:acteriza~tlon of the product or concept under review. While designers may .d .on tOpICS to be covered in the focus group, they rarely participate dre'::~ e;:~~group d ISCUSSIons usually include some consideration of the PartiCipan~' kacuvi nes , environment , and future needs Although th di d f worvideotaoes someti . e une tte rocus group

apes sometimes can be obtained for independent analysis , most frequently

9Blomberg ( 198Sb) notes "WhiJ [ bTinterface design or with the f~nct iona~tyo::i~:blest:l :y reveal some ~roblems wi~h the userlaboratory environment has distinct characteristics W~Ch d~ff,~olo~y•.~heY fad to recogmze that thements in which the technology will be used d I er 10 srgm icant ways from theenviron-

on a ay-to-day basis in the labo I .user has none o f the soc ial resources that are il bl " " " " . " ra ory sem ng theavar a e In most work environments."

7. ETHNOGRAPHIC FIELD METHODS 145

designers are given video taped summary documents which highl ight significant

issues talked about during the focus groups.Designers and market researchers rece ntly have been exploring the use of

storyboards, scenarios of use , mock-ups, models, and concept videotapes toelicit comments from focus group participants about the relevance of new tech ­nology concepts to the ir work. Customers are asked to react to these representa­tions of technology concepts from the standpoint of the ways in which thesetechnologies might support current work or might enable them to accomplishwork in the future . The use of such representations addresses the problem of howto provide adequate verbal descriptions of a technology concept such that focusgroup participants might imagine uses for the technology.

Field Trips

Field trips provide desig ners with one of their only opportunities to observeand interact directly with users in the users' workplace . Field trips typically are

of two kinds; field visits and field tests .

1. Field Visits. Field visits are intended to familiarize product developerswith customers and the use of products within particular market segments. Thesevisits , typically lasting no longer than two hours at each of several customersites , take place anytime during the design cycle . The product developmentteams help set up the visits but, once on site , designers and other team membersinteract on their own with customers. Questionnaires , interviews , and brief ob­servations are used to obtain information about the use of current products at the

site , as well as possible future customer requirements.

2. Field Tests. Field tests usually take place following the placement of anew product in the customer's work place and are used to gauge the success of aproduct as well as to identify opportunities for its improvement. Such tests areconducted by the organizations responsible for product sales and installation,with support from the product developmen t teams. Questionnaires and informalinterviews are used to obtain users' views on product performance. The results ofthe tests are made available to the product development teams.

Each of the traditional approaches sketched above has strengths and weak­nesses. Some are more appropriate in the early phases of the design cycle , whenknowledge about the customer and potential uses of a product can be used toshape the characteristics and definition of the product. Often in later stages of thecycle, the designer can do little more than verify that the design is acceptable tousers or make minor changes to those aspects of the product with which the usersare most uncomfortable . Issues of cost , impending design schedule , and the

146BLOMBERG. GIACOMI. MOSHER. AND SWENTON-WALL

availability of key people all are .of one approach over another. Important factors that can influence the selection

Contrasts Between Traditional andEthnographic Approaches

Traditionally, the User interface desi ner' . .functionality of the technology avail gbl s rhole has been limited to maki ng thebe . I a e to t e end-user Id 11 d .considered user-advocates in that th . ea y, esrgners mightthe sys tem by articulating their ed

eyredpresent users as the human element in

d . ne s an req uirem ents . .es ign solutions with the devel . In uegouations about

oprnenr team AIl too ft hcontact designers have with .. ' 0 en, .owever, the only

. users IS 10 laboratory 0 biliproduct IS placed in the user" . pera rty tests or after a

. s environment Since th .dun ng critical phases of design th desi e user IS out of the loopexperience and imagination Cre'all"e eSlgnerfmu~t represent the user based on

B • ng a sort 0 " virtual" userY contrast ethnography provides for an on oi . . .

based on designers' firs thand kId g mg relationship with Users. now e ge of the users' rk .109 are som e of the ways tradi tional wo setn ng. The follow-ethnography differ. approaches to understanding users and

The Context of the Designer-User Interaction'

