COLLECTING AND INTERPRETING QUALITATIVE MATERIALS
»ftp*
2. Also recall Fine, Weis, Weseen, and Wong's discussion of focus groups in Volume 1, Chapter 4.
♦ References
Agar, M. (1991). The right brain strikes back. In N. G. Fielding Sc R. M. Lee (Eds.), #         Using computers in qualitative research (pp. 181-194). London: Sage,
f Akeroyd, A. V. (1991). Personal information and qualitative research data: Some practical and ethical problems arising from data protection legislation. In N. G. Fielding &. R. M. Lee (Eds.), Using computers in qualitative research (pp. 89-106). London: Sage. Atkinson, P., & Coffey, A. (1997). Analysing documentary realities. In D. Silverman (Ed.), Qualitative research: Theory, method and practice (pp. 45-62). London: Sage.
Coffey, A., Sc Atkinson, E (1996). Making sense of qualitative data: Complementary research strategies. Thousand Oaks, CA: Sage.
Fielding, N. G., & Lee, R. M. (1998). Computer analysis and qualitative research. London: Sage.
Hayano, D. M. (1979). Auto-ethnography: Paradigms, problems, and prospects. Human Organization, 38, 113-120.
Lee, R. M., & Fielding, N. G. (1991). Computing for qualitative research: Options, problems and potential. In N. G. Fielding & R. M. Lee (Eds.), Using computers in qualitative research {pp. 1-13). London: Sage.
Lieblich, A., Tuval-Mashiach, R., & Zilber, T. (1998). Narrative research. Thousand Oaks, CA; Sage.
Oakley, A. (1981). Interviewing women: A contradiction in terms. In H. Roberts (Ed.), Doing feminist research (pp. 30-61). London: Rourledge & Kegan Paul.
Schwandt, T. A. (1997). Qualitative inquiry; A dictionary of terms. Thousand Oaks, CA: Sage.
Seidel, J. (1991). Method and madness in the application of computer technology to qualitative data analysis. In N. G. Fielding 8c R. M. Lee (Eds.), Using com-putets in qualitative reseatch (pp. 107-116). London: Sage,
Silverman, D. (1998). Harvey Sacks; Social science and conversation analysis. Cambridge: Polity.
Trotter, R. T, & Schensul, J.J. (1998). Methods in applied anthropology. In H. R. Bernard (Ed.), Handbook of cultural methods in cultural anthropology (pp. 691-736). Walnut Creek, CA: AltaMira.
60
1*1-
2
The Interview
From Structured
Questions to Negotiated Text
Andrea Fontana and James H. Frey
Hamlet;   Do you see yonder cloud that's almost in shape of a camel? Polonius: By the mass, and 'tis like a camel, indeed. Hamlet:  Methink it is like a weasel Polonius: It is backed like a weasel. Hamlet:   Or like a whale?
Polonius: Very like a whale.
—William Shat&speare, Hamlet, dct 3, scene 2
Hamlet's interview. . . approximates the threefold ideal of being interpreted, validated and communicated. . . .
The interview appears as a display of the power relations at a royal court. . . .
Hamlet's interview may... be seen as an illustration of a pervasive doubt about the appearance of the world. [Or, we would like to add, the interview can emerge as an example of a negotiated text.}
—Kvale, InterViews, 1 996
a.    Asking questions and getting aiiswers is a much harder task than it
may seem at first. The spoken or written word has always a residue
of ambiguity, no matter how carefully we word the questions and how
carefully we report or code the answers. Yet interviewing is one of the
6'
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
JIM!'
most common and powerful ways in which we try to understand our fellow human beings. Interviewing includes a wide variety of forms and a multiplicity of uses. The most common form of interviewing involves individual, face-to-face verbal interchange, but interviewing can also take the form of face-to-face group interchange, mailed or self-administered questionnaires, and telephone surveys. It can be structured, semistructured, or unstructured. Interviewing can be used for marketing research, political opinion polling, therapeutic reasons, or academic analysis. It can be used f£>r the purpose of measurement or its scope can be the understanding of an individual or a group perspective. An interview can be a one-time, brief event—say, 5 minutes over the telephone—or it can take place over multiple, lengthy sessions, at times spanning days, as in life history interviewing.                                                                                              ;.
