HIGHER RATES OF PARTICIPATIONpalys/ResDec4_Ch6_InterviewOralHist-QualExcerpt.pdf · Merton, Fiske, and Kendall (1956) originally coined the term “focus group” (see Fontana & Frey

researcher’s research objectives, theoretical interests,and epistemological stance.

H IG H E R RATE S OF PARTICI PATION Many of thedisadvantages of questionnaires are handledadmirably by the interview. Participation ratesamong people approached for a face-to-face inter-view are often around 80 percent or even 90 per-cent; this is comparable to some self- andgroup-administered questionnaires and consider-ably better than mail-out questionnaires. Thus, vol-unteer bias is generally less of a problem withinterviews.

B E N E FITS OF CONTACT The interaction of inter-viewer and respondent also offers benefits that canenhance the quality of the data gathered. The inter-viewer can ensure that the appropriate person com-pletes the interview, can clarify immediately anyconfusion about particular questions, and canencourage verbally stingy respondents to embel-lish further. Also, since the interviewer asks ques-tions and writes down responses, the respondentneedn’t be literate. And although some participantsmay feel less anonymity in the personalized inter-view setting than with the impersonal question-naire, skilled interviewers can often build sufficientrapport to alleviate such misgivings. Finally, therapport that’s built may have longer-term benefitsfor researchers engaging in longitudinal research,since respondents may be more willing to partici-pate in panel studies involving repeated inter-views.

TI M E AN D COST The biggest disadvantages of face-to-face interviews are their cost and the timerequired to complete a large-scale interview study.While it wouldn’t be unusual for a questionnairestudy to cost $4 or $5 per participant or for a tele-phone interview study to cost $20 to $25 per par-ticipant, conducting the same study by face-to-faceinterview might well cost $50 or even $100 ormore per participant. But comparing these esti-mates may be deceptive, since they reflect alterna-tives that aren’t equally likely to occur. Asquantitative researchers seeking large data sets have

leaned more and more toward using telephone sur-veys, the realm of the face-to-face interview has, forthe most part, been vacated to more qualitativelyoriented researchers, who are more likely to do theinterviewing themselves than to hire teams ofresearch assistants to amass huge sets of aggregatedata.

W H E R E TH E ACT ION I S This very shift tells ussomething about how the process of face-to-faceinterviewing is conceived differently from ques-tionnaires or telephone surveys. Large-scale quan-titative researchers place great emphasis on thedesign phase of survey questionnaires; the “highestpriced talent” designs the survey instrument, whilethe survey’s actual execution is often relegated to the“lowest priced talent,” that is, part-time assistantswho often have little commitment to the data andmay not even know who’s sponsoring the study orwhat its purpose is (beyond what can be deter-mined from the superficial content). In contrast,more qualitatively oriented researchers typicallyvalue the process of data gathering as much as thedesign of the research instrument itself. As a result,both the study’s design and its execution are under-taken by the “highest priced talent.”

Embedded here are two very different views ofthe data-gathering process. Recall that among quan-titative researchers, the role of the researcher-as-expert is to determine what is important to ask; afterthat, anyone can gather data, as long as he or she isproperly trained to be neutral, aloof, and standard-ized in delivery. Qualitative researchers view thedata-gathering process itself as informative, main-taining that one must be open to any new directionsthat may emerge in the context of the interviewbecause of the unique perspective of the partici-pant(s). Doing so effectively requires people whoare familiar with the participants and with the phe-nomenon under study, and who clearly understandthe research objectives so that they can make respon-sible decisions about what to do in the unique situ-ations that emerge with every interview.

Because of other preferences held by those whoengage in qualitative study (e.g., a phenomenological

C H A P T E R 6 – I N T E R ACT I V E M E T H O D S : S U R V E YS , I N T E R V I E WS , A N D O R A L H I STO RY T E C H N I Q U E S 1 5 7

NEL

perspective, an inductive approach, an emphasis oncase study analysis in context), face-to-face inter-views look very different from telephone surveys ormail-out surveys simply because the qualitativeresearcher is much less likely to be trying to amasshuge sets of standardized data for nomotheticanalysis. Face-to-face interviews tend to be longerand more detailed, tend to seek greater depth ofresponse, and tend to be more open-ended in theirconstruction to allow for phenomenological inputfrom respondents. They’re also more likely to be sit-uated in some context (e.g., an organization, agroup, a limited geographical setting such as a neigh-bourhood) that plays an important role in theanalysis.

H U MAN IZ I NG TH E PROCE SS This more intimateconnection between the researcher and respondentmay have many benefits associated with it, such asgreater likelihood of developing rapport, but theinterview’s interactive nature also means that onemust be more careful about reactive bias.Interviewees can be very attentive to cues that theinterviewer emits, since they want to know whetherthey are “doing well” as participants. Thus, whatyou choose to write down out of their verbalresponses and even your supportive and encouraging“uh-huhs” or nods of the head may be taken as cuesabout what the interviewee “should” be talkingabout. One must avoid leading the interviewee.

Finally, although a sensitive interviewer may reas-sure the respondent about confidentiality, interviewsclearly generate less of a feeling of anonymity thando impersonal questionnaires. Considerable effortmust therefore be made to ensure that rapport iscreated and that the interviewee legitimately believesthere is no reason to feel threatened. We say “legiti-mately” here on the assumption that that is true.Obviously, where possibilities for repercussionsexist—whether to the individual or to his or hergroup—because of the results of the study, ethicalpractice requires that respondents be informed ofthat possibility and that researchers build in safe-guards to the extent possible.DEALI NG W ITH TH E DATA Lastly, the interviewermust make some choices about how to retain

responses, and each choice has its advantages anddisadvantages. If the choice is to write down ver-batim responses to questions, the process canbecome quite tedious and may well interrupt theflow of the interview while the respondent waits forthe interviewer to finish writing down each response.The researcher may therefore choose to write downonly summaries or major points from the response,but then distortion may occur, or matters that aresubsequently found to be important may be left out.

Another possibility is to tape (using audio or videorecording) the interview. This approach frees theinterviewer to pay attention to the interviewee,although some would advise that the interviewershould occasionally jot down notes in any event,because doing so helps the interviewer retain theflow of the interview, because most respondentsexpect you to write something down every so often,and because notes give you some backup in case atechnical foul-up renders the tape useless and youmust regenerate the content of the interview frommemory. On the other hand, transcribing taperesponses is expensive, and the permanence andunforgiving accuracy of tape may inhibit candour.

FOCUS GROUP INTERVIEWS

The focus group interview is essentially a group ver-sion of the face-to-face interview. Focus groups havean extensive history in marketing research, but haveonly more recently been discovered by the broadersocial scientific community (D. L. Morgan 1988;Morgan & Spanish 1984). Such groups normallyinvolve a target sample or purposive sample of inform-ants brought together to discuss the phenomenon inwhich the researcher is interested. As Fontana andFrey (1994) note, “The group interview is not meantto replace individual interviewing, but it is an optionthat deserves consideration because it can provideanother level of data gathering or a perspective onthe research problem not available through indi-vidual interviews” (364).