All of the approaches listed earlier exce t field .the users' work place Product ' p tnps , are conducted outsideare viewed and eval ~ated in a~o~:::ts, ~ototypes or ear ly engineering modelstasks . The quality of the infonnaf o

g lOelseltmg, often utilizing hypothetica l

I n IS un ited by the des ' , bilirepresent the concept in relation to ' . igner s a I ity toability to envision themselves using t~~~~~:~7..med workplace and the users'

rs«). .ocus IS on the Technolog y, not the Work

Traditional approaches to understandindriven. The focus is on obtainin ans g user needs are largely technologyability of a particular techno! g wers to specific quesu ons about the accept­understanding the relationshi;g~~oncep~ prototype or product, rather than onsign-ed to support Technology ~ ised i e technology and the work it is de-d . . - ocus techniqaes provide Iittl

eslgners to learn abo ut the everyd k nracti I e opportunity foray wor prac tices of potential users .

Users are not Collaborators in theTechnology Development

Traditional approaches provide little room for .ers and users over the evolving desi but i collaboratIOn between design­verbalize their needs Or to expos es~gnd ut Instead rely on the users' ability to

e ina equacres of the des ign in isolated tests

7. ETHNOGRAPHIC FIELD METliODS 147

which take place on single occasions (e. g .• in focus group discussion or oper­ability tests).

Ethnography on the other hand has the potential of providing a context where ­in mutual understanding between users and designers can evo lve . Armed withknowledge of user work practices gained through direct observation of users atwork, designers are in a much better position to accurately, and more fully,incorporate users' perspectives in the design , with the potenti al of improvingexisting products as well as identifying opportunities for new products . However,incorpo rating the ethnographic approach into design and product developmentefforts requires some reorientation. Des igners must develop skills in interview­ing, observation, analysis, and interpretation, while development teams must bewilling to shift their emphasis to support ear ly and co ntinued user involvement.These are considerable investments, but the potential benefit s also are high .

A PROJECT TO LINK ETHNOGRAPHY AND DESIGN

Researchers at the Palo Alto Research Center (PARC) and designers in theIndustrial Design/Human Interface (ID/ HI) department at Xerox recentl y bega ncollaborating to explore new ways of directly linking ethnogra phy and design.This project, the Parti cipatory Design (PO) Project, brings together individualswith backgrounds in anthropology, graphic design, human factors, and industrialdesign . Together with users from selected field sites , they form the ParticipatoryDesign team .

Goals of the Project

The PD project has multiple , interrelated goals which invol ve understanding userwork practices, deve loping new ways of incorpora ting such an understandinginto everyday design practice , and integrating the lessons learned from thisproject into Xerox product development.

Characterizing User Work Practices

The PD project is concerned with giv ing designers new and better ways ofgaining an understanding of users' everyday work practices, with a foc us on therelation between technology and human activity. Of parti cular concern are theways current technologies support work activities and how work practices inte­grate a collection of technologies into a system of activity. The application ofethnogra phic field methods is the vehicle by which such an understand ing isbeing developed .

148 BLOMBERG. GIACOMI . MOSHER. AND SWENTON-WALL7. ETHNOGRAPHIC FIELD METHODS 149

Developing New Design Practices

The PO project also is aimed at changing current design practices to allow foruser parti cipation from the begin ning of product design . The use of eth­nographic-sty le field work places designers in the users' work setting whichprovides an opportunity for continued involvement of users in the desi gn processand for design iteration in relation to actual situations of use.

Integration With Product Development

The PO project also is focused on building on the experiences of this project tohelp shape new, more participatory product development processes througboutthe co mpany. The strategy is to introduce the benefi ts of a participatory approac hto produ ct design by exam ple and by involving individuals from other organ iza­tions (Marketing and Product Planning) in some of the work.

New Ways of Working

The PO project has requ ired that designers learn new skills for acquiring anunderstanding of users' wo rk practices and that anthropologists learn new, 000­

text based, ways of representing the insights gained from an ethnogra phic study.Because most trad itional appro aches to understanding users are not field workbased, des igners involved in the projec t wanted some grounding in ethnographicfield methods before they embarked on this projec t. Two workshops were held onthe topics of Ethnographic Field Methods and Qualitative Data Analysis . Thefirst workshop provided designers with a perspec tive on ethnography, as well aspract ical skills related to observ ing work practices , co nducting ope n-ended inter­views, and making audio and video recordings in the field (see Appendix 1-forfield exercises used in Ibis workshop). The second workshop focused on ways ofanalyzing and interpreting the information acquired from interviews, observa­tions . and video recordings .