The use of interviewing to acquire information is so extensive today that it has been said that we live in an "interview society" (Atkinson & Silverman, 1997; Silverman, 1993). Increasingly, qualitative researchers are realizing that interviews are not neutral tools of data gatheting but active interactions between two (or more) people leading to negotiated, contextually based results. Thus the focus of interviews is moving to encompass the hows of people's lives (the constructive work involved in producing order in everyday life) as well as the traditional whats (the activities of everyday life) (Cicourel, 1964; Dingwall, 1997; Gubrium 8c Holstein, 1997, 1998; Holstein 6í Gubrium, 1995; Kvale, 1996; Sarup, 1996; Seídman, 1991; Silverman, 1993, 1997a).
In this chapter, after discussing the interview society, we examine interviews by beginning with structured methods of interviewing and gradually moving to more qualitative types, ending with interviews as negotiated texts. We begin by briefly outlining the history of interviewing, then we turn to a discussion of the academic uses of interviewing. Although the focus of this volume is qualitative research, in order to demonstrate the full import of interviewing, we need to discuss the major rypes of interviewing (structured, group, and unsrructured) as well as other ways to conduct interviews. A caveat: In discussing the various interview methods, we use the language and rationales employed by practitioners of these methods; we note our differences with these practitioners and our criticisms later in the chapter, in our discussion of gendered and other new types of qualitative interviewing. Following our examination of structured interviewing, we address in detail the various elements of qualitative interviewing. We then discuss the problems related to gendered interview-
62
The interview: From Structured Questions to Negotiated Text
ing as well as issues of interpretation and reporting, and we broach some considerations related ro erhical issues. Finally, we note some of the new trends in qualitative i n tet vi e wing.
♦ The Interview Society
Before embarking on our journey through interviewing per se, we want to comment briefly on the tremendous reliance on interviewing in U.S. society today, which has reached such a level that a number of scholars have referred to the United States as "the interview society" (Atkinson 6i Silverman, 1997; Silverman, 1993). Both qualitative and quantitative researchers tend to rely on the interview as rhe basic method of data gathering, whether the purpose is to obtain a rich, in-depth experiential account of an event or episode in the life of the respondent or to garner a simple point on a scale of 2 to 10 dimensions. There is inherent faith that the results are trustworthy and accurate and that the relation of the interviewer to respondent that evolves in the interview process has not unduly biased the account (Atkinson & Silverman, 1997; Silverman, 1993). The commitment to and reliance on the interview to produce narrative experience reflects and reinforces the view of the United States as an interview society. It seems that everyone, not just social researchers, relies on the interview as a source of information, wirh rhe assumption that intetviewing results in true and accurate pictures of respondents' selves and lives. One cannot escape being interviewed; interviews are everywhere, in the forms of political polls, questionnaires about doctor's visits, housing applications, forms regarding social service eligibility, college applications, talk shows, news programs—the list goes on and on. The interview as a means of data gathering is no longer limited to use by social science researchers or police detectives; it is a "universal mode of systematic inquiry" (Holstein St Gubrium, 1995, p. 1). It seems that almost any type of question-personal, sensitive, probing, upsetting, accusatory—is fair game and permissible in the interview setting. Almost all interviews, no matter their purposes (and rhese can be varied—to describe, to interrogate, to assist, to test, to evaluate), seek various forms of biographical description. As Gubrium and Holstein (1998) haveíioted, the interview has become a means of contemporary storytelling, where persons divulge life accounrs in response to interview inquiries. The media have been especially adept at using this technique.
03
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
As a society we rely on the interview and by and large take it for granted. The interview and the norms surrounding the enactment of the tespondent and teseatcher roles have evolved to the point where they are institutionalized and no longer require extensive training; rules and roles are known and shared. However, there is a growing group of individuals who increasingly question the traditional assumptions of the interview— we address their concerns in our later discussion of gendered interviewing and new trends in interview. Many ptactitioners continue to use and take * for granted traditional interviewing techniques. It is as if interviewing is \ now part of the mass culture, so that it has actually become the most feasible mechanism fot obtaining information about individuals, groups, and organizations in a society characterized by individuation, diversity, and specialized role relations. Thus, many feel that it is not necessary to re-invent the wheel for each interview situation, as "inteiviewing has become a routine technical practice and a pervasive, take n-for-granted activity in our culture" (Mishler, 1986, p. 23).
