82 + FORMS OF INTERVIEWING . 1999. "Response Accuracy When Interviewers Stray from Standardization." Pp. 940-45 in Proceedings of theJoint StatisticalMeeting of the American StatisticalAssociation, Section on Sur- vey Research Methods, 1998. Alexandria, VA: American Statistical Association. Schober, M. F., F. G. Conrad, and S. S. Fricker. 2000. "When and How Should Survey Interviewers Clarify Question Meaning?" In Proceedings of theJoint StatisticalMeetingof theAmerican Statis- ticalAssociation, Section on SurveyResearchMethods, 1999. Alexandria, VA: American Statistical Association. Schuman, H. andJ. M. Converse. 1971. "Effects of Black and White Interviewers on Black Responses in 1968." Public Opinion Quarterly 35:44-68. Schuman, H. and S. Presser. 1981.Questions andAnswers in Attitude Surveys: Experiments on Ques- tion Form, Wording, and Context. San Diego: Academic Press. Schwarz, N., R. M. Groves, and H. Schuman. 1998. "Survey Methods." Pp. 143-79 in Handbook of SocialPsychology, 4th ed., Vol. 1, edited by D. T. Gilbert, S. T. Fiske, and G. Lindzey. New York: McGraw-Hill. Schwarz, N. and S. Sudman, eds. 1994. Autobiographical Memory and the Validity of Retrospective Reports. New York: Springer-Verlag. ,eds. 1996.Answering Questions: Methodology for Determining Cognitive and Communica- tive Processes in Survey Research. San Francisco: Jossey-Bass. Singer,E., H.-J. Hippler, and N. Schwarz. 1992. "Confidentiality Assurances in Surveys: Reassurance or Threat?" International Journal of Public Opinion Research 4:256-68. Singer, E., J. V; Hoewyk, N. Gebler, T. Raghunathan, and K. McGonagle. 1999."The Effect of Incen- tives on Response Rates in Interviewer-Mediated Surveys." Journal of Official Statistics 15:217-30. Singleton, R. A,, Jr. and B. C. Straits. 1999.Approaches to SocialResearch. 3d ed. NewYork: Oxford University Press. Sirken, M. G., D. Herrmann, S. Schechter, N. Schwarz, J. M. Tanur, and R. Tourangeau, eds. 1999. Cognition and Survey Research. New York: John Wiley. Smit,J. H., W Dijkstra, and J. van der Zouwen. 1997. "Suggestive Interviewer Behavior in Surveys: An Experimental Study." Journal of Official Statistics 13:19-28. Steeh, C. G. 1981."Trends in Nonresponse Rates, 1952-1979."Public Opinion Quarterly 45:40-57. Suchman, L. and B. Jordan. 1990. "Interactional Troubles in Face-to-Face Survey Interviews." Jour- nal of the American StatisticalAssociation 85:232-41. Sudman, S. and N. M. Bradburn. 1974. Response Effects in Surveys: A Review and Synthesis. Chi- cago: Aldine. Sudman, S., N. M. Bradburn, and N. Schwarz. 1996. Thinking about Answers: The Application of Cognitive Processes to Survey Methodology. San Francisco: Jossey-Bass. Tanur,J. M., ed. 1992.Questions about Questions: Inquiries into the Cognitive Bases of Surveys. New York: Russell Sage Foundation. Tourangeau, R. and T. W Smith. 1996. "Asking Sensitive Questions: Impact of Data Collection Mode, Question Format, and Question Context." Public Opinion Quarterly 60:275-304. Traugott, M. W, R. M. Groves, and J. M. Lepkowski. 1987. "Using Dual Frame Designs to Reduce Nonresponse in Telephone Surveys." Public Opinion Quarterly 51:522-39. van der Zouwen, J., W Dijkstra, and J. H. Smit. 1991. "Studying Respondent-Interviewer Interac- tion: The Relationship between Interviewing Style, Interviewer Behavior, and Response Behav- ior." Pp. 419-37 in Measurement Errors in Surveys, edited by l? I? Biemer, R. M. Groves, L. E. Lyberg, N. A. Mathiowetz, and S. Sudman. New York: John Wiley. Weeks, M. F. 1992. "Computer-Assisted Survey Methods Information Collection: A Review of CASIC Methods and Their Implications for Survey Operations." Journal of Official Statistics 8:445-65. Weinberg, E. 1983. "Data Collection: Planning and Management." Pp. 329-58 in Handbook of Sur- vey Research, edited by l?H. Rossi,J. D. Wright, and A. B.Anderson. NewYork: Academic Press. IVE INTERVIEWING The emphasis of this chapter is on the epistemology of the tives for an understanding of the process w tends to be more con- and the relevancies of qualitative inter- rs, such as myself, frame it illustration, especially my own. 84 0 FORMS OF INTERVIEWING Qualitative Interviewing 0 85 s A Note on Perspective Donna Luff (1999:701) refers to perspec- tives as "fractured subjectivities." Applied to interviewing, Luff's characterization suggests that participants--both research- ers and respondents-speak to each other not from stable and coherent standpoints, but from varied perspectives. These include the structured and historically grounded roles and hierarchies of their society, partic- ularly those of gender, race, and class (Campbell 1998). Extending this to more local considerations, it also suggests that the perspectives relevant to the qualitative interview encompass the social positions that emerge in the interview itself, apparent in talk and interaction between interviewer and respondent. For example, during an in- terview, the perspective of the respondent may shift from one standpoint in her expe- rience to another, as she speaks, say, as a former child, then as a mother, as a care- giver, then as an employee, or even as one who watches the local news (Holstein and Gubrium 1995; see also Gubrium and Hol- stein, Chapter 1, this volume). Although situational, these perspectives shape the flow of the interview and, in its qualitative version, are taken into account by the inter- viewer in understanding the meaning- making process. In most texts on qualitative interview- ing, the perspective of the interviewer is talten to be that of the discipline: she or he is interviewing in order to write, publish, and contribute to a body of knowledge and literature. The ways in which this disciplin- ary task are conceived is historically grounded, with the planning, conduct, and interpretation of interviews shaped by changing rules and expectations. What was viewed as improper procedure at an earlier time might now be de rigueur, as changing concepts of the interview task become ac- cepted (see in this volume Platt, Chapter 2, as well as all contributions to Part VI). In- deed, even the significance of perspectives is historically grounded, with the current recognition that perspectives other than those drawn from the discipline come into play for the interviewer as well as the re- spondent, especially in qualitative inter- viewing2 Much has been written on the respon- dent's perspectives in the qualitative inter- view, especially in relation to gender (Arendell 1997; Warren and Hackney 2000). The chapters in Part I1 of this vol- ume are, in some sense, an outline of an ac- cumulated discourse on types of respon- dents, including the respondent as ethnic, gendered, aged, classed, and identified with one or another sexual community. An important point to emphasize here is that these are not only distinctive respondents but various perspectives that can be taken up by a single respondent within a single in- terview. Perspective is especially significant in qualitative interviewing, where meaning making is center stage in the interpretive process. My own disciplinary and research expe- rience, for example, forms a perspective, one that gives shape to how I present the qualitative intervievc3 I write this chapter from the perspective of a seasoned sociolo- gist who has done qualitative interviewing and extensive writing about interviews. During the 1980s, I interviewed respon- dents for two projects, one on older women married to younger men (Warren 1996) and the other on patients, relatives, psychi- atrists, and hospital administrators in- volved with electroconvulsive therapy (ECT), formerly known as electroshock therapy or EST (Warren and Levy 1991; Kneeland and Warren forthcoming). My ethnographic study of a gay community in the late 1960s also included interviewing (Warren 1972).In another study, I analyzed 30,000 pages of interviews with 17women diagnosed as schizophrenic in the late 1950sand early 1960s, and their husbands. This was known as the Bay Area study (Sampson, Messinger, and Towne 1964), material from which I used to write a monograph titled Madwives on the inter- section of