Scope of the Project

The PO team has worked primarily with one user co mmunity that was se­lected beca use (a) a broad range of technologies (fax, computers , printers, type­wri ters , etc.) were in use at the site , (b) the work was shaped by a rich array ofdocume nts (both paper and electron ic), (c) groups were linked through a varietyof media, and (d) peop le at the site were eager to take part in the projec t. Thiscombination of characteristics enabled the team to exp lore how information wasorganized and disseminated, how various activities we re coordinated, how docu ­ments helped struc ture activities, and how work practices integrated stand-aloneoffice technologies into a coherent system. These were all important issues forthe kinds of technologies members of the PO team were being asked to design.

VIDEO ANALYSIS ' WAl l ' _

Layered documentation of ideas, issuesand opportunities for change resultingfromvideo analysis

Post-it -lH.1Wno'"

Fcamccreboard

Acetate overlays todraw on and addmore notes

FIG. 7.4. One of the techniques developed by designers to aid in vid­

eo analysis.

The team began its field work by interviewing users . These open-endedinterviews served as introd uctions to people, activities, and technologies . Dunngthese interviews users were asked to give a guided tour or walk through of theiroffices which included a description of the artifacts they used , the peo ple withwhom 'they interac ted , and the activities in which they engaged. The interv iewswere videotaped and were the basis upon which the team chose one group toobserve more closely and involve in the actual codesign effort .

BLOMB ERG. GIACOMI. MOSHER. AND SWENTON-WA l l

THE"ENYISIO NER"At the same time the team was developing new ways of representing both the

work at the sites and possible new technologies to support it. They explored theuse of concept videotapes, 3-D representations. graphical depictions. story­boards, and scenarios of use to represent possible relationships between workand emerging technology concepts . These representations aided communicatio namong team members and between them and users. One such representation tooldeveloped by the team was the Envisioner, which allowed a 3-D scaled (Y. scaleand Y, scale to show different levels of detail) model of the studied work settingsto be constructed (see Fig. 7 .5). The layout of particular offices and the equip­ment and other artifacts used were depicted with foam core pieces which hadmagnetic bases that sat on a magnetic grid and could be easily moved andrearranged. A sheet of acetate sat on top of a Plexiglass roof so designers couldmake annotations, draw connections between objects, people. technologies. andexplore design ideas . The Envisioner supported looking beyond the design of aparticular technology by making the relationship between work, people andtechnology the focus of the representation .

The Participatory Design project continues at the time of this writing with userco-design sessions being the current focus of activity. The team has agreed on atechnology design direct ion and is now exploring possible design solutions withusers. Concurrent with this activity members of the team are becoming involvedas work practice analysts and participatory design experts in some mainstreamXerox product development projects. With little time to reflect on their recentexperiences. project members are being asked to adapt their new partic ipatorydesign skills and practices to the requirements of product development . Whilethe value of closer, more extended contact with users is beginning to be under­stood outside the small participatory design project described here , how suc­cessful we will be in migrating the approach into Xerox product developmentprojects remains to be seen.

150

A 3 dimensional representationof an office layout

Foamcorepieces

Plexiglas top withacetateoverlay todraw on

7. ETHNOGRAPHIC FIELD METH ODS 151

FIG. ~.5 . Phot~g~aph of industrial design techniques adapted to workpractice analysis.

The activities of the group selected were observed and videotaped over severalweeks . By jo intly analyzmg both the interview and observational videotapes , thePD team began to build a shared understanding of the work at these sites .Drawing on skills In graphic and industrial design, the team began to construct acollage of Ideas, Issues. and opportunities for change that surfaced during thejoint VIdeo analysis (see Fig . 7.4). The collage helped the designers articulateand translate what they were learning into possible design concepts.