This is not to say, however, that the interview is so technical and the ptocedures so standardized that interviewers can ignore contextual, societal, and interpersonal elements. Each interview context is one of interaction and relation; the result is as much a product of this social dynamic as it is a product of accurate accounts and teplies. The interview has become a routine, almost unnoticed, part of evetyday life. Yet response rares continue to decline, indicating that fewer people are willing to disclose their "selves" or that they are so overburdened by requests for interviews that they are becoming more selective regarding which interviews to grant. Social scientists are more likely to tecognize, however, that interviews are inreractional encounters and that the nature of the social dynamic of the interview can shape the nature of the knowledge generated. Interviewers with less training and experience than social scientists may not recognize that intetview participants ate "actively" constructing knowledge around questions and responses (Holstein 6c Gubrium, 1995).
We turn now to a btief history of interviewing to frame its roots and development.
♦ The History of Interviewing
One form of interviewing or another has been with us for a very long time. Even ancient Egyptians conducted population censuses {Babbie, 1992). In
64
The interview: From Structured Questions to Negotiated Text
more recent times, the tradition of inteiviewing evolved from two trends. First, interviewing found great popularity and widespread use in clinical diagnosis and counseling, where the concern was with the quality of responses. Second, during World War I interviewing came to be widely employed in psychological testing; here the emphasis was on measurement (Maccoby & Maccoby, 1954).
The individual generally credited with being the first to develop a social survey relying on interviewing was Charles Booth (Converse, 1987). In 1886, Booth embarked on a comprehensive survey of the economic and social conditions of the people of London, published as Life and Labour of the People in London (1902-1903). In his early study, Booth embodied what were to become separate interviewing methods, because he not only implemented survey research but triangulated his work by relying on unstructured intetviews and ethnogtaphic observations:
The data were checked and supplemented by visits to many neighborhoods, streets and homes, and by conferences with various welfare and community leaders. From time to time Booth lived as a lodger in districts where he was not known, so that he could become more intimately acquainted with the lives and habits of the poorer classes (Parten, 1950, pp. 6-7)
Many other surveys of London and other English cities followed, patterned after Booth's example. In the United States a similar pattern ensued. Among others, an 1895 study attempted to do in Chicago what Booth had done in London (see Converse, 1987), and in 1896, self-admittedly following Booth's lead, the American sociologist W E. B. Du Bois studied the black population of Philadelphia (see Du Bois, 1899). Surveys of cities and small towns followed, most notable among them R. S, LyndandH. M. LynďsMiddletousn (1929) andMiddletown in Transition (1937).
Opinion polling was another early form of inteiviewing. Some polling took place well before the start of the 20th century, but it really came into its own in 1935 with the formation of the American Institute of Public Opinion by George Gallup. Preceding Gallup, in both psychology and sociology in the 1920s thete was a movement toward the study (and usually measurement) of attitudes. W I. Thomas and Florian Znaniecki used the documentary method to introduce rhe study of attitudes in social psychology. Thomas's influence, along with that of Robert Paik, a former reporter who believed sociology was to be found out in the field, spaiked a
05
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
number of community studies at the University of Chicago that came to be known collectively as the works of the Chicago school. Many other researchers were also greatly influential, such as Albion Small, George H, Mead, E. W Burgess, Everett C. Hughes, Louis Wirth, W Lloyd Warner, and Anselm Strauss (for a recent discussion of the relations and influence of various Chicagoans, see Becker, 1999).
Although the members of the Chicago school are reputed to have used the ethnographic method in their inquiries, some disagree, and have noted • that many of the Chicago school studies lacked the analytic component of '_,. modern-day ethnography, and so were, at best, "firsthand descriptive studies" (Harvey, 1987, p. 50). Regardless of the correct label for the Chicagoans' fieldwork, they clearly relied on a combination of observation, personal documents, and informal interviews in theit studies. Interviews weie especially in evidence in the work of Thrasher (1927/1963), who in his study of gang members relied primaiily on about 130 qualitative intet-views, and in that of Nels Anderson (1923), whose classic study of hoboes teiied on informal, in-depth conversations.