psychiatry, gender, and marital roles during that era (Warren 1987).I drew on the interpretive, feminist perspectives of the 1980s for my reinterpretation of this material. I have also written or coauthored a number of methodological articles on in- terviewing based on Madwives and other research (Warren 1985; Harkess and War- ren 1993; Karner and Warren 1995) and, like many of my colleagues, have super- vised several generations of student re- search using qualitative interviewing and ethnography.4All of this shapes this presen- tation, which, together with other perspec- tives, both within and outside the qualita- tive interview, will bring up noteworthy points throughout the chapter. g, Preliminary Considerations Qualitative interviewing is a kind of guided conversation (Kvale 1996; Rubin and RLI- bin 1995)in which the researcher carefully listens "so as to hear the meaning" of what is being conveyed (Rubin and Rubin 1995:7).James Spradley (1979:8) extends the concept of listening to include dis- tinctly disciplinary concerns. According to Spradley, the purpose of interviewing is to make "cultural inferences," thick descrip- tions of a given social world analyzed for cultural patterns and themes. These are of typical anthropological interest, which is Spradley's own disciplinary context. Spradley explains that qualitative research- ers make cultural inferences from three sources: what people say, the ways they act, and the artifacts they use. Taken together, these sources implicate qualitative inter- viewing's sister research genre, ethnogra- phy (see Atkinson and Coffey, Chapter 38, this volume.) QUALITATIVE INTERVIEWING AND ETHNOGRAPHY Qualitative interviewing has long been linked to ethnographic fieldwork, a tradi- tional staple of anthropological research. Today, it is linked to many other disciplin- ary contexts. Qualitative interviewing and fieldwork are often classified together, along with documentary analysis, as quali- tative or interpretive methods (ICvale 1996:Y; Rubin and Rubin 1995:34-35).Yet the "cultural inferences" that the qualita- tive methods of ethnography and inter- viewing provide give us subtly different lenses on the world. Ethnography's lens is that of lived experience, set in an eternal present. The lens of the intensive interview is verbal-what people say and mean-but its temporal range is biographical, extend- ing into the past and the future. In this re- gard, contrast Erving Goffman's (1961) ethnography of a late-1950s mental hospi- tal with my own interview-based study of mental patients (Warren 1987). From Goffman, we see staff-patient and patient- patient interaction in the context of that decade's eternal present. From my work, we see the meaning of mental-patienthood in the context of 1950s housewifery. Researchers often choose qualitative in- terviews over ethnographic methods when their topics of interest do not center on par- ticular settings but their concern is with es- tablishing common patterns or themes be- tween particular types of respondents. As Rubin and Rubin (1995) note, interview topics come from many sources: "employ- ers; life experiences . . . the researcher's personality; from ethnic, racial, or sexual identity. Some subjects attract researchers' curiosity; others appeal to researchers' po- litical or social values" (p. 49). Where both settings and individuals are available, and are mutually pertinent, re- searchers often combine ethnographic data with interview data, illuminating both the culture and the biographical particulars of members' worlds. Social researchers use ethnographic interviews and other field- based methods to "fill in" the biographical meanings of observed interactions (Sprad- ley and Mann 1975; Esterberg 1997).' These methods hearken back, in sociology, to the Chicago school and its methods, 86 4 FORMS OF INTERVIEWING Qualitative Interviewing 4 87 which combined surveys, case studies, doc- umentary analysis, and qualitative inter- viewing. These methods were brought to- gether in the service of understanding the varieties of experience that made up the Chicago urban experience in the 1930sand 1940s. Although Chicago school scholars were short on methodological treatises and ruminations (in general they just did their job, but see Palmer 1928), certain aspects of contemporary qualitative interviewing, and its penchant for ethnographic linkages, can be seen as linear inheritors of the Chi- cago school. DESIGNING QUALITATNE INTERVIEW RESEARCH Steiner Kvale (1996) writes that the original Greek meaning of the word method is "a route that leads to the goal" (p. 4). Extending this concept by way of a traveler's metaphor to the qualitative inter- view researcher, Kvale adds, "The inter- viewer wanders along with the local inhab- itants, asks questions that lead the subjects to tell their own stories of the lived world, and converses with them in the original Latin meaning of conversation as 'wander- ing together with' '' (p. 4). The design of qualitative interview research, for Kvale, is open-ended in the sense that it is more con- cerned with being attuned to who is being traveled with, so to speak, than with setting out a precise route for all to follow, as in survey research. As with ethnography in earlier decades, the wanderings of qualitative interviewing became systematized into texts and mono- graphs during the 1990s (Arksey and Knight 1999; Holstein and Gubrium 1995; Kvale 1996; Rubin and Rubin 1995; Weiss 1994). IGale (1996:88) proposes that, like Shakespeare's "man," interviewing has seven stages: thematizing, designing, inter- viewing, transcribing, analyzing, verifying, and reporting. By thematizing, he means thinking about the topic of interest to the researcher and its fit with the interview method; qualitative interviewing is de- signed with the aim of thematizing the respondent's experience as well. Of course, designing the research may involve reviewing the existing qualitative (andperhaps quantitative)literature on the topic to determine whether a new qualita- tive interview study would add anything to it. The researcher also considers the time available to complete the study, access to respondents, and the financial and emo- tional costs of conducting the study (Rubin and Rubin 1995:54). Emotional costs are particularly relevant in qualitative inter- viewing because of its open-ended, explor- atory character; probing for details and depths of experiences (seeJohnson, Chap- ter 5, this volume) can be stressful for all participants. At the same time, beyond the standard issues such as reviews of the existing litera- ture and the practical matters of time and access, qualitative researchers' concern with meaning making causes them to be rather skeptical of standard design strictures. For example, the constructionist epistemolog- ical leanings undergirding much of qualita- tive research beg the researcher to move ahead and interview open-endedly. The goal is to unveil the distinctive meaning- making actions of interview participants. As such, the design of qualitative interview research necessarily places limits on stan- dardization and the working relevance of existing literature. This is not to say that the research litera- ture is unimportant. It is, but its relevance for the design of interviewing is confined to the first steps, if it is taken into account at all. From the "research questions" gener- ated by a possible review of the literature, the interviewer develops 10 to 12 specific questions, together with a face sheet cover- ing such descriptors as respondent age, race, and gender. Rubin and Rubin (1995: 145-46) note that the qualitative interview uses three kinds of questions: main ques- tions that begin and guide the conversation, probes to clarify answers or request further examples, and follow-up questions that pursue the implications of answers to main questions. But, equally important, the qual- itative interviewer remains flexible and at- tentive to the variety of meanings that may emerge as the interview progresses. This open stance includes being alert to develop- ing meanings that may render previously designed questions