In marketing, a group of typical consumers mightbe brought together to discuss their preferences,what they value in an existing product, or what theymight like to see in a developing product. General

C H A P T E R 6 – I N T E R ACT I V E M E T H O D S : S U R V E YS , I N T E R V I E WS , A N D O R A L H I STO RY T E C H N I Q U E S1 5 8

NEL

Motors, for example, might be interested indesigning a new truck and might hire a researchagency to bring together a group of truck owners todiscuss what features they believe should be includedin the design. Marketing researchers also have usedfocus groups to assess other products and their pack-aging, such as politicians.

In the social sciences, however, the “product” theresearcher wants to develop may be a questionnaireor a completed study. The participants would beinvited because of their relevance to the phenom-enon of interest to the researcher. An early exampleof this (cited by D. L. Morgan 1988) would beMerton and Kendall’s (1946) effort to evaluate thepersuasiveness of wartime propaganda; morerecently, Morgan and Spanish (1985) created focusgroups to generate discussions concerning percep-tions of risk factors involved in heart attacks, and D.L. Morgan (1986) brought together focus groups ofwidows as part of his investigation into bereave-ment. One can envision innumerable other uses,including government administrators discussingorganizational processes or policy objectives, filmdirectors discussing how they reconcile their ownbeliefs with the market-oriented pressures of fundingagencies, and groups of parents discussing educationpolicies in relation to their children’s schools.

U N IQU E ADVANTAG E S Such groups may serve sev-eral purposes for researchers. D. L. Morgan (1988)explains that focus groups can be used productivelywhen

orienting oneself to a new field; generatinghypotheses based on informants’ insights; evalu-ating different research sites or study populations;developing interview schedules and question-naires; [and] getting participants’ interpretationsof results from earlier studies. (11)

Focus groups also may provide provocativeand/or insightful information to the exploratoryresearcher who is looking for unanticipated conse-quences to organizational interventions; is interestedin determining issues of importance to those in theresearch setting or in acquiring new insights about

the phenomenon from those who have experiencedit; and/or is trying to develop research instruments(e.g., questionnaires, interview schedules, samplingstrategies) that have integrity with respect to thephenomenology of those under study.

But the usefulness of focus groups is not limitedto exploratory research. After a piece of research iscompleted, for example, and the researcher has ana-lyzed and interpreted the data, this informationmight be imparted to additional focus groups fordiscussion. Such discussions may help the researchergather alternative interpretations for further consid-eration and generate additional hypotheses and/orresearch questions on which to focus subsequentresearch. Indeed, it was for exactly this purpose thatMerton, Fiske, and Kendall (1956) originally coinedthe term “focus group” (see Fontana & Frey 1994).

Although focus groups have much in commonwith the traditional in-person interview, D. L.Morgan (1988) argues that their inherent socialdynamic gives them at least two unique advantages.First, instead of simply taking an inventory ofopinion through individual interviews, the focusgroup setting places opinions “on the table” wheredifferences between perspectives can be highlightedand negotiated. This process allows participants toembellish on positions, discuss related dynamics,and articulate the rationale(s) underlying their per-spective. Blumer (1969) advocates identifying asmall number of informed participants who areacute observers in any social setting of interest, andthen states that “a small number of such individualsbrought together as a discussion and resource groupis more valuable many times over than any represen-tative sample” (41). D. L. Morgan (1988) adds thata second major advantage of focus groups is theopportunity to “witness” (as opposed to “influence”)extensive interaction on a topic within a relativelylimited time frame.

AN D U N IQU E COM PLICATION S, TOO While someadvantages accrue from the social composition offocus groups, this characteristic can also pose prob-lems. Some people will be less comfortable thanothers in expressing their opinions publicly; peoplewith more extreme or unique views may be reluctant


NEL

to expose them to possible ridicule; and people willundoubtedly be more concerned about maintainingtheir image in a public setting than in a one-to-oneinterview. Fontana and Frey (1994) reaffirm Mertonand colleagues’ (1956) view that three skills are par-ticularly important in the group interviewer’s reper-toire:

First, the interviewer must keep one person or asmall coalition of persons from dominating thegroup; second, he or she must encourage recalci-trant respondents to participate; and third, he orshe must obtain responses from the entire groupto ensure the fullest possible coverage of the topic.(365)

Thus, focus groups may be seen as a usefulvehicle for encouraging the embellishment andnegotiation of public opinion, while the traditionalsingle-person interview or questionnaire acts as acomplementary expression of privately held opin-ions or “secret ballots.”

The focus group also accords a less central but noless important role to the researcher. In an interviewinvolving one researcher and one participant, therespondent can look only to the researcher for direc-tion. The researcher in the focus group setting gen-erally plays a more facilitative and less directive role.Although the researcher can set up a structured sit-uation, he or she typically acts only to initiate,prompt, and referee the discussion. Accordingly, onemight infer that the results will be influenced moreby the group than by the researcher.

In sum, focus group interviews are compatiblewith an array of research objectives, and while theypossess their own limitations, they also offer uniquestrengths. They thus seem a currently under-utilizedaddition to the social science researcher’s proceduralrepertoire.

ORAL HISTORIES

There are two types of oral history, only one ofwhich will be considered in detail here. Be sure youunderstand the difference between them. The onethat receives only passing attention in this book is

the type of oral history remembered and practised byAboriginal peoples in North America.

A B O R I G I N A L O R A L H I S TO R I E S Until relativelyrecently, most North American Aboriginal cultureswere primarily oral cultures. Consistent with thisemphasis, Aboriginal peoples made extensive use oforal history, where the role of particular individualswould be to remember certain stories about theirpeople’s history, rather like walking archives.

These memories weren’t merely the recollectionof stories, as might have happened when your par-ents told you a story when you were a child, but werein fact “lived memorizations” and verbatim accountsthat would be repeated in the same manner 50 yearsfrom now as they would be today. Indeed, contem-porary oral histories are often found to be identicalto those recorded by anthropologists at the turn ofthe 20th century (e.g., see Mills 1994). Part of theirintegrity came from the fact that one of the “jobs”of each new generation required accurately learningand remembering the stories handed down by pre-vious generations.

Another part of their integrity came throughtheir having survived the rigorous process of ongoingchallenges to their accuracy. Among the Aboriginalpeoples of the northwest coast, for example, thisprocess occurred in the context of the feast (pot-latch) system. On appropriate occasions, eachspeaker recounted, in this most public of settings,the history of his or her clan, the boundaries of theclan’s territories, and the way its crests and songshad been acquired. Anyone at the feast could chal-lenge this oral history; the lack of challenge signalledacceptance that the account was valid (e.g., seeGisday Wa & Delgam Uukw 1992; Mills 1994).