CONCLUSION

Linking ethnographic field methods and design has the poten tial both to providedesigners with new ways of gaining a deeper understanding of user work prac­tices and to provide a context for designers to collaborate with users over thedesign of new technologies. However, realizing the benefits of a link betweenethnography and design presents many challenges including learning how totranslate the insights from an ethnographic study into terms relevant to design,providing designers with the skills necessary to be reasonably accomplished fieldworkers , and altering the mind set of product planners and developers so thatextensive, in-depth user involvement is viewed as necessary throughout thedesign and development process. It is our hope that this paper, and the PD project

152 BLOMBERG, GIACOMI, MOSHER, AND SW ENTON-WA LL7. ETHNOGRAPH IC FIELD METHODS 153

repo rted upon here, wilJ help move us closer to a successful linkage between thetwo undertakings.

REFERENCES

Agar, M . (1980). The professional strange r, New York: Academic Press .Agar. M . (1986). Spe aking of ethnography. Qualitative Research Methods. 2. 11-78.Arensberg . C .• & Niehoff, A . (197 1). introducing social chan ge: A manualfo r community develop­

ment . Chicago: Aldine l Athenan .Atkin son , J. M ., & Heritage , J. (Eds .). (l 984). Structures a/social action: Studies in conversation

analysis . Cambridge . UK: Cam bridge University Press.Blombe rg , 1. (1987). Soci al intera ction and office comm unication : Effects on user evaluation of new

technologies: In R. Kraut (Ed.), Technology and the transfonnation of white col/ar work ( pp.195- 210). HiIldaJe, NJ: Lawrenc e Erlbaum Assoc iates.

Blomberg , J. (1988a). The variable impact of computer technologies on the organization of workacti vities. In I. Greif (Ed.), Computer-supp orted cooperative work: A book 0/ readings ( pp. 771­78 1). San Mateo , CA: Morgan Kaufman n.

Blomberg . J. (1988b , February ). Soc ial aspects of operabi lity: Ethnography cfphotocopiers. PaperPrese nted at the American Associat ion for the Advancement of Science Annua l Meeting , Boston.

Blom berg . J. (1989). Workshop on etbnographic field methods and their relation to design . Present­ed to Xerox Industri al Des ign/ Human Interface Department. Roch ester, NY.

Blomberg , J.• & Henderson . A. (1990). Reflections on part icipato ry design: Lessons from theTrilliu m community. In J. C. Chew & 1. Whiteside (Eds .). Human / actors in comp uting systems .CHI '90 Co nferenc e Proceedings (pp. 353-359). New York: ACM .

Briggs, C . ( 1986). l1aming how to ask: A sociolinguistic appraisal of the role of the interview insoc ial science research. Camb ridge , UK: Cambridge Universi ty Press .

Cicourel, A . (1982). Interviews . surveys, and the problem of eco logical validity. American So­cio logist , 17, 11-20.

Clifford , J. (1988). The predicament of culture: Twent ieth -century ethnography. literature. and art.Cambridge. MA: Harvard University Press .

Clifford, J. , & Marcus , G .B. (Ed s.). (1986). Writing culture. Berkeley: University of Californ iaPress .

Council of the American Anthropological Associ ation (1971). Statements on ethics: Principles ofprofessional responsibili ty. In J. Casse ll & S . Jacobs (Eds.}, Handbook on ethical issues inAnthropo logy ( pp. 96- 1(0). Washington, DC: American Anthropological Assoc iation .

Crane, J. , & Angros ino , G . ( 974 ). Field projects in ant hropology : A student handbook . 2ndedition. Prospec t Heights, lL: Waveland Press.

Elm, P. (1988). Work -oriented design of computer artifacts, Stockho lm: Gummessons.Geertz , C . (1973). The interpretation 0/ cultures. New York.: Basic Books .Gee rtz , C. ( 1983 ). Local know ledge : Further essays in interpretive anthropology . New York : Basic

Books.Good win, C ., & Heritage , J. (1990). Conversation analysis. Annual Review of Anthropology , 19 ,

283-307.Gre if, I. (Ed.). (1988). Computer-supported cooperative work : A boo k of read ings . San Mateo , CA:

Morgan Kaufmann.Harris, M . (1979). Cu ltural materialism: The struggle f or a science 0/culture . New York: Random

House .Heritage . J. , & Atkinson . J. M. (1984). Introduction: In 1. M. Atkinson & J. Heri tage (Eds.),