irrelevant in light of the changing contexts of meaning6 FINDING RESPONDENTS Whom does one interview? In the logic of survey research, interviews are con- ducted with a representative sample of a larger population, drawn systematically in order that the findings will be generalizable to that population. In qualitative interview studies, respondents may be chosen based on a priori research design, theoretical sampling, or "snowball" or convenience design, or particular respondents may be sought out to act as key informants (Hol- stein and Gubrium 1995; Spradley 1979). In the Bay Area study, respondents were se- lected by a priori research design. Inter- viewers were to approach Caucasian, mar- ried women with children who were first admissions to Napa State Hospital within one week of their admission (Sampsonet al. 1964; Warren 1987). Such a priori stric- tures, of course, do not always work out. One respondent was found to have had prior psychiatric admissions, but she was kept in the sample because, by the time this discovery was made, a great deal of time and effort had been expended in interview- ing her. Using a theoretical sampling strategy, the interviewer seeks out respondents who seem likely to epitomize the analytic crite- ria in which he or she is interested (see Glaser and Strauss 1967; see also Charmaz, Chapter 32, this volume). Because the ob- ject of qualitative interviewing is to discern meaningful patterns within thick descrip- tion, researchers may try to minimize or maximize differences among respondents -say, according to race or class-in order to highlight or contrast patterns. In gen- eral, with one-time interviews, the more comparisons to be made between sets of patterns, the more respondents are likely to be interviewed. For example, a researcher studying male caregivers of elderly Alzhei- mer's patients may decide on 20 or 25 in- terviews, whereas a researcher comparing male and female caregivers may seek 35 or 40.' Theoretical sampling may be carried out through a "snowball" process: One respon- dent is located who fulfills the theoretical criteria, then that person helps to locate others through her or his social networks (Arksey and Knight 1999:4; Biernacki and Waldorf 1981; Weiss 1994:25). But there are many other ingenious ways in which qualitative researchers find respondents to interview. For example, one sociology graduate student at the University of South- ern California who was interested in the topic of interracial marriage approached her respondents during her working hours as a supermarket checker. Any time she checked the groceries of an apparently in- terracial couple, she asked them if they would be willing to be interviewed. Most of them agreed, to my supervisorly surprise. One of the problems in seeking respon- dents for an interview study may be, in Hil- lary Arksey and Peter Knight's (1999:70) terms, not being able to find anyone to talk to. This can be a problem, especially when the topic of the interview is stigmatizing or when the occurrence of needed respon- dents is rare in a population. Both were true for our study of elderly ECT patients (War- ren and Levy 1991). For other topics, such as that of Laurel Richardson's The New Other Woman (1985), finding respondents is less difficult, even if personally stigmatiz- ing. As Richardson says: t@K ksz Finding "other women" to interviewfqj @ was not difficult. . . . I announced my research interest to nearly everyone I ag met-conferees, salesclerks, travel ac- $%! quaintances, and so on. Women I met in !8p 114these different circumstances volun- 88 Q FORMS OF INTERVIEWING Qualitative Interviewing Q 89 teered to be interviewed, or put me in contact with women who were in- volved with married men. (I?x) In ethnographic interviews, informants may be chosen for their communicative competence or access to information rather than their personal epitomization of some topic-related characteristic of interest to the researcher (Briggs 1986). As Spradley (1979)notes: "I use the term ifzforfnant in a very specific way, not to be confused with concepts like subject, respondent, friend, or actor. . . . Informants are first and fore- most native speakers" (p. 8), one connota- tion of which is that they have inside knowledge of some social world. Where in- terviewer and interviewee share the same life world, however, the selection of an in- formant may be based more on the particu- lar standpoint from which the individual can interpret cultural meanings. As James Holstein and Jaber Gubrium (1995) state, "The term informant no longer conveys a distinct difference in narrative compe- tence; instead it signals more a difference in point of view" (p. 24). Indeed, because of their interest in the construction, not just the substance, of meaning making, Hol- stein and Gubrium propose that, where there is a choice, qualitative interviewers should select "respondents because they are assumed to be capable of narrative pro- duction" (p. 24), thus dignifying them as people and orienting to the interview proj- ect as narrative collaboration. Both positivist and constructionist dis- cussions of respondent selection tend to as- sume that the interviewer and respondent will be strangers; indeed, the title of a re- cent text on qualitative interviewing is Learning from Strangers (Weiss 1994). However, this may not be the case. Rich- ardson (1985), for example, included fel- low conferees and acquaintances among her 55 respondents. In ethnographic stud- ies, where the researcher is a member of the community she or he is studying, respon- dents may even be a part of the inter- viewer's own social circle. Kristin Esterberg (1997) describes her theoretical sampling of members of a community with which she was quite familiar: The initial interviewees were selected, in part, for their location in the commu- nity; I actively sought out those who were seen by others at the "center" and at the "margins" of community. I also sought out women, with varying degrees of success, in "under-represented" cate- gories: old women, bisexual women, working-class women, and women of color. (Pp. 177-78) In some cases, sampling begins with ac- quaintances and moves on to strangers. This is typical of snowball sampling. In the ECT study (Warrenand Levy 1991),we ini- tially posted flyers in nursing homes seek- ing respondents, with absolutely no luck. In discussing the study with colleagues and friends, however, we found that many had elderly relatives who had had ECT. Simi- larly, in our study of older women married to younger men, respondents included uni- versity colleagues, friends, and even a cleaning woman who worked for one of the researchers. INFORMED CONSENT As with other ltinds of research involv- ing human subjects, qualitative interview- ing requires researchers to deal with profes- sional ethical codes, in particular federal and university human subjects regulations. These have become more formalized over the past several decades, to the point where some say that they unduly constrain the conduct of social research or protect the re- searcher more than the subjects of the re- search (see Adler and Adler, Chapter 25, this volume). Institutional review boards (IRBs) translate federal policy into local standards for the protection of human sub- jects from physical and emotional harm by requiring researchers to obtain informed consent from research subjects. From an IRB perspective, human sub- jects regulation of interview research seeks to protect respondents from such things as invasion of privacy, breaches of confidenti- ality or anonymity, and distress caused by topics raised in the interview process itself. But from the standpoint of understanding qualitative interviewing, what is interesting about these strictures is not so much the ways they are implemented by the re- searcher, but the ways they are interpreted by the respondent. Among dangers or harms in intensive in- terviewing research from the perspective of the respondent is the act of listening itself. Listening to another speak, for example, is an act that reflects the self back to the re- spondent, and