We make these points for two main reasons. First,we want to encourage respect for the oral historiesof Aboriginal peoples. Because of Europeans’ andother non-Aboriginal peoples’ reliance on writtendocumentation and reverence for materials inwritten archives, many Europeans and non-Aborig-inals have ethnocentrically assumed that no docu-mentary history meant “no history” (e.g., see Wolf1982), and that oral histories are little more than


NEL

some sort of quaint, ever-changing cultural fairytale. This is far from the truth. Yet such views havebeen used by colonial powers to dismiss Aboriginalhistories and thereby deny Aboriginal rights.1

Second, we’ve made this point to put the oral histo-ries we’ll deal with in this chapter into perspective.

O R A L H I S TO RY I N S O C I A L S C I E N C E Broadlydefined, “history” is everything that happenedbefore you read this sentence. And now even thatsentence has receded into history. We can neverknow everything about history, but that hasn’tstopped historians and the rest of us who are inter-ested in history from trying to understand it. Intrying to do so, we realize one of the challenges ofunderstanding history: it isn’t here anymore. Wethus cannot study history directly, but must do soby looking at those pieces and remnants of historythat remain.

WHAT’S I N TH E BOX? Now, let’s imagine that thehistory we can study is all contained in a huge box.Of all that happened in that huge period of time weknow as “history,” the only things we can base ourstudy of history on are the things inside that box,because those are the only things that remain. Andwhile many things get inside the box, many thingsdo not. For example, last weekend, Ted’s son Felixpitched an absolutely great game against the bestteam in his Little League; it was the only game thatteam lost all year. Felix was extremely happy aboutit, and Ted was glad he was there to see Felix play.It was an enjoyable moment for both of them asfather and son.

But 100 years from now, when some historian sitsdown to write something about early 21st centuryhumanity—even if he or she is writing about“father–son relationships in the early 21st cen-tury”—there isn’t a chance in the world that thisfuture historian will write about the day Ted’s sonpitched such a great game and Ted was there towatch. Why? Because that bit of human history,while as real as the fact that Ottawa is the capital ofCanada, will never make it into the box of humanevents out of which future historians will manufac-ture history.

Or will it? Ironically, our describing that experi-ence here makes it possible that some historian in2107 actually will see some dilapidated old copy ofthis book and discover the fact about Felix and hisdad and Felix’s great day of pitching. Because we’veput it in writing, a fact that would otherwise berecalled by no one besides Felix and his dad (neitherof whom will be around by then) is now part of thecontents of the box of study-able human history. It’sactually a very interesting example of selectivedeposit, a phenomenon we’ll discuss in more detailin Chapter 8. This term reflects the recognition thatsome things have a higher likelihood of being putinto the box than others, and that some people andgroups have better access to the box than others do.It’s interesting, for example, that, in Ted’s role asdad, he has very little likelihood of accessing thebox. No one outside his immediate family will prob-ably ever have any sense of him as a parent. But inhis roles as university professor and author, he hassomewhat more access to the box, as evidenced bythis book, which is now part of the written histor-ical record.

One of the tragedies of history is that so muchthat would be interesting to know will remain for-ever beyond our grasp because it was never placedinside the box. It’s interesting to consider what sortsof biases have entered into that process. What peopleor groups have been systematically less likely to havehad a chance to put something in the box? And whatpeople or groups have had much better access to thebox, allowing them to influence our sense of historyby placing their experiences into the box?

Clearly, some people have had better access to thebox than others. Governments, the rich, the pow-erful, the upper classes, and the educated have allhad better access to the box than individual citizens,the poor, the vulnerable, the lower classes, and theilliterate. Similarly, because of the history of patri-archy in many non-Aboriginal traditions,2 it is alsothe case that men—because until relatively recenttimes, it was primarily they who formed the govern-ments, controlled the property and wealth, and hadbetter access to education—have had better access tothe box than women. When historians open the box


NEL

to try to understand history, the “facts” they look atare therefore not all the facts, or even a representa-tive sample of facts, but only the facts placed thereby those who had access to the box. So when we tryto look at what life was like in, say, 17th-centuryEngland or 15th-century Spain, we’re relying mosttypically on the views of the rich, the powerful, andthe educated and on the views of men. Even whenwe find material about others whose experience wealso would like to understand—the daily life of theaverage 15th-century Spanish peasant, say—it’srarely through their eyes that we see the worldaround them, but rather through the eyes of non-peasants who had access to the box.

R ECTI F Y I NG TH E I M BALANCE OF W R IT TE N

H I STORY Oral history is partly a way of trying todeal with the problems of access just outlined. Itrecognizes and to some extent shares the generalEuropean bias in favour of written documentation,and therefore tries to get material into the box thatwouldn’t otherwise be there. Oral history is conse-quently seen by many as “an interview techniquewith a mission.” Fontana and Frey (1994), forexample, note that “oral history does not differ fromthe unstructured interview methodologically, but inpurpose” (368), where the purpose is to take mate-rial that otherwise might have been forgotten andmake it part of the written record. Reinharz (1992)adds that “oral history … is useful for getting atpeople less likely to be engaged in creating writtenrecords and for creating historical accounts of phe-nomena less likely to have produced archival mate-rial” (131; emphasis in original). By interviewingpeople about their past, we “recover” parts of historythat might otherwise have been lost; by interviewingpeople about their present, we help ensure that theirrecord is available for future generations.

Although examples of collected oral history nar-ratives go back to antiquity, “its modern formalorganization can be traced to 1948, when AllanNevins began the Oral History Project at ColumbiaUniversity” (Starr 1984: 4, cited by Fontana & Frey1994: 368). That quotation, of course, has a certaindelicious irony, since it’s another example of how

people with access to the box of history (like aca-demics at prestigious universities) are the ones whosecontributions we remember and can cite becausethey’re part of the written record.

Because of the nature of the mission associatedwith oral history, you shouldn’t be surprised to dis-cover that the technique has been particularly pop-ular among people who are on the margins ofsociety—minorities, the poor, street people, andwomen, for example—and/or among those who areinterested in engaging in research with such people.Oral history narratives exist en masse for many of the“common people” of history whose experiencewould otherwise be ignored or forgotten. Examplesinclude collections in which people talk about theirworking lives (e.g., Terkel 1975), as well as morespecific projects that focus on Pennsylvania steel-workers and their families, women working in Bal-timore canneries (e.g., Olson & Shopes 1991), theexperience of Blacks in the Vietnam War (Terry1984), Palestinian women engaged in resistanceactivities (e.g., Gluck 1991), or on a staggering arrayof other groups. No doubt many others haven’t yetseen the light of day; as Fontana and Frey (1994)note, “often, oral history transcripts are not pub-lished but may be found in libraries, silent memoirsawaiting someone to rummage through them andbring their testimony to life” (368).