It was left to Herbert Blumer and his former student Howard Becker to formalize and give impetus to sociological ethnogtaphy in the 1950s and 1960s, and interviewing began to lose both the eclectic flavor given to it by Charles Booth and the qualitative accent of the Chicagoans. Understanding gang members or hoboes through interviews lost importance; what became televant was the use of interviewing in survey research as a tool to quantify data. This was not new, as opinion polls and market tesearch had been doing it for years. But duting Wotld War II rhete was a tremendous increase in survey research as the U.S. armed forces hired great numbers of sociologists as sutvey researchets. More than half a million American soldiers were interviewed in one manner or anothet (Young, 1966), and their mental and emotional lives were reported in a four-volume survey titled Studies in Social Psychology in World War II, the first two volumes of which were directed by Samuel Stouffet and titled The American Soldier. This work had tremendous impacr and led the way to widespread use of systematic survey research.
What was new, however, was that quantitative sutvey research moved into academia and came to dominate sociology as the method of choice for the next three decades. An Austrian immigrant, Paul Lazatsfeld, spearheaded this move. He welcomed The American Soldier with great enthusiasm. In fact, Robert Merton and Lazatsfeld (1950) edited a book of reflections on The American Soldier. Lazatsfeld moved to Columbia in 1940,
66
Mt
Trie interview; From Structured Questions to Negotiated Text ■-,
taking with him his market reseatch and other applied grants, and became instrumental in the directing of the Bureau of Applied Social Reseatch. Two other "survey organizations" were also formed: one in 1941, by Harry Field, the National Opinion Research Center, first at Denver and then at Chicago; and one in 1946, by Likert and his group, the Survey Research Center ar Michigan.
Academia at the time was dominated by theoretical concerns, and there was some resistance toward this applied, numbers-based kind of sociology. Sociologists and other humanists were critical of Lazarsfeld and the other survey researchers. Herbert Blumer, C. Wright Mills, Arthur Schlesinger, Jr., and Pitirin Sorokin, among others, voiced their displeasure. According to Converse (1987), Sotokin felt that "the new emphasis on quantitative wo t k was obsessive, and he called the new practitioners 'quantophrenics'—with special reference to Stouffer and Lazarsfeld" (p. 253). And Converse quotes Mills: "Those in the grip of the methodological inhibition often tefuse to say anything about modem society unless it has been through the fine little mill of the Statistical Ritual" (p. 252). Schlesinger, Converse notes, called the survey researchers "social relations hucksters" (p. 253).
But the survey researchers had powerful allies also, such as Merron, who joined the Survey Center at Columbia in 1943, and government moneys were becoming increasing available fot sutvey research. The 1950s saw a growth of survey research in the universities and a proliferation of survey research texts. Gradually, survey research increased its domain ovet sociology, culminating in 1960 with the election of Lazatsfeld to the ptesidency of the Ametican Sociological Association. The methodological dominance of survey research continued unabated through the 1970s, 1980s, and 1990s, although othet methods began to erode the ptomi-nence of survey methods.
Qualitative interviewing continued to be practiced, hand in hand with participant observation methods, but it too assumed some of the quantifiable scientific tigot that so pteoccupied sutvey tesearch. This was especially visible in grounded theory (Glaset &c Strauss, 1967), with its painstaking emphasis on coding data, and in eth no methodology, with its quest for invariant properties of social action (Cicourel, 1970). Other qualitative researchers suggested variations. John Lofland (1971) criticized grounded theory for paying little attention to data gathering techniques, Jack Douglas (1985) suggested lengthy, existential one-on-one interviews lasting one or more days, and James Spradley (1980) ttied to
■57
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
clarify the difference between ethnographic observation and ethnographic interviewing.
Recently, postmodernist ethnographers have concerned themselves with some of the assumptions present in interviewing and with the controlling role of the interviewer. These concerns have led to new directions in qualitative interviewing focusing on increased attention to the voices of the respondents (Marcus & Fischer, 1986), the interviewer-respondent relationship (Crapanzano, 1980), the importance of the researcher's gen-
*der in interviewing (Gluck & Parai, 1991), and the roles of other elements,
/■such as race, social status, and age (Seidman, 1991).