Structures of social action : Studies in conversation analysis (pp. 1-15). Cambridge , UK: Cam-

bridge University Press. .Jackson, J. (1990). " Deja en tendu" : Th e liminal qualities of anthropological fieldnotes . Journal of

Contemporary Ethnography , 19. 8-43. . . .,Jordan , B . (1987). Method in sociocultural anthropology. Course notes , Mlchl~an Sta te U~lverslty.Jordan. B., Henderson, A. , & Tatar, D. (in preparation). Interaction analysts: Foundati ons and

practice. Palo Alto: Institute for Research on Leaming and Xerox Palo Alto Research Cente~.Kluckhohn , C. (1940 ). The conceptual structure in middle American studies. The Maya and their

neighbors. New York: Appleton-Century. .Mehan, H. (1979) . Learning less ons : Social organ izat ion in the classroom . Cambndge, UK: Cam-

bridge University Press . .Mishler, E. G . (1986). Research intervi ewing: Context and narrative . Cam bridge, MA : Harvard

Univers ity Pres s. . 'Nader, L. (1974). Up the anthropologist: Perspect ives gamed by studyi ng up: In D. Hymes (Ed. ),

Reinvent ing anthropology ( pp. 284-3 11). New York: Vintage Books .Naroll , R ., & Cohen. R . (Eds .). (1970). Handbook o/ method in cultural anthropology. New York:

Columb ia University Press .Niehoff, A . (1966 ). A casebook of social change . Chi cago: A ldine .Pelto . P. J. (1970). Anthropological research . New York: Harpe r and Row.Penniman , T. K. (1974). A hundred years of anthropology . New York: Morro w.Strathem , M. (1987). The limits of auto-anthropology: In A. Jackson (Ed .), A nthropology at home

( pp. 16-37). London : Tavistock Publishers . . . . .Suchma n, L.• & Jordan, B. (1990). Validity and the collabo rative construction of mean ing Jnfa~e.

to-face survey interviews : An interaction analysis. Palo Alto: Institute for Research on Learning

Report No. IRL 90-00 19. .Sucbma n, L. , & Trigg , R. (1990 ). Unders tanding practice: Video as a medium for refl~hon ~nd

desi gn . In J. Greenbaum & M . Kyng (Eds.), Design at w.0rk: Approaches to cottaborauve design

(pp. 65- 89). Hillsdale , NJ: Lawrence Eatlbaum ASSOCiates . .Tyler, S . (1979). The said and the unsaid: Mind . meanin~ an~ culture".New York: Academic Pre ss.Van Maanen, J. (1988). Tales 0/ the field . Chicago: Univers ity of Chicago Press .Whiting, B ., & Wh iting, 1. (197 0). Methods for observi ng and recording behavior: In R. Naroll ~

R . Cohen (Eds .) , Handbook of method in cultural anthropology ( pp. 282-315). New York.

Columbia Univers ity Press .Wolcott , H . F. (1990). Makin g a study " more ethnographic. " Journal o/ Conte mpo rary Ethnogra-

phy, 19, 44-72.

APPENDIX 1: FIELD EXERCISES

The following is a list of field exercises developed for a workshop on Eth­nographic Field Methods. Although these exercises are most profitably employedin conjunction with a general discussion of ethnogra phy and ethnographI c .fieldmethods, they are presented here to provide some ideas for practical acnvmes In

which people interested in developing skills in ethnography might engage (fromBlomberg , 1989: adapted from Cra ne & Angro sino (1974) and Jordan ( 1987).

I . Through observation and careful note taking investigate how space isemployed in some area of this building. Explore the extent to which the area you

154 BLOMBERG, GIACOMI , MOSHER, AND SWENTON-WALL7. ETHNOGRAPHIC FIELD METHODS 155

study and the furniture within it tend to keep people apart or to draw themtogether. How does the arrangement of walls , doorways, hallways, furni ture,machines , etc. effect the patterns of interaction you observe?

. 2. Cons~ct a map of an office area . Your map should show where peopleSit, thCLC relationship to one another, the activities that go on at various loca tionswhere " significant" objects are located . Describe how you gathered the informa­non represented on your map (through observation, interview, available docu­ments, etc). You' ll probably want to include a key or legend to help othersInterpret your map.