Recently, oral history methods have found partic-ular favour among many feminist researchers, whosee oral history methods as a way to rectify thegender imbalance in the largely male-dominateddocumentary archives of history. Gluck and Patai(1991) note that

The first major body of literature on women’soral history appeared in late 1977 in a specialissue of Frontiers: A Journal of Women’s Studies.This ground-breaking issue served as the key ref-erence on women’s oral history for many years,and the suggested outlines for women’s oral his-tory interviews that appeared at the back of thejournal were xeroxed, dittoed, and mimeographedby women in communities and classroomsaround the country. (4)


NEL

Emphasizing the gathering of women’s oral his-tories is thus a way to include women’s voices in his-tory: “Refusing to be rendered historically voicelessany longer, women are creating a new history—using our own voices and experiences” (Gluck 1984:222). Reinharz (1992) suggests that women’s oralhistory actually serves a threefold function: drawingwomen out of obscurity, repairing the historicalrecord, and providing stories of people with whomwomen readers and authors can identify.

QU E STION S ABOUT R E S EARCH R E LATION S H I PS

While feminist engagement in oral history is thusrewarding in its own right, involvement in thismassive oral history project has also prompted fem-inists to lead the way in contemporary reconsider-ations of oral history in particular, interviewing ingeneral, and, even more broadly, the whole set ofrelations between respondents and researchers. Theconcern is not only that men’s voices dominate his-tory, since this imbalance could be addressedsimply by using techniques like oral history to rec-tify it, but also that the methods available to usnow, since they were conceived in a traditionallymale-dominated social science, are often particu-larly “male” in the way they’re conceived, designed,and executed. Many feminist researchers who tookthe plunge into oral history research found thatstraight application of the methods they’d beentaught in graduate school needed reconsideration.

Minister (1991), for example, observes that

the male subcommunication subculture isassumed to be the norm for social science inter-viewing ... If women aspire to become approvedoral historians, they must learn to control topicselection with questions, must make certain thatone person talks at a time, and must encouragenarrators to “take the floor” with referential lan-guage that keeps within the boundaries of selectedtopics. (31)

Minister clearly doesn’t intend to encourage womento aspire to those essentials. Instead, she argues thatwomen must carve out their own version of oral his-

tory, a version that’s more sensitive to, and a betterreflection of, women’s ways of communicating.

We will not hear what women deem essential totheir lives unless we legitimate a female socio-communication context for the oral history situ-ation. As Sue Armitage says, “We will learn whatwe want to know only by listening to people whoare not accustomed to talking.” (31–32)

Anderson and Jack (1991) agree that morefemale-reflective approaches are required. Theybelieve, for example, that men tend to look at theinterview purely as an information-gathering ses-sion, so that designing an interview study becomesa strategic question of how best to order and com-pile questions and answers. In contrast, they believewomen are more attuned to relationships andprocess and that woman-to-woman interviews mustreflect that difference:

Realizing the possibilities of the oral history inter-view demands a shift in methodology from infor-mation gathering, where the focus is on the rightquestions, to interaction, where the focus is onprocess, on the dynamic unfolding of the subject’sviewpoint. (23)

A major problem crops up when interviewingwomen, according to Anderson and Jack (1991).Because men’s experience has defined much of con-temporary existence, women have become used totalking in dual narratives: using concepts that reflectmen’s cultural domination, but focusing on theirown experience, which may or may not be ade-quately captured by those schemas. Feminist oralhistory researchers must therefore be particularlysensitive to reading between the lines:

We need to hear what women implied, suggested,and started to say but didn’t. We need to interprettheir pauses and, when it happens, their unwill-ingness or inability to respond. We need to con-sider carefully whether our interviews create acontext in which women feel comfortable


NEL

exploring the subjective feelings that givemeaning to actions, things, and events, whetherthey allow women to explore “unwomanly” feel-ings and behaviours, and whether they encouragewomen to explain what they mean in their ownterms. (17)

This process is clearly an interpretive one. How-ever understandable the position, it’s also fraughtwith complexity and paradox. At issue here is thequestion of “voice” and, especially, whose voice (ifeither) dominates the final look of the text that’sproduced from the interview. An excellent self-crit-ical analysis by Borland (1991) of an oral historyinterview she did with her grandmother highlightsthe dilemma well. The article focuses on an experi-ence for which Borland’s interpretations of “whatwas really going on” and “what was really being said”were completely different from her grandmother’s.Rejection of the interpretation by her grandmotherafter reading the first draft led to further discussion,further revising, and some movement on the parts ofboth women as they tried to reach jointly satisfactoryresolution of the meaning of the original episode.

One of the central tenets of feminist research isthat women must be able to say things in their ownvoice and that voice must be heard. Male-dominatedscience is seen as an inappropriate model to theextent that it embodies hierarchical relationsbetween the researcher and those researched, wherean “expert” researcher “extracts” data from “subjects”and then reinterprets it according to a “culturallysanitized” (i.e., male-dominated) “spin.” Instead oftreating each other like researcher and respondent,many feminists aim to do research in the same egal-itarian and respectful manner in which one mightinterview one’s friend or grandmother. Andersonand Jack’s (1991) assertions seem to question thatsort of face-value acceptance, seeing the task as oneof reading between the lines and finding what thewomen being interviewed are “really” saying andwhat they “really” mean by it.

Many feminist researchers (including Borland1991, as noted) have wrestled with these issues.Their analyses can make us all aware of how our pro-

fessional zeal may lead us to usurp others’ voices andto take for granted some things that perhapsshouldn’t be accepted. Black feminists, for example,argue that sisterhood has its limits if it meanshomogenizing women’s experience in a way thatdoesn’t do justice to the equally meaningful andsimultaneously marginalizing experience of race(e.g., see Collins 1991; Etter-Lewis 1991; Fine 1994;Olesen 1994), while others make the same pointregarding Latina (e.g., see Benmayor 1991), Aborig-inal (e.g., see Greschner 1992; Monture-Okanee1993; Petersen 1994), and Third World women(e.g., see Hale 1991; Patai 1991; Salazar 1991). Allcall for research considerations that acknowledgeand respect cultural differences

Computer-Assisted Social Research(CASR)

The Internet is expanding at an unprecedented rate;its growth has eclipsed all other technologies pre-ceding it (Dahlen 2002). Between 1994 and 199850 million people logged on to the Internet world-wide; it took 38 years for radio and 13 years for tel-evision to acquire the same user base (UnitedNations 2004). It is clear that the Internet is thefastest growing information and communicationsmedium to date, and it will not be long before it isas common as the television or the telephone(Dahlen 2002). At present, Internet traffic doublesevery 100 days; in December 2005 one billionpeople worldwide were online (United Nations,2004).