♦ Structured Interviewing
In structured interviewing, rhe interviewer asks all respondents the same series of preesrablished quesrions with a limited set of response categories. There is generally little room for variarion in responses except where open-ended questions (which are infrequent) may be used. The interviewer records the responses according to a coding scheme that has already been established by the project director or research supervisor. The interviewer controls the pace of the interview by treating the questionnaire as if it were a theatrical script to be followed in a standardized and straightforward manner. Thus all respondents receive the same set of questions asked in the same order or sequence by an interviewer who has been trained to treat all interview situations in a like manner. There is very little flexibility in the way questions are asked or answered in the structured interview setting. Instructions to interviewers often include some of the following guidelines:
♦   Never get involved in long explanations of the study; use the standard explanation provided by the supervisor.
♦   Never deviate from the study introduction, sequence of questions, or question wording.
♦  Never let another person interrupt the interview; do not let another person answer for the respondent or offer his or het opinions on the question.
♦   Never suggest an answer or agree or disagree with an answer. Do not give the respondent any idea of your personal views on the topic of the question or the survey.
68
The tnterview:Trom Structured Questions to Negotiated Text   ^
♦   Never interpret the meaning of a question; just repeat the question and give instructions or clarifications that are provided in training or by the supervisors.
♦   Never improvise, such as by adding answer categories or making wording changes.
Interviews by telephone, face-to-face interviews in respondents' households, intercept interviews in malls and parks, and interviews generally associared with survey research are most likely to be included in the structured interview caregory.
This interview context calls for the interviewer to play a neutral role, never interjecting his or her opinion of a respondent's answer. The interviewer must establish what has been called "balanced rapport"; he or she must be casual and friendly on the one hand, bur directive and impersonal on the other. The interviewer must perfect a style of "interested listening" that rewards the respondent's participation but does not evaluate the responses (Converse & Schuman, 1974).
In a structured interview, hopefully, nothing is left to chance. However, response effects, or nonsampling errors, that can be attributed to the questionnaire administration process commonly evolve from three sources. The first of these is respondent behavior. The respondent may deliberately try to please the interviewer or to prevent the interviewer from learning something about the respondent. In order to do this, the respondent may embellish a response, give what is described as a "socially desirable" response, or omit certain relevant information (Bradburn, 1983, p. 291). The respondent may also err due to faulty memory. The second source of error is found in rhe narure of the task: the method of questionnaire administration (face-to-face or Telephone) or rhe sequence or wording of the questions. The third source of error is the interviewer, whose characteristics or questioning techniques can impede proper communication of the questions (Bradburn, 1983). It is the degree of error assigned to the interviewer that is of greatest concern.
Most strucrured interviews leave little room for the interviewer to improvise or exercise independent judgment, but even in the most structured interview situation not every contingency can be anticipated, and not every interviewer behaves according to the script (Bradburn, 1983; Frey, 1989). In fact, one study of interviewer effects found rhat interviewers changed the wording to as many as one-third of the questions (Bradburn, Sudman, & Associates, 1979).
69
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
In general, research on interviewer effects has sľĽown interviewer characteristics such as age, gender, and interviewing experience to have relatively small impact on responses (Singer Sc Presser, 1989). However, there is some evidence that student interviewers produce larger response effects than do nonstudents, higher-status interviewers produce larger response effects than do lower-status interviewers, and the race of an interviewer makes a difference only on questions specifically related to race (Bradburn, 1983; Hyman, 1954; Singer, Frankel, ÔC" Glassman, ,    1983).
The relatively minor impact of the interviewer on response quality in structured interview settings is directly attributable to the inflexible, standardized, and predetermined nature of this type of interviewing. There is simply little room for error. However, those who are advocates of structured interviewing are not unaware rhat the interview takes place in a social interacrion context and that it is influenced by that context. Good interviewers recognize this fact and are sensitive to how interaction can influence responses. Converse and Schuman (1974) observe, "There is no single inrerview style that fits evety occasion or all respondents" (p. 53). This means that interviewers must be aware of respondent differences and must be able to make the proper adjustments called for by unanticipated developments. As Raymond Gordeu (1992) states, "Intetviewing skills are not simple motor skills like riding a bicycle: rather, they involve a high-order combination of observation, empathic sensitivity, and intellectual judgment" (p. 7).