this may unfold in ways un- foreseen by IRBs or researchers themselves. In reflecting on repeat interviews with ex-patients in the Bay Area conducted in the 1950s and Vietnam veterans in the 1990s, I found that the interviewer becomes dangerous by the simple act of listening: when the speaker has put on the mantle of a new self seeking to bury the old in an un- marked grave, yet must confront the presence of an interviewer who has knowledge of the past self. The listener is also dangerous as a participant in the retelling of the past by a respondent who feels unable to escape from that past and the self constituted by it. (ICarner and Warren 1995:81) Some subjects may not see written con- sent forms as at all protective. In a study conducted by a University of Southern Cal- ifornia graduate student, respondents ex- pressed repeated exasperation with con- sent forms. This particular study focused on lesbian identities. The researcher's re- quests for interviews-which included clear promises of confidentiality yet re- quired signed consent forms-were uni- formly met with exasperated refusals by prospective respondents. The contradic- tion between requiring signed consent forms, which prospective respondents per- ceived as going to the government funding agency, and promising confidentiality was too great. The researcher resolved the problem by shifting to oral, tape-recorded consent. In the team qualitative interview study in which I participated in the late 1980s (Warren and Levy 1991), in which we sought interviews with elderly ECT recipi- ents, their relatives, hospital psychiatrists, and hospital administrators, none of the patients or relatives took issue with the consent forms. But most of the hospital psychiatrists and administrators waved them away as "too official." They were willing to talk with us, but they were not willing to put their names to any docu- ments that might involve them in future liti- gation. Curiously enough, such responses are often not discovered until after the in- terview process has begun, the start of which the consent form is meant to regu- late. The logic of informed consent presumes that the respondent will understand the in- tent of the research, as it is explained by the researcher or a consent letter. However, there are many indications in the literature on qualitative interviewing that the re- searcher's understanding may not match the interviewer's from the start, may shift over time, or may be "confused." The fol- lowing extract from an interview with a Bay Area ex-patient-whom the researcher had interviewed at least 50 times over a 36-month period-il!ustrates the dynamics involved: She began by asking what kind of a psy- chologist I was. . . . "You said that you were working on a project. I was won- dering what your field was. ...at times, as I said, I was confused about what your interest was in the family, whether you were prying or whether you were just surveying to see how the family was getting along, with your connection with the hospital in your field, whether it has helped out or whether it was part 90 * FORMS OF INTERVIEWING Qualitative Interviewing + 91 of it-it wasn't really that, it was just simple explanation of the confusement of it all." (Warren and Karner 1990: 123) SETTING UP THE INTERVIEW Once the researcher identifies respon- dents, she or he must ask them if they will agree to be interviewed, a process that usu- ally accompanies obtaining informed con- sent. In particular, the time and place of the interview needs to be decided. The received wisdom on how to accomplish this is highly varied, with some commentators advising particular venues and specific kinds of scheduling (see Seidman 1998) and others leaving this largely an open question (for example, see Kvale 1996). In my experience, the continuum of re- sponses to these preliminary matters can range from outright refusal to welcoming agreement, with every variation in be- tween. In the original Bay Area study, one husband refused to be interviewed at all. More generally, a willingness, even an ea- gerness, to talk about oneself in interviews is quite commonly reported, at least in the American context. Indeed, as Rubin and Rubin (1995) note: At a basic level, people like to talk about themselves: they enjoy the sociability of a long discussion and are pleased that somebody is interested in them. ...you come along and say, yes, what you know is valuable, it should not be lost, teach me, and through me, teach oth- ers. (l? 103) Setting up the interview and actually making it happen are two different things. Generations of qualitative interviewers have been admonished to schedule inter- views at times and in places convenient to respondents, but they may find that even this is problematic. For example, an under- graduate sociology student at the Univer- sity of Kansas had, with great difficulty, scheduled a focus group session for six stu- dents to talk about the issue of going to school and working at the same time. When the scheduled time for the group to meet expired, he ran into my office and breath- lessly announced that not one of the six had appeared. Although this incident may be extreme, it is not uncommon for respon- dents to forget, simply not show up, or in other ways delay or prevent the actual com- pletion of the interview. But let us continue with those interviews that do move ahead. Armed with a list of questions, a fact sheet for demographic in- formation, the informed consent letter, and the requisite tape recorder and backup pen- cil and paper, the interviewer meets the re- spondent at the agreed-upon location. The location itself may have been negotiated. In the BayArea study, the female respondents, once out of the hospital, did not know quite where to meet their male interviewers. The home seemed out of the question-what would the neighbors say? And the same might be said for the coffee shop across the road. On the other hand, a journey to the researchers' offices, although far from the gaze of prying eyes, was logistically diffi- cult, given child-care and household re- sponsibilities. These ex-patient interviews were replete with discussion and discom- fort over the issue of where to meet; in the summer, interviewers sometimes resolved the problem by meeting with respondents in the outdoors, in a garden or on a park bench. Most interviews were eventually completed, but their locations were far from being the result of a well-defined method of procedure. In retrospect, it is ev- ident that the negotiation of perspectives on this matter filtered many of these pre- liminary issues, just as many seasoned qual- itative researchers have noted that such ne- gotiations indeed reverberate throughout the interview process itself. A respondent is, by definition, someone who responds-someone who is willing and able to talk to the interviewer. But the respondent is also raced, classed, and gendered as well as being situated in the present moment, with anticipatory notions of what an interview might entail. All this, too, will reverberate in the forthcoming in- terview. Nancy Ammerman's (1987) eth- nographic study of fundamentalist Chris- tians, for example, illustrates the religious, class, and educational perspectives from which her respondents anticipated inter- views: My role as an interviewer often placed an initial distance between me and my subjects that was not present in my role as a participant observer. . . . a good many people approached the interview full of apprehension about what it would be like to be interviewed by someone who was getting a Ph.D. from Yale. After they had cleaned their houses, prepared special food, and even bought new clothes, some still worried about whether they would know the "right" answers and why I had chosen them instead of someone who was a stronger Christian or had been in the church longer or who had a more inter- esting testimony. (l?13) Clearly, the procedural staging of the quali- tative interview develops both extempora- neously and methodically within the social relations of the participants. + The Qualitative Interviewing Process We now turn to