Prior to the popularization of the Internet,Kiesler and Sproull (1986) provided a roadmap forusing networked computers to conduct social sci-ence research. While they were optimistic about thepotential of computer technology, they felt that“until such time as computers and networks spreadthroughout society, the electronic survey will prob-ably be infeasible” (p. 403). The available statisticson the growth of computer technologies such as theInternet provide ample evidence that the techno-logical revolution that we have witnessed over thepast 20 years makes it impossible to ignore the


NEL

research possibilities envisioned by Kiesler andSproull.

PRACTICAL BENEFITS AND BURDENS

CASR brings with it a number of practical benefitsand burdens. One of the most cited is the effect thatimplementing a computer-assisted design has on thespeed and duration of the research process; CASRcan be much faster than comparable traditionaldesigns. In network environments the footwork ofthe design and administration process is done bynetwork connections. Research teams can create,edit, and finalize the research instrument withoutthe burden of scheduling and attending physicalmeetings, and the research team can administer thedesign without having to physically connect with theparticipant.

Once in motion, CASR allows researchers tomove from observation to analysis much morequickly than conventional research designs. Finally,when research is conducted over wide area networkssuch as the Internet, observations can be made anddata collected 24 hours a day, 7 days a week. Thisbeing said, in some cases the initial design andadministration of the research can be prohibitivelytime-consuming. When the research team is inexpe-rienced with the use of technology or the technolog-ical infrastructure for the research is not already inplace, extra training and the installation and testingof hardware and software may be necessary. Further-more, researchers who are new to the technologyare more likely to make errors during the adminis-tration and observation stages that can result in evengreater time delays in the research.

When it comes to the cost of materials and labourthere are several distinct differences between CASRand conventional research methods. The hardware,software, and scripting that is required for the obser-vation and data collection portion of a CASR projectcan cost researchers thousands of dollars. However,researchers can offset the software and scripting costsby using freely available open-source software andscripts instead of high-priced commercial applica-tions. Furthermore, while equipment and designcosts can be high, these costs are generally recouped

through savings on paper, postage, transcription,mileage, lodging, the renting of research venues, andrepeated research. While it is uncommon to findhard-to-estimate human labour costs factored intodiscussions of many traditional data collectionmethods, the introduction of computer program-mers and graphic and web designers into the CASRdesign process has made many social scienceresearchers begin to account for the cost of labour.The hourly rate that most programmers anddesigners charge can reach as high as $120. Theresult is that in some situations CASR can be quitea cost-effective solution for the North Americanresearcher. This advantage does however depend onresearchers having enough experience with tech-nology that they can implement solutions thatrequire specialized user and programming skills.

It is important to recognize that there can beconsiderable start-up material and personnel costsassociated with CASR, but repeated research is lesscostly since much of the investment is saddled by thefirst project and the cost of upgrading vital researchmaterials is much less than first-time creation expen-ditures. Moreover, researchers should keep in mindthat some research designs are more cost-effectivethan others. For example, the simple e-mail–basedquestionnaire, where questions are placed in thebody of the e-mail, costs almost nothing to admin-ister, while complex computer-assisted experimentaland quasi-experimental designs that require specialhardware and software can cost tens of thousands ofdollars.

In addition to considerations of time and money,many CASR designs also dramatically alter the phys-ical research environment. Research such as experi-mental, focus group, personal interview, orethnography that would normally require a physicalspace can be conducted in a networked environment(Birnbaum & Wakcher 2002; Buchanan & Smith1999; Evans et al. 2003; Karr 2000). Additionally,with the increasing availability and affordability ofhandheld and laptop computers and wireless net-works, researchers are not confined to any one phys-ical locale in order to collect and analyze data.However, problems associated with powering the


NEL

computer can limit how long a researcher can stayout in the field. This is particularly challenging forresearchers conducting research in remote settings.Moreover, while the introduction of computer tech-nology into the research setting has enhanced ourability to use highly controlled experimental designsthat can be reproduced with ease and precision (Karr2000), the inability for the researcher to be physicallypresent in some virtual experimental settings can bequite prohibitive.

For research designs that are not dependent uponthe physical presence of a researcher, network-basedCASR facilitates the solicitation and recruitment oflarge, geographically and demographically diversesamples (Fox et al. 2003; Gosling et al. 2004). Theselarger samples make it possible to amass largeamounts of data in a relatively short time.

The benefits of bypassing physical and geo-graphic limitations extend beyond the acquisition ofresearch participants to the coordination of theresearch process itself. Team meetings and researchcoordination are much more efficient as investiga-tors are able to conduct virtual as opposed to phys-ical project meetings where they can communicatetheir progress and observations or make changes tothe research design without introducing con-founding variables to the procedure.

BOURGEONING POSSIBILITIES

Although using computers and network technologyto collect data and make observations is clearly notappropriate for all social science research situations,there are many examples of situations where imple-menting technology is a viable methodologicaloption. In primary data collection the goal is to col-lect original information from an identifiable pop-ulation; techniques used include self-administeredquestionnaires, face-to-face interviews, telephoneinterviews, panel studies, and focus group designs.All of these conventional approaches to the collec-tion of primary social science data can be adapted toCASR designs. In fact, many of these methodolog-ical approaches have been adapted by socialresearchers.

The computer-assisted self-administered ques-tionnaire (CASQ) comes in three main varieties—disk-based, e-mail, and web-based (Truell et al.2002). Like its traditional counterpart, this mode ofdata collection is perhaps the most widely usedwithin the social science research community. Theoldest and most basic form of CASQ is disk-basedadministration. With this technique participants aregiven a floppy disk, CD, or portable drive that con-tains a questionnaire that can be completed on apersonal computer (PC). Responses are either storedback on the disk or they are sent automatically acrossa network. With the increased availability of theWorld Wide Web (WWW) and e-mail, this mode ofCASQ is not used as often; however, someresearchers have found it a useful option for targetedpopulations (Hampton & Wellman 1999).

E-mail is one of the simplest and most used net-work information and communication components.Eighty-nine percent of networked communicationsare accomplished via e-mail (Jackson & De Cormier1999). As a result, e-mail is seen by many as the mosteffective mode of CASQ (Hessler et al. 2003;Sheehan & Hoy 1999). E-mail questionnaires canbe divided into three distinct types: simple e-mailmessages with questions in the body of the e-mailtext (either plain or XHTML formatted text), sepa-rate documents that are attached to an e-mail, andembedded URLs in the body of an e-mail messagethat direct the user to a web-based CASQ (Bradley1999). In all of these variants of the self-adminis-tered questionnaire, responses are either sent directlyback to the researcher via e-mail in text form orattached in a separate file, or they are received via aweb-based interface.