It is not enough to understand the mechanics of interviewing, it is also important to understand the respondent's world and forces that might stimulate or retard response (Kahn & Cannel, 1957). Still, the structured interview proceeds under a stimulus-response format, assuming that the respondent will truthfully answer quesrions previously determined to reveal adequate indicators of the variable in question, as long as those questions are properly phrased. This kind of interview often elicits rational responses, but it overlooks or inadequarely assesses the emotional dimension.
#'Group Interviews
The group interview is essentially a qualitative data gathering technique (see Madriz, Chapter 10, this volume) that relies upon rhe systematic
70
The Interview: From Structured Questions to Negotiated Text ■->
questioning of several individuals simultaneously in a formal or informal setting. Thus this technique straddles the line between formal and informal interviewing.
■ The use of the group interview has ordinarily been associated with marketing research under rhe label of focus group, where the purpose is to gather consumer opinion on product characteristics, advertising themes, or service delivery. This format has also been used to a considerable extent by political parties and candidates who are interested in voter reaction to issues and policies. The group interview has also been used in sociological research. Bogardus used it to test his social distance scale in 1926, Zuckerman (1972) interviewed Nobel laureates, Thompson and Demerath (1952) looked at management problems in the1 military, Morgan and Spanish (1984) studied health issues, we investigated older-worker labor force reentry (Fontana & Frey, 1990), and Merton and his associates studied the impact of propaganda using group interviews (see Frey Sc Fontana, 1991). In fact, Merton, Fiske, and Kendall (1956) coined the term focus group to apply to a situation in which the researcher/ interviewer asks very specific questions about a topic after having already completed considerable research. There is also some evidence that established anthropologists such as Malinowski used this technique, although they did not report it (Frey & Fontana, 1991). Today, all group interviews are often generically designated focus group interviews, even though there are considerable variations in the natures and types of group interviews.
In a group interview, the interviewer/moderator directs the inquiry and the interaction among respondents in a very structured fashion or in a very unsrructured manner, depending on the interview's purpose. The purpose may be exploratory; for example, the researcher may bring several persons together to test a methodological technique, to try out a definition of a research problem, or to identify key informants. An extension of this exploratory intent is the use of the group interview for the purpose of pretesting questionnaire wording, measurement scales, or other elements of a survey design. This is now quite common in survey research (Desvousges & Frey, 1989). Group interviews can also be used successfully to aid respondents' recall of specific events or to stimulate embellished descriptions of events (e.g., a disaster or a celebration) or experiences shared by members of a group. Group interviews can also berused for triangulation purposes or can be used in conjunction with other data gathering techniques. For example, group interviews could be helpful in the process of "indefinite triangulation," by putting individual responses into a conrexr (Cicourel,
71
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
f 974). Finally, phenomenological purposes may be served whether group interviews are the sole basis for gathering data or they are used in association with other techniques.
Group interviews can take different forms depending on their purposes. They can be brainstorming sessions with little or no structure or direction from the interviewer, or they can be very structured, as in nominal, Delphi, and marketing focus groups. In the laner cases the role of the interviewer is very prominent and directive. Fieldwork settings provide Ttoth formal and informal occasions for group interviews. The field researcher can bring respondents into a formal setting in the field context and ask very directed questions, or a natural field setting, such as a street corner or a neighborhood tavern, can be a conducive setting for casual but purposive inquiries.
Group interviews can be compared on several dimensions. First, the interviewer can be very formal, taking a very directive and controlling posture, guiding discussion strictly, and not permitting digression or variation from topic or agenda. This is the mode of focus and nominal/Delphi groups. In the latter case participants are physically isolated but share views through a coordinator/interviewer. The nondirective approach is more likely to be implemented in naturally established field settings, such as a street corner, or in controlled settings (e.g., research labs) where the research purpose is phenomenological, to establish the widest range of meaning and interpretation for the topic. Groups can also be differentiated by question format and purpose, which in the case of gtoup inter-views usually means exploration, pretest, or phenomenological. Exploratory interviews are designed to establish familiarity with a topic or setting; the interviewer can be very directive (or the opposite), but the questions are usually unsttuctured or open-ended. The same format is used in interviews with phenomenological purposes, where the intent is to tap intersubjective meaning with depth and divetsity. Pretest interviews are generally structured in question format and the interviewer is directive in style. Table 2.1 compates the types of group intetviews on various dimensions.