the interview process itself, in particular to the meaning making in- volved as it relates to the social interaction of the participants. This has been a com- mon topic in the interview methods litera- ture for years (see DeSantis 1980; Suchman and Jordan 1990; Peneff 1988). Meaning making is especially pertinent to qualitative researchers because their constructionist leanings bring the interview process itself within the purview of the designated re- search topic. The social contexts of the interview process are not viewed as some- thing to be controlled, as they are in stan- dardized survey interviews, but instead are seen as an important part of meaning mak- ing in its own right. Qualitative research- ers, in other words, treat the unfolding so- cial contexts of the interview as data, not as somethingthat, under ideal conditions, can be eliminated from the interview process. To illustrate these unfolding social con- texts, I begin at the very start of the inter- view, when the tape recorder is set up, and end after it is over, with the "echoes" that can follow the respondent and researcher into their other lives. Between the begin- ning and the echoes, interviews can take many directions. Here, I depict two such directions: currents of the clinical and the sociable-of loyalty and disloyalty-that occurred in situations where one inter- viewer interviewed spouses (separately), and issues of gender and power in feminist interviewing. THE TAPE RECORDER AND ITS MEANINGS The interview often begins as the inter- viewer's tape recorder is set up amid friendly greetings, creating a particular so- cial context for the interview conversation. For several decades, the conventional wis- dom has been that qualitative interviews should be audiotaped, and perhaps even videotaped (Holstein and Gubrium, 1995:78).But does the respondent remain basically unaffected by this? Not only might turning on a tape recorder alter the ensuing conversation, creating a particular context for what is said, but the meanings of audio- or videotaping may be different to different respondents, whose perspec- tives on the matter are likely to vary by so- cial class and age, for example. Tape recording has historical reso- nances. The tape recorder itself, ubiquitous in recent decades, was a novelty at the time 92 @ FORMS OF INTERVIEWING Qualitative Interviewing O 93 of the Bay Area study; indeed, the first half of the study was conducted with the eight interviewers taking handwritten notes. When tape recorders were introduced, they were a source of exclamation and discus- sion on the part of the respondents, who would bring their children into the room to examine and discuss the then-bulky instru- ments. But times and expectations change. In my study of older women married to younger men, the one working-class His- panic couple I interviewed met me at their front door with exclamations of disap- pointment over my small and insignificant tape recorder. Their concept of the "inter- viewn-shaped by the TV program Eye on L.A.-had led them to expect me to arrive with a video camera, perhaps even a TV camera crew. What Paul Atkinson and Da- vid Silverman (1997)call "the interview so- ciety" seems to have constructed a new, postmodern, social context for interview data, perhaps making the interview itself the characteristic format for personal nar- ratives (see Gubrium and Holstein, Chap- ter 1, this volume). In the process of conducting qualitative interviews, many of us have encountered the "on and off the record" associations that respondents have with recording de- vices. In perhaps the majority of interviews that I have conducted, supervised, or ana- lyzed, from the 1960s through the 1990s, respondents have continued to speak after the tape recorders have been turned off. This seems to occur for two reasons: (a) The respondent wants to talk about his or her own, rather than the interviewer's, con- cerns; and (b) the respondent does not want to talk "on the record" about issues that might be dangerous or personally dam- aging. For example, my notes from the Bay Area study show that an interview with one husband was extended past its conclusion, with the husband offering some telling re- marks about ECT that had not been forth- coming in the interview proper: As I packed up the tape recorder, Mr. IN asked me what ECT does for people. I muttered something about, "I wish I knew." He responded with, "Well, what's it supposed to do?" In another instance, as I turned off the tape recorder in a 1980sinterview with a hospi- tal psychiatrist concerning ECT, the psychi- atrist said, "Now that we can talk off the re- cord, I will tell you about billing." It is a hallmark of qualitative interviewing that "unrecorded" data of this kind are as im- portant as those derived from tape record- lngs. SHIFTING CONTEXTS Whatever the training and intentions of the interviewer, the social interaction of the qualitative interview may unfold in unex- pected ways. This unfolding is even more complexwhen interviews are repeated over time. Once again, the interview process it- self can be treated as an important source of data. In the Bay Area interviews, for exam- ple, "clinical" perspectives emerged in in- terviews with the women patients and ex-patients, whereas "distancing" perspec- tives emerged with both the ex-patients and their husbands after the wives' release from hospitalization. Although interviewers were trained to be nonpartisan with these husbands and wives, they nevertheless were at times treated as partisan. In their training, the Bay Area study in- terviewers were instructed not to act as cli- nicians during the study. One psychiatrist warned that "any sort of regular relation- ship was bound to be therapeutic (or antitherapeutic) notwithstanding our 'in- tentions' " (Warren 1985:74). And from the point of view expressed in interviewers' accounts, it was apparent that this psychia- trist's warnings were appropriate. The women patients and ex-patients asked the researchers for help, advice, and opinions, as did their husbands. "Transference" also seemed to affect the interactions between respondents and interviewers (Laslett and Rapoport 1975; Warren 1985).After one interviewwith patient Joyce Noon on April 3, 1959, the interviewer commented: # 1had originally anticipated that I would $93~ stop the lntervlew after about one tape, but since Joyce seemed to be getting some benef~tfrom talking to me and ex- pressing her feellngs, I went on for an- other tape to glve her further opportu- nlty to do so. (Warren 1985:80) In the ex-patient phase of the Bay Area study, the issue of the interviewer as dan- gerous listener was especially salient. Some of the women and their husbands sought to distance themselves from the women's "old selves," a distancing that extended to the researcher. This, in turn, affected the social interactions within these interviews. For example, ex-patient June Mark said that she cannot fully participate in the re- search simply because the research in it- self signifies the stigma of deviance which she is struggling to avoid. . . . "You keep asking a lot of questions .. . things Iwant to forget about. ...It's not normal, my talking to you. . . . It's just that I am reminded I'm a patient. If you're a patient, you're always a pa- tient." (Field notes) Despite her strong reservations, June continued to participate, as did all but one of the respondents. However, they did try to redirect the interviews into more socia- ble, everyday-in June Mark's word, "nor- mal"-channels. For example, in response to one interviewer's "How are you?" the respondent answered "How are you?" in a pointed attempt at role reversal (Warren 1987:261). This rather explicit attempt at reconstructing the interaction not only al- tered the social context of the interview, changing it from an interview with an ex-patient to one with another person, it presented itself as data in the sense that it documented, on that occasion, the normal- izing work of everyday life for this popula- tion. This is one of those many points in qualitative interviews when the interview becomes