The massive expansion of access to and use of theWWW in public (Internet) and private (intranet)networked environments over the past 10 years,accompanied by improvements in user interface andscripting support, has made web-based CASQs avery popular format for the collection of social sci-ence data (Dahlen 2002; Gosling et. al. 2004; Rosset. al. 2003). Web-based CASQs generally come inthree varieties: an open website where no control is


NEL

questionnaire, or experiment (Barrett & Barrett2001; Schmidt 2001). Flexible ordering options canbe implemented to reduce, or at least control for,order effects. Additionally, the strategic use of inter-face design and scripting allows researchers to incor-porate adaptive questioning into the researchinstrument in a manner that is neither obvious nordisturbing to the research participant. With adaptivequestioning, answers to specific questions help influ-ence which subsequent questions will be asked(Bauman et al. 1998; Liu et al. 2001). The obviousadvantages of this design feature are that individualrespondents answer fewer questions that are irrele-vant to them, and the complexity of the overallinstrument is reduced for respondents since they nolonger need to read and follow skip patterns andinstruction sets.

In addition to enhancing the complexity of thedata collection design, a well-constructed comput-erized instrument can help ensure that questions arecompleted and completed accurately (i.e., helpsreduce invalid responses, item non-response bias,and interviewer error; Liu et al. 2001). Unlike ahuman interviewer who may forget to ask a specificquestion, with CAPI or CASI all questions are askedbecause the computer always follows the pro-grammed routine (Peiris et al. 2000). In less struc-tured interview and observation instruments,researchers can script in pop-up dialogue boxes thatautomatically request that the interviewer probe fur-ther if a certain number of keystrokes are not presentin a particular answer or if the respondent has notgiven a very detailed response. Finally, in CASQs itis quite common for researchers to build in pro-grammed checks of the responses provided to ensurethat all required questions have been answered andthat the information provided corresponds to theexpected format.

Perhaps one of the most promising possibilitiesthat computerized instruments offer in the way offormat and design comes in the form of improvedaccess. Multi-modal participant input devices can becreated to facilitate the participation of people whohave physical disabilities, limited reading or com-puter skills (Black & Ponirakis 2000), and atten-

tion deficits. Additionally, instruments can be cus-tomized to adapt to language and cultural differ-ences. For instance, CASQ participants can beallowed to fill out forms presented in a number ofdifferent languages or the digital voice that is usedin a CAI may be changed to one that the participantis more culturally familiar with (Black & Ponirakis2000). It is also possible to build in instructions orconstruct elaborate help or frequently asked ques-tions (FAQ) sections that can be made available toa research participant at the click of a mouse (Karr2000). Finally, the instrument can also be set up toprovide feedback or instructions to the respondentwhen he or she has problems navigating the instru-ment, filling out questions, or submitting responses(Bauman et al. 1998; Woong, Yun, & Trumbo2000).

TECHNOLOGY AND THE RESEARCH SETTING

At present, the most significant technological limi-tations facing researchers wishing to use computersand network technology in the field are that hand-held computers do not have sufficient disk-drivespace and memory, secure wireless network access issporadic, and the range of mobile research-relatedsoftware applications is limited (Greene 2001;Woong, Yun, & Trumbo 2000). Experimentalresearchers have found that differences among com-puters in graphic display (Horswill & Coster 2001;MacInnes & Taylor 2001; Schmidt 2001), data pro-cessing speed and hardware timing (Eichstaedt2001; Finney 2001), and keyboard and mouse per-formance (Eichstaedt 2001) hinder the experimentaldata collection process. They have also found thatnetworked environments where participants do notinteract directly with the researcher introduce manymore uncontrollable environmental variables,thereby reducing the internal validity of the research(Ross et al. 2003; Wolfe & Reyna 2002). Some fieldresearchers have expressed similar concerns that theuse of technology in field research may influence arespondent’s reactions, with a negative effect on dataquality (Gravlee 2002).

A few researchers have found that the lack ofphysical presence in the research setting leaves them


NEL

with less control over interactions with the partici-pants or the setting (Epstein & Klinkenberg 2001;Petit 2002). When conducting network-adminis-tered focus groups, differences among participantsmay be magnified and variations in the setting maybe overlooked. It is generally not possible for theresearcher or participant to pick up on audio orvisual cues that emerge during the interview orobservational process (Black & Ponirakis 2000).Because there is no visual contact the researcher isnot privy to facial expressions and body language.Nor is the researcher able to pick up on voicetonality (Tse 1999). Furthermore, it is much moredifficult to develop personal rapport in interview orfocus group settings (Black & Ponirakis 2000). Notbeing able to have physical connect with the partic-ipant or pick up on context effects significantlylimits the type and range of data that are available toresearchers.

While there are certainly problems associatedwith implementing technology in the research set-ting, there are also some unique general andmethod-specific advantages. The provisional natureof digital research instruments makes it relativelyeasy to make changes or modifications on the flywithout disrupting the flow of the research (Petit2002). Additionally, some researchers have foundthat the speed (Gravlee 2002) and the privacy andflexibility (Black & Ponirakis 2000) of observationand data collection in computer-assisted settingsincrease respondent’s willingness to participate inthe process. In anonymous CASR environmentsparticipants are more likely to self-disclose (Epsteinet al. 2001; Gravlee 2002) and are thus more likelyto take part in studies of reactive, socially taboo, orhighly sensitive topics. Several researchers havefound that the social-desirability effects are lower inanonymous computerized environments (Evans etal. 2003; Fox et al. 2003; Gravlee 2002). Further-more, in interview and observational designs thevisual separation between the participant and theresearcher can help to reduce potentially biasinginterviewer effects. (Epstein & Klinkenberg 2001).

Some virtual environments have also been foundto enhance existing face-to-face settings and activi-

ties. Virtual environments force researchers torethink the nature of the field setting and ourapproach to selecting and studying such sites. Theinteractions that can be captured in the networkedfield setting are becoming more complex (Best et al.2001; Ruhleder 2000). Networked field environ-ments can be asynchronous or synchronous and theycan contain an array of virtual social artifacts such aspictures, animations, movies, audio recordings, anddocuments.

HANDLING INFORMATION AND OBSERVATIONS

Technical problems can seriously interfere with therecording and storing of research data. They arisefrom many different sources, including network fail-ures, software failures, and scripting errors. Anunstable server or network connection often resultsin data loss or corruption (Bauman et al. 1998;Bosnjak & Tuten 2001; Campbell, Campbell, &Maglio 1999). Different versions of software havebeen found to create confusion among researchersand participants (Bradley 1999; Epstein & Klinken-berg 2001; Sheehan & Hoy 1999) that resulted inmissing or incomplete observations and sample attri-tion. Finally, errors in scripting often lead to prob-lems with the data. Instead of relying on trainedresearchers to record observations CASR relies onpotentially “buggy” script. The former scenario isusually more readily detectable; the latter can bemore difficult, if not impossible, to detect. Extensivepre-testing is costly and time consuming, but vital.Finally, problems with networks, servers, software,and scripting can also result when respondentsintentionally or unintentionally bias research resultsby changing their answers, submitting answers onmultiple occasions, or participating in portions ofthe research that are not applicable to them(Bauman et al. 1998; Sheehan & Hoy 1999;Woong, Yun, & Trumbo 2000).