The skills that are required to conduct the gtoup interview are not significantly different from those needed fot individual interviews. The interviewer must be flexible, objective, empathic, persuasive, a good listener, and so on. But the group format does present some problems not found in the individual interview. Merton et al. (1956) note three specific
72
The Interview: From Structured Questions to Negotiated Text ■-.
TABLE 2.1	Types of Group Interviews and Dimensions			
Type	Setting	Rote of Interviewer	Question Format	Purpose
Focus group	formal-preset	directive	structured	exploratory pretest
Brainstorming	formal or informal	nondirective	very structured	exploratory
Nominal/ delphi	formal	directive	structured	pretest exploratory,
Field, natural	informal spontaneous	moderately nondirective	very structured	exploratory phenomenological
Field, formal	preset, but in field	somewhat directive	semi-structured	phenomenological
problems: First, the interviewer must keep one person or small coalition of persons from dominating rhe group; second, the interviewer must encourage recalcitrant respondents to paiticipate; and third, the intet-viewer must obtain responses from the entire group to ensure the fullest coverage of the topic. In addition, the interviewer must balance the directive, interviewer role with the tole of moderator, which calls for rhe management of the dynamics of the gtoup being interviewed; the group interviewer must simultaneously worry about the script of questions and be sensitive to the evolving patterns of group interaction.
Gtoup interviews have some advantages over individual interviews: They are relatively inexpensive to conduct and often produce rich data that are cumulative and elaborative; they can be stimulating for respondents, aiding recall; and the format is flexible. Group interviews are not, however, wirhour problems: The results cannot be generalized; rhe emerging group culture may interfere with individual expression, and the group may be dominated by one person; antr'group think" is a possible outcome. The requirements for interviewer skills are greater than those fot individual interviewing because of the group dynamics that are present. In
73
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
j nív addition, it is difficult to research sensitive topics using this technique.
Nevertheless, the group interview is a viable option for both qualitative
and quantitative research.
♦ Unstructured Interviewing
Unstructured interviewing can provide a greater breadth of data than the t- other types, given its qualitative nature. In this section we discuss the tradi--f: tional rype of unstructured interview: the open-ended, ethnographic (in-depth) interview. Many qualitative researchers differentiate between in-depth (or ethnographic) inrerviewing and participant observation. Yet, as Lofland (1971) points out, the two go hand in hand, and many of the data gathered in participant observation come from informal interviewing in the field. Consider the following report, from Malinowski's (1967/1989) diary:
Saturday 8 [December 1917]. Got up late, felt rotten, took enema. At about 1 I went out; I heard cries; [people from] Kapwapu were bringing uri to Teyava. I sat with the natives, talked, took pictures. Went back. Billy corrected and supplemented my notes about wasi. AtTeyava, an old man talked a great deal about fishes, but í did not understand him too well. Then we moved to his bwayama. Talked about iili'u. They kept questioning me about the war—In the evening I talked to the policeman about bwaga 'u, Iili'u and yoyova. I was irtitated by their laughing. Billy again told me a number of interesting things. Took quinine and calomel, {p. 145)
Malinowski's "day in the field" shows how very important unstructured interviewing is in the conduct of fieldwork and clearly illustrates the difference between sttuctured and unstructured intetviewing. Malinowski has some general topics he wishes to know about, but he does not use closed-ended questions or a formal approach to intetviewing. What's more, he commits (as most field-workers do) what structured interviewers would see as two "capital offenses": (a) He answers questions asked by the respondents, and (b) he lets his petsonai feelings influence him (as all field-workers do), thus he deviates from the "ideal" of a cool, distant, and rational interviewer.
Malinowski's example captures the differences between structured and unstructured interviewing: The formet aims at capturing precise data of a
74
Trie Interview: From Structured Questions to Negotiated Text
codab 1 e natute in order to explain behavior within preestablished categories, whereas the latter attempts to understand the complex behavioi of members of society without imposing any a priori categorizarion that may limit the field of inquiry.
In a way, Malinowski's interviewing is still structured to some degiee— that is, there is a setting, there are identified informants, and the respondents are clearly discernible. In othet types of interviewing there may