ethnographic. Building a context for sociability, rather than data gathering, was especially appar- ent in posthospital interviews with Ann Rand. One of the interviewer's notes in this case reads: Repeated that she would only see [me] again if she would have her over to her house. While the interviewer was eva- sive, Ann said, "Then I suppose you still see me as a patient. To me you are either a friend or some kind of authority, now which is it?The way I see it, you either see me as a friend or a patient." (Warren 1987:261) Another note, this one concerning Jack Oren's interview, reads, "Mr. Oren asked me if I wanted to join them for dinner, and was rather insistent about this despite my repeated declining" (Warren 1987:261). Other respondents turned the psychiat- ric tables on the interviewers, interpreting them clinically, as the following note indi- cates: [Referring to the interviewer], Jack Oren said, "I think that you're a kid that missed happiness somewhere along the line." He then started speculating about my past life and thought that something had happened to me. ..to make me feel like that. Mr. Oren first was critical about my interviewing technique, then started to question me about my life, and so on. (Warren 1987:62) But not all of the ex-patients sought to re- lease themselves from the researcher's grasp on their past selves, or saw this as dangerous. Some continued to therapeu- tize the researcher and the interviews, as the following notes about two respondents suggest: V T {& I had the feellng that Irene James was q :*@ ,gdesperately trylng to gain some control ia6/j/ over her feellngs and thoughts by talk- 94 + FORMS OF INTERVIEWING Qualitative Interviewing + 95 -6% pq ing about them to me. ...Irene saysthatc11v @J when I arrive for my interview that @$seems reassuring. (Warren 1987:262) """'j I asked Shirley Arlen if she would see a psychiatrist and she said no, she could- 8 n't afford it, then all she would do is I talk, and she feels she would do better ,,' just talking to me. (Warren 1987:262) In the Bay Area study, each of the eight female and male interviewers spoke with both the mental patients and their hus- bands. Each wife and husband knew that the other was being interviewed by the same person, forming a triadic relation- ship. One consequence of this arrangement was that the interviewer was incorporated into the respondents' attempts to find out and pass on information and opinions con- cerning, mainly, the wife's mental condi- tion at the time. In such a situation, the in- terviewer is supposed to be, in Georg Simmel's (1950)words, a nonpartisan who either stands above the contrasting interests and opinions [of the dyad] and is actu- ally not concerned with them, or . . .is equally concerned with them. . . . the non-partisan may make the interaction between the parties, and between him- self and them, a means for his own pur- poses. (Pp. 149-50) Regarding the latter point, it is characteris- tic of qualitative interviewing that it is structured to take these options seriously, generating new data in the process. It was clear in the Bay Area transcripts that re- spondents took varied perspectives in the interview, some of which were far from be- ing neutral sources of information. Interviewed husbands often asked the interviewer about their wives, and when they did not hear what they wanted, some became testy: Mr. Sand told me that he didn't see any point going on [with the interviews]. ... He asked me if I had talked to his wife that day and when I did not answer at once he repeated the question and I fi- nally told him that I did. ...He told me that this wasn't going to help him any- way, and besides which, I knew things about what was going on at the hospital with his wife, and I didn't tell him a thing about it. (Warren 1987:266) The respondents' varied perspectives in these triadic relationships-perhaps cen- tered on secrecy in relation to oneself or loyalty in relation to another-are as signif- icant for what they reveal or conceal, in terms of data, as they are indicators of in- terview rapport. Here, again, the ethno- graphic character of the qualitative inter- view is evident. For example, Bay Area ex-patient Joan Baker agreed to continue with her interviews but kept them secret from her husband. She felt he would inter- pret her being interviewed as evidence that she was still mentally ill (Warren 1987: 267). Similarly, in a different research con- text, a woman sociologist, commenting on a draft of an article on interviewing, con- veyed her thoughts about an interview she had just completed and her husband's forthcoming one: I was conscious all through the inter- view of trying to be honest with [the in- terviewer] but not to say anything that would seem disloyal to [husband]. She was going to interview him next, and I kept wondering if she would say any- thing to him that might make him feel I had been disloyal to him. (Harkess and Warren 1993:334) GENDERAS A SOCIAL CONTEXT Although race, ethnicity, nationality, sexual orientation, and age have received increasing attention in the interviewing lit- erature, it is gender to which qualitative re- searchers have been most attentive in soci- ology (Benney, Hughes, and Starr 1956; Luff 1999). In the early years of the Chi- cago school, the authoritative, ques- tion-asking status of the interviewer was unproblematic, the gender of the inter- viewer either unacknowledged or pre- sumptively male. In time, however, the interviewer with "no gender," like the ethnographer as "any person," ceded place during the century to the male interviewer interviewing both women and men (Kinsey's model. .. ). ...During the modern era, accounts of what made an interview go smoothly and produce valid data was contested terrain: any polite and dignified inter- viewer (Palmer 1928), a male inter- viewer (Cressey 192011986) or a fe- male interviewer with a female respondent (Oakley 1981). . . . Not to mention the female "sociability special- ist" of the 1980s wresting secret infor- mation from reluctant male and female nude beach habitues. (Warren and Hackney 2000:37-38, 42) In a historical shift in disciplinary per- spectives, feminist interviewers have sought, over the past several decades, to change the social interactions of the inter- view from being authoritative, sociable, or therapeutic to being expressly egalitarian. By the 1970s,women interviewers were be- ing encouraged to interview other women from the empathic standpoint of gender. By the 1980s, it was commonplace to speak of a special genre of "feminist interviewing" (DeVault 1986; Oakley 1981).In the late 1990s, however, exceptions to, and cri- tiques of, the idea of feminist interviewing appeared and the consideration of respon- dent subjectivity became more complex (see in this volume Reinharz and Chase, Chapter 11; DeVault and McCoy, Chapter 36). The standpoints of race, ethnicity, na- tionality, and sexual orientation were pro- posed as de-essentializing femaleness. Thus "women interviewing women" was com- plicated by whether or not one participant was Third World and one First, one lesbian and one heterosexual, or one religious and radical right and the other left-leaning and feminist (Blee 1991; Luff 1999). Even where both interviewer and re- spon dent are women, interviews may not be with "those whose standpoints the re- searcher shares" in terms of "religious/ secular, feministIantifeminist, or liberal/ heterosexist" (Luff 1999). For example, Luff (1999) discusses how her preconcep- tions affected her interviews with British "moral right" (what American sociologists might call "moral majority") women. She points out that the disciplinary perspectives of sociologists are often secular, feminist, antihomophobic, and politically left-leaning. Her respondents reversed all of these per- spectives; they were religious, antifeminist, homophobic, and politically right-wing. Fur- thermore, as middle-class, semipublic fig- ures, they were "relatively powerful" as well as potentially hostile (p. 687). Retell- ing "moments of rapport" in her research with these women, Luff concludes that "the researcher, as