Concerns are often expressed about respondentsbiasing research results by submitting answers onmultiple occasions or without being part of the pop-ulation of interest. Although this can be a very realproblem, researchers have a number of methodolog-ical, scripting, and database options at their disposal


NEL

to ensure that it is less likely to occur. For example,by using a combination of client and server-sidescripting, secure relational participant and responsedatabases, and randomly assigned numeric pass-words it is possible to exercise complete control overthe processing of questionnaire responses.

STRATEGY AND PROCESS

The most frequent situation in survey and interviewresearch involves a researcher’s approaching a respon-dent that he or she may never have (and in all like-lihood hasn’t) met before. Respondents knownothing about you or your study’s purpose otherthan what you tell them, and their participation is afragile gift that can be withdrawn at any time.

So far, we’ve concentrated on the discrete ele-ments of questionnaires and interview schedules:choosing content, considering different ways toword and structure individual questions. Now it’stime to consider the research instrument as a whole.How do you put a sizable number of these indi-vidual questions together in a way that will make theexperience of completing the survey or interview asenjoyable and free of frustration as possible forrespondents, while ensuring that your own objec-tives as a researcher—to maximize response rates,minimize error, and obtain candid responses—arealso met?

General Organizational Issues

For the most part, general matters of organizationare similar in the preparation of both questionnairesand interviews, but some elements are also clearlyunique. For the questionnaire, because respondentsmust complete the document on their own, aes-thetic concerns like the questionnaire’s “look” andapparent ease of completion play an important role.For the interview, because the interviewer is presentand there’s a constant interchange back and forthbetween the interviewer and respondent, the choicesthe interviewer makes—whether and when to probe,how to rephrase a question, when to hold back andwhen to go ahead with a more intimate question—

will have a big impact. In the discussion below, wewill deal with general questions of organization thatare applicable for both interviews and questionnairesbut will also occasionally pursue matters that areapplicable only to one or the other.

AESTHETIC APPEAL

Many texts, particularly those that deal with morequantitatively oriented survey research, emphasizethe creation of a questionnaire that is aestheticallyinviting and easy to follow (e.g., see Gray & Guppy1994). The general impression given by this litera-ture is that if a questionnaire looks pretty and pro-fessional, is well laid out and easy to follow, doesn’tinclude too many open-ended questions, and doesn’tseem too big, you’ll increase the likelihood of snag-ging a respondent. This view may be overly sim-plistic, but certainly if the converse is true—if yourquestionnaire is not well laid out, seems difficult tofollow, has too many open-ended questions, and isbig enough that it looks like it will take a sizablechunk of time to complete—only the most moti-vated of respondents will complete the question-naire, and your response rates will suffer.

ANTICIPATING A CONVERSATION

Beyond the first impression, authors of survey textssuggest that your next challenge is to organize thequestionnaire or interview and its constituent partsso that it follows a logical sequence and, ideally,reflects and anticipates a social conversation.Numerous principles can be used to guide thesequencing of questions. If a chronological sequenceis involved in the phenomenon being addressed(e.g., the way information is processed in an organ-ization, the development of a romantic relationshipfrom first date to some state of mutual commit-ment, questions on child rearing that ask about thedifferent stages of child development), the question-naire’s sections can merely follow that chronology.Alternatively, questions can be arranged by topic,grouping together the ones that are thematicallyrelated, from general to specific, from most impor-tant to least important, and/or from least threat-ening to most threatening (Gray & Guppy 1994).


NEL

The big trick with creating a questionnaire is totry to organize it in a way that mirrors a conversa-tion you might have with a respondent. The same istrue of the interview, of course, but the very natureof the interview means that you can adapt somewhatto the unique social dynamic that arises with eachrespondent; in contrast, once the questionnaire isphotocopied, you’re stuck with the standardizedsetup you’ve created. Preliminary exploratory work,through exploratory interviews, focus group discus-sions, and/or participant observation, can play a cru-cial role in helping us know how best to organize ourresearch instruments.

THE RESPONDENT’S PERSPECTIVE

It is that actual or imaginary conversation, preferablywith an emphasis on the way the respondent wouldprobably organize things, that should guide thestructure of the interview or questionnaire, and notyour anticipated analysis. Variables can always bereorganized when you start analyzing the data. Toomany novice researchers let their perspective andinterests dominate the interview or questionnaire,instead of putting their own structures “on hold”and letting the respondent’s schema organize theshow. Of course, you’ll want to ensure that all thequestions that are important to you are addressed.But the order in which they are addressed should begoverned by the respondent’s convenience, notyours.

SOCIAL CIVILITY: UNWRITTEN RULES

Perhaps the place to begin is by discussing brieflysome of the unwritten “rules” of conversation, sothat we can then consider how they’d apply to thequestionnaire or interview setting. Imagine thatyou’re at a party or some other social gathering.Somewhere between the chip dip and the petits foursyou find yourself standing next to someone you’venever met, and the two of you begin a conversation.Such conversations never begin with requests forintimate details. (“Hi. Did you ever consider suicideduring your adolescence?”) Instead, they usuallystart with an exchange of pleasantries and chitchatthat does little more than serve as a warm-up and

allow time to determine whether you’ll pursue theconversation further. These beginning portions ofthe conversation typically deal with basic, superficialdetails of our lives that we’re prepared to share withanyone. (“So what do you do?” or “So are you herebecause you know the host or the hostess?”) As longas the person you’re chatting with isn’t completelyweird (in which case the exit/escape sequence isenacted), these initial moments often involve asearch for common interests that can provide thebasis for further conversation.

COMMON GROUND

Once the first “real” basis of conversation is tacitlyagreed on (e.g., you find that you both like outdooractivities such as hiking and camping), the two ofyou will typically “go with it” for a while, perhapscomparing notes on favoured hiking trails or outfit-ters and/or trading stories. This sort of conversationcan consume the whole evening if you find anintense compatibility of interest. More often,though, the first topic soon becomes exhausted; theconversation then either terminates or moves intoanother phase.

Phase 2 may simply be another topic, related tothe first (e.g., you move on to canoeing) or not (e.g.,you find that you also have a common interest inimpressionist painters). But now that you’ve“checked each other out” and decided that you sharesome things in common and that you each seemlike a reasonable human being, the exchange oftenmoves to a more intimate level. Rather than dealingonly with activities and interests, you start to askabout and share more in the way of feelings andopinions. Two people who hadn’t met half an hourearlier are suddenly trading information about somefrustrating aspect of child rearing, comparing theirfears about having vasectomies, or talking aboutwhat dipsticks men (or women) can sometimes be.