3. Find a specia list (accountant, machine operator, drafts person, etc.) whowould be willing to talk to you about the work they do. Ask them to show youhow they accomplish some aspect of their work (complete a transfer of fundsprepare and distribute a memo for a group meeting, create a drawing). On thebasis of your observati ons and questioning of the specialist, prepare a descriptionof how the activity IS accompli shed, the materials used , the people consulted ,etc. so that someone who is completely unfamiliar with the activ ity would beable to get a good idea of what's involved . The object of this exercise is tosharpen your powers of observation and description, and to develop your abilityt~ askmore and moreprecisequestionsof yoursubject. Audiorecordthe interac-non. .

4. Collect the work life history of someone working in your organization.Collec tmg hfe histories requires that the interviewer be as nondi rective as pos­SIble, interrupting as little as possible, and allowing the person prov iding theinformati on to decide what 's important to tell . You might want to start by askinga nondirective question like, " Tell me about the jobs you 've had ?" If the conver­sation fal ters, you might ask about specific jobs, what was involved in each, whythey moved on, what they learned at each position, what they hope to be doing inthe future etc . AudIO record the interaction so that you are able maintain anappropriate interactio nal relationship with the. subject.

5 . Do a semantic analys is of the kinds of documents found in someone 'soffice , or the kinds of telephone calls they make or receive. Ask them to give youthe name of each type of document found in their office or the calls they make orrecei ve. If they don 't have a name for some category of documents or calls, ju stmclude. a desc ription of the category. Be careful not to impose you're owncategones and names . Ask if there are sub categories within part icular classes( personal cal l of less than 2 minutes , routine requests , etc.) Ask them thecharacteristicsof each document or call included in a particu lar category and askthem to explain how these characteristics are used to distinguish one type ofdocument from another. If possible, locate actual documents or inquire aboutactual calls made or received while you were present and ask which categoryeach falls into and why. Their idealized classification system may need to beadjusted to accommoda te actual instances . If you have time you might want to

repeat the procedure with another informant and then compare their class ificati onsystems.

6 . This exercise should be coo rdinated with exercise 7 . Locate a publi c areawhere a video camera can be set up. Decide on the best location for the camera.Make note of the issues you considered when deciding where to locate thecamera (facial expressions captured, lighting, movement in and out of the space ,close-up shots of buttons pushed etc.). Record no more than 30 minutes of theactivity at the chosen location . While the camera is running, observe what'sgo ing and take notes , paying particular attention to things the camera is unable torecord.

7. In the same location selected for exercise 6 , take 10 photographs. Choosethe shots so that they are representative of the activities and the physical arrange­ment of objects relative to one another. Attempt to capture interaction and thephysical and social context of the activity. Describe eac h photograph (when itwas taken , what it represents , its relationship to the other photograph s, etc .).Together with the person doing exercise 6, compare what you were able tocapture with the video camera, your observations , and the still photos.

8 . Locate an area where you can unobtrusively observe some activity. Ob­serve the activity for 30 minutes or so. Take detailed notes of the what's goingon, including a sketch of the area , time of day, movement in and out of the scene ,participants' relation to one another. You might want to devel op a table tofacilitate recording your observations. Review your notes and prepare a descrip­tion of what you observed .

9. Interview someone with the goal of developing a description of their" social network." Describe the working relationship (lines of authority, rela­tionship between tasks each engages in, etc .) that exi sts among co -workers .Represent this information in an "organization chart" where the nature of therelationship between co-workers is represented.

10. Go to a public area where people meet (cafeteria, coffee room, etc.).Select a small group of peop le to observe (2-6). Based only on your powers ofobservation try and determine who these people are (age , occ upation , education,etc .), their relationship to one another, what they're talking about, etc . Afterobserving them for a few minutes and writing down your hunches, interviewthem and find out to what extent your impressions were related to what youlearned from interviewing . Reflect on how you misread certain cues , the infor­mation unobtainable through observation alone , etc .

I I . Interview someone about the " life history " of some small co llection ofdocuments (i .e . , the ones currently on their desk or in their " in" basket}. Talkwith other people involved in creating , processing , or acting upon the docu­ments . Prepare a description of each document's life history.