much as the participant, draws on her own conflicting, often contra- dictory aspects of identity as resources in the interaction," adding that the emphasis on power-sharing and the vulnerability of the researched that has characterized much feminist methodol- ogy . . . may come from tendencies within feminist research to study the "powerless" and therefore may not be transferable, indeed may be coun- ter-~roductive,to the development of feminist theory and practice in research with the "powerful." (l?692) By the 1990s, some feminist researchers had come to recognize that women inter- viewing women might not work (Hertz 1996) or might be ethically problematic (Luff 1999).In her interviews with military men and their wives concerning gender in- tegration in the military, Rosanna Hertz (1996) found that the men were uncom- fortable "trying to explain. ..their position to the two female interviewers who were 96 O FORMS OF INTERVIEWING Qualitative Interviewing O 97 outside of male camaraderie" (p. 256). But Hertz also found that the women respon- dents had even less to say than the men; she surmised that this was because the status she shared with them as women was over- shadowed by educational, social class, and marital differences (p. 256). Indeed, Luff (1999:698) points out that rapport-and trust-enhancing interview strategies such as not arguing, saying "I see" and "um," smil- ing, and maintaining a polite tone of voice-can make even (liberal) women in- terviewing (right-wing) women seem de- ceitful, "falling somewhere between the co- vert and overt" in social research. There is general agreement in the quali- tative interviewing literature that women interviewing men presents special prob- lems, given the obduracy of the interper- sonal dominance involved (Arendell 1997; Warren and Hackney 2000). This gender problem was exacerbated for Terry Arendell (1997) in a study where the topic of the interview was divorce. The topic cre- ated an interaction in which male respon- dents spoke forcefully of their betrayal by women to another woman who was the in- terviewer. Arendell found that from the ini- tial point of contact, the interviewbecame a proving ground for masculinity and a site for the exercise of male definitions and dominance displays against ex-wives (and sometimes against all women). These men immediately "took charge" of the interview process and topic and attempted to "place" Arendell as married or unmarried, available or not, male basher or nice girl. Their "as- sertion of superiority" involved both the denigration of women in general and the assumption that their knowledge and in- sights were superior to Arendell's. Their handling of the interview (for it was they who handled) ranged from chivalry to sex- ual harassment (Warren and Hackney 2000:37). POSTINTERVIEW ECHOES Like most things, qualitative interviews come to an end, with respondents and in- terviewers returning to their respective life worlds. For the respondent, there may be no more thoughts of the interview (DeSantis 1980); for the interviewer, the main thoughts may be of the way in which the interview fits into the overall analysis. But sometimes-perhaps especially where interviews are combined with ethno- graphic research-there may be echoes of the interview within the life worlds of the interviewer, the respondent, or both. This possibility was recognized in the 1970s and 1980s literature on feminist interviewing; Luff (1999)refers to this early "assumption that feminists can, or indeed should have a powerful affect [sic] on participants' lives" as "patronizing" (p. 692). Nevertheless, such echoes can occur. Two lesbian sociology graduate students at the University of Kansas who did ethnographic and interview research on their own communities concluded that the interviewing experience created an emo- tional distance between themselves and their respondents. In one case, this ex- tended to emotional distance between the researcher herself and her lesbian identity (Warren 2000). In the research on ECT re- cipients and their families (Warren and Levy 1991),several of our collaborators in- terviewed university colleagues-friends or acquaintances-concerning their el- derly, hospitalized parents. In more than one instance during the interviews, divin- ing our possible critique of the use of ECT on elderly mental patients, our colleagues became upset with us, accusing us of not understanding their situation and, in one case, of no longer being a friend. We sus- pected, too, that one or two respondents simply did not tell us the truth about their family members, avoiding the sort of con- frontation we had had with others. In one case, a prior friendship between an inter- viewer and respondent was severely strained for many months following the interview. Qualitative interviewing is distinctive in this regard. Interviewers do not necessarily end their relationships with respondents at the conclusion of their interviews, as is typ- ically done in survey interviewing. Rather, the perspectives of, and information con- veyed in, interviews echo in the ongoing re- lations of research participants. o Interpretation, Self; and Others The interviewer, like the respondent, par- ticipates in the interview from historically grounded biographical as well as disciplin- ary perspectives. Biographical perspectives may frame entire analyses or affect the se- lection of illustrative quotes. In her book Worlds of Pain (1976),for example, Lillian Rubin tells the reader that her interpreta- tion of working-class life was shaped by her experiences as a working-class child, left- wing political activist, and clinical practi- tioner. She saw pain, and only pain, in working-class lives: "Often people im- plored, even commanded me, to believe they had happy home lives as children. I tried . . . [but the] dominant memories of childhood for me, as for the people I met, are of pain and deprivation" (p. 46). When, in the late 1980s, I was analyzing transcripts of ECT experiences, I saw my- self in respondents' comments, something that was highly emotional for me. In sifting through the many thousands of pages of in- terviews, I chose the following extract from ex-patient Shirley Arlen's case material to illustrate and exemplify the negative as- pects of the biographical memory loss at- tendant upon EST: [Shirley Arlen], although she had been reminded by others of her son's exis- tence, appeared to have lost her affec- tive memory of him as her child: "I guess I feel sort of strange with him. ... I just don't even feel like he's mine, for some reason. . . . I think he's nine months now . . . I really don't ltnow. I can't even remember when he was born." (Warren 1988:295) This comment was particularly poignant for me because while I was writing Mad- wives (Warren 1987) I was a new mother myself, and could imagine nothing more horrible than the emotional separation from a baby. Extending the metaphor of the qualita- tive interviewer as a traveler to strange lands (Kvale 1996), we see that the inter- view, like the ethnography, is about self as well as other (Warren 2000). As Rubin (1976)says of her interview research about working-classpain: "No matter how far we travel, we can never leave our roots behind. I found they claimed me at unexpected times, in unexpected places" (p. 13). As I noted at the start of this chapter, the purpose of qualitative interviewing (and as- sociated fieldwork) is to understand others' meaning making. As many qualitative re- searchers report, I came early on to the point at which I viewed those meanings as intersecting with my own story. Yet, even with our knowledge of the different per- spectives from which respondents and re- searchers talk and write, the empathic ap- preciation of others' meanings is not an easy task, especially across various cultural divides. In Learning How to Ask (1986), Charles Briggs cautions researchers against importing one set of linguistic and cultural assumptions into another when interview- ing between cultures. But it is evident that even within the same culture, meanings that seem clear to the interviewer can be