OVERSTEPPING BOUNDARIES

Of course, few conversations ever go completelysmoothly. At times, one person may feel comfortableenough to stretch the bounds to a more intimatetopic or level: “So, I hear you and Kim separated


NEL

recently; how’s that going for you?” If such anattempt is made prematurely, the other person iscaught off guard. Perhaps you’ve touched a nerve.Perhaps your motives are unclear: are you asking tobe caring, because you want to ask the other personfor a date, or because you’re about to disclose yourown experiences (which the other person isn’t yetready to hear)? Whatever the reason, an avoidanceritual begins. Sensitive listeners notice these things—the slight blush, the superficial response, the subtlechange of topic—and, respecting that they’vecrossed an inappropriate boundary, back off a bit.Later, when the other person understands ourmotives better or simply feels more comfortable withus, we or the other person may return to that issue,and this time the conversation will continue to flow.But for now, the matter is put on hold and anothertopic is addressed.

WITHDRAWAL

When conversations turn intense, we rarely termi-nate them abruptly. Instead, a “shutdown” sequenceis often enacted: we withdraw gradually, often byreturning to a more superficial level of conversationthat reconnects us with what’s happening aroundus. (“Oh ... I see some people are starting to dance;do you like to dance?”) Sometimes this processmarks the beginning of a new friendship orromance. Other times we merely go our separateways and, despite having enjoyed the interaction,may never see the other person again.

Additional Considerations Unique toResearch

A FORMAL INTRODUCTION

Many of these same “rules” are followed in organ-izing a questionnaire. Your first task in a question-naire or interview should always be to introduceyourself with a brief statement about who you areand the purpose of your study. Any promises you’reprepared to make, for example, that a brief summaryof results will be sent to participants following com-pletion of the study if they’re interested, should also

be made here. In the case of a questionnaire, respon-dents should be told whether they should write theirnames down or complete it anonymously; for inter-views, or when respondents’ names are obtained ona questionnaire, you should clearly specify what stepsyou will take to safeguard their confidentiality. Thiscan all be accomplished in a few sentences or a shortparagraph, for example,

My name is Pat Wallace, and I’m a graduate stu-dent in sociology at Provincial University. Thisquestionnaire is part of my master’s thesis, whichdeals with how different parents teach “appro-priate behaviour” to their children, so it includesa number of questions that ask about your par-enting practices. The whole thing should take nomore than about 20 minutes to complete. I hopeyou’ll answer all the questions, but feel free toleave out any that you feel uncomfortable about.Finally, please note that responses to this ques-tionnaire are intended to be anonymous. If you’dlike to receive a brief summary of the results ofthe study after it is completed, please fill out thesmall card at the end of the questionnaire andsubmit it separately from your completed ques-tionnaire. Thank you very much for agreeing toparticipate.

GETTING TO KNOW YOU

After the basic introduction, the first topic that’sasked about is often relatively trite and superficial,devoted to acquiring preliminary information, forexample, to ensure that the respondent is an eligibleparticipant in the study and perhaps to ask about afew demographic details (although one generallyavoids such “threatening” demographics as incomeand education at this point). Because respondentsare often looking for cues, we give them signpoststhat tell them what we are doing (e.g., “First I needto ask just a few general questions so that we have arecord of how many kids you have and how longyou’ve been a parent”). The first section also sets thepace for the interview or questionnaire, establishinga rhythm of query and response, query and response.


NEL

TRANSITIONS

After passing this “getting to know you” phase, theinterview or questionnaire bridges to the first set ofquestions about the main phenomenon of interest.Again, some sort of signpost is often given, both tokeep the respondent informed and to provide a bitof a mental break before digging into the next sec-tion. For example, a transition might be “Okay, thatcompletes the first section regarding some of yourearly experiences as a parent; now I’d like to ask a fewquestions on how you handle different kinds of sit-uations that can arise with young children.”

If only one set of questions deals with the mainphenomenon of interest, the questions in the setwould normally be ordered from least to most threat-ening; if there are several sets of questions, then the setswould also be ordered from least to most threatening.Each time there’s a change of theme, another signpostshould be offered to help make the transition.

LOOSE ENDS AND THE FINAL WORD

The final section of the interview or questionnaireshould tie up loose ends and leave some positive res-olution. For example, a final section often includessome “basic” demographic items that will help youdescribe the sample and perhaps engage in subgroupanalyses. An additional benefit of including suchitems is that when dealing with populations withknown characteristics, one can use responses to thesequestions to assess the sample’s representativeness. Inany event, the final section should leave respondentswith a good taste in their mouths. Conclude with athank-you, asking respondents (in interviews)whether they have any questions they’d like to askand (for both interview and questionnaire respon-dents) whether they have anything they wish to addor further comments to make.

Perspectives on the Research Interview

Although the above discussion pretty much coversthe “general organization” issues that pertain to themore structured survey methods, interviewresearchers, particularly those who engage in semi-structured or unstructured interviews, have gone

much further in analyzing the process of interactionin the interview setting. Some apparently hopemerely to make prospective interviewers aware ofthe various cues respondents can give (e.g., non-verbal cues) as to whether they’re feeling comfort-able, becoming defensive, or whatever (e.g., Gorden1980; Gray & Guppy 1994). Other texts are writtenmore along the lines of a strategic manual, where thename of the game is to control the setting in such away that the respondent tells all and feels comfort-able about doing so.

A DRAMATURGICAL PERSPECTIVE

Berg (2001), for example, offers what he refers to asa “dramaturgical analysis of the interview,” analyzingin detail the different roles that interviewer andrespondent occupy and the expectations that eachcommonly has of the other. He also discusses how theresearcher can get maximal information with minimaldefensiveness through sensitive attention to bothverbal and nonverbal cues. Resistance in the respon-dent is thus a challenge to be overcome. For example,Berg spends considerable time discussing the “eva-sion tactics” enacted when we step over the line andask about things that are too personal or painful.

Such evasion tactics may involve a word, phrase,or gesture that expresses to another participantthat no further discussion of a particular issue (orin a particular area) is desired. Conversely, peoplealso usually acquire the ability to recognize theseevasion tactics and, in a natural conversationalexchange, to respect them. (84)

But the interview isn’t a “natural” encounter, anddeferring to people’s evasion tactics all the timewould mean that much data of interest would belost. Berg makes no bones about the mission:

This sort of deference [ceremony] simply cannotbe permitted during the course of a research inter-view. In fact, the emergence of evasion tacticsduring the course of an interview is among themost serious obstacles to overcome—but over-come them you must! ... The interviewer must


NEL

Documents

HIGHER RATES OF PARTICIPATIONpalys/ResDec4_Ch6_InterviewOralHist-QualExcerpt.pdf · Merton, Fiske, and Kendall (1956) originally coined the term “focus group” (see Fontana & Frey