unshared (see in this volume Dunbar, Ro- driguez, and Parker, Chapter 14; Briggs, Chapter 44). In a study of "affirmative ac- tion" in the South in the 1970s, an em- ployer, when asked his definition of the term, replied: Uh ...try to get a job done in as orderly a manner and please our customers . . . so it's firm as possible . . .to get a day's work for a day's pay. . . .And it would be affirmative action. And it's almost impossible. (Harkess and Warren 1994:273) 98 * FORMS OF INTERVIEWING Qualitative Interviewing + 99 Indeed, even the most seemingly com- monplace terms may vary surprisingly in meaning in the context of particular life worlds. In the Bay Area study, sociologist and interviewer Sheldon Messinger talked approximately 25 times between Novem- ber 1957 and July 1958 with ex-patient Kate White (Messinger and Warren 1984). Among the "delusions" that precipitated Kate White's diagnosis and hospitalization was the idea that she and her husband were "homosexual." In the commonsense mean- ings of the 1950s, homosexuality refer- enced, as it does now, same-gender erotic preferences, attraction, or behavior (al- though there would be differences now in the social sensibilities associated with the category). But as Messinger delved into the meaning Kate assigned to the term, it be- came clear that what she was talking about was not desire or eroticism at all, but a so- cial role. She wanted to work outside the home and men did that, so she talked of herself as homosexual. During her hospi- talization, her husband had enjoyed keep- ing house and taking care of the chil- dren-ostensibly a woman's role-so perhaps he was also homosexual. For Kate White, homosexuality referenced gender roles, not sexual desire; in fact, she was having an extramarital heterosexual affair at the time she was interviewed. Messinger and Warren (1984)also point out that stories such as that of Kate White's "homosexuality" are grounded in impor- tant relationships and adaptations that ex- ist outside the purview of the interview. This observation, of course, highlights the necessity of using ethnographic linkages to flesh out the social contexts of meaning making. The social situation of the inter- view may not be the most important one for researchers who are trying to understand the meanings ("frameworks or labels") used by respondents. "These frameworks or labels must be examined in their interac- tion contexts. It is there that they do their work" (Messinger and Warren 1984:205), not in the interview or with the interviewer. So we return full circle to the close rela- tionship between qualitative interviewing and ethnography. I have always found ex- periences and stories such as Kate White's to point me in the direction of multiple rather than one-shot interviews, or of eth- nography combined with interviews rather than interviews alone. But, as Holstein and Gubrium (1995) point out, even in the one-shot interview, the respondent may shift viewpoints and tell different tales. In a 1970sethnographic study of Weight Watchers (Laslett and Warren 1975), I had noticed that a large portion of each meeting was taken up with the discussion of food-what was permitted, how to cook, and so on. This came as no surprise to me. Flush with the then-current ardor for "tri- angulation," I embarked on interviews to "validate" my observation^.^ When I asked my first respondent, "Do you think that the meetings focus on food?"she responded, to my astonishment, with a definite "Oh no!" About one and a half hours later, how- ever-much of which was spent discussing food-she said, "About that earlier ques- tion of yours-well, it does seem like we spend an awful lot of time discussing food doesn't it!" Among the ethnographic quali- ties of the qualitative interview itself is that the interview unfolds reflexively as each participant looks at the world through the other's eyes, incorporating both self and other into the process of interpretation. Although asking, listening, talking, and hearing are important, so are seeing and feeling as means of apprehending the social world. Although the frame of talking and listening may be apt for conceiving tele- phone interviews, the frame of social inter- action accords better with the face-to-face qualitative interview. In the social interac- tion of the qualitative interview, the per- spectivesof the interviewer and the respon- dent dance together for the moment but also extend outward in social space and backward and forward in time. Both are gendered, aged, and otherwise embodied, one person (perhaps) thinking about her topic, questions, rapport, consent forms, and the tape recorder, not to mention feel- ing nervous. The other is (perhaps) preoc- cupied with her relationships outside the interview, pressing tasks left undone, seek- ing information, getting help, or being loyal. These are the working selves and oth- ers at the center of qualitative interviewing. And that is just the beginning. Notes 1. Although interviews may be conducted with more than one interviewer and more than one respondent, I confine this discussion to the dyadic interview situation. See Chapter 7 of this volume for a discussion of group interviewing. 2. Some approaches to interviewing, notably those taking a postmodern perspective, focus more on the interviewer's viewpoint than on the respondent's. Sometimes they fuse these per- spectives. Norman Denzin's (1987) study of self-help groups of which the interviewer or eth- nographer is a member is a case in point. 3. Consider the differences in presentation apparent in the following diversely authored de- pictions: Spradley (1979), Seidman (1998), Weiss (1994), Holstein and Gubrium (1995), Kvale (1996), and Rubin and Rubin (1995). tered in most interview research, and repeat in- terviews, which can be considered a kind of longitudinal design. 5. In a section of his book titled "When Not to Interview," Kvale (1996)notes, "In recent so- cialresearch there has been an inflationary use of interviews; also in areas better covered by other methods." He adds, "If you want to study peo- ple's behavior and their interaction with the en- vironment, the observations of field studies will usually give more valid knowledge than merely asking subjects about their behavior" (p. 104). 6. Indeed, the folk wisdom of qualitative re- search regarding design includes.the. caution that researchers should not consult the literature un- til after the research has gotten under way and they have apprehended a sense of the subject matter. This, of course, works against design as formally understood. 7. Although there are few reasons set forth for the numbers of respondents appropriate in qualitative studies, there seem to be norms. To have a nonethnographic qualitative interview study published, the minimum number of inter- views seems to fall in the range of 20 to 30. Re- spondent groups also generally come in round numbers, such as 20 or 35. 8. The idea of triangulation was discredited in the 1980s,but it is apparently staging a come- back (seeArkseyand Knight, 1999;relatedly, see also Atkinson and Coffey, Chapter 38, this vol- 4. This research includes both one-shot in- ume" terviews, which I suspect is the form encoun- 0 References Ammerman, N. T. 1987.Bible Believers: Fundamentalists in theModern World. New Brunswick, NJ: Rutgers University Press. Arendell, T. 1997. "Reflections on the Researcher-Researched Relationship: A Woman Interviewing Men." Qualitative Sociology 20:341-68. Arksey, H. and I!Knight. 1999.Interviewing for Social Scientists: An Introductory Resource with Ex- amples. Thousand Oaks, CA: Sage. Atkinson, l?and D. Silverman. 1997. "Kundera's Immortality: The Interview Society and the Inven- tion of Self." Qualitative Inquiry 3:304-25. Benney,M., E. C. Hughes, and S. Star. 1956. "Age and Sex in the Interview." AmericanJournal of So- cioiigy 62:143-52. Biernacki. l?and D. Waldorf. 1981. "Snowball Sampling: Problems and Techniques of Chain Referral Sampling." Sociological Methods &Research 10:141-63. Blee, K. M. 1991.Women of the Klan: Racism and Gender in the 1920s. Berkeley: University of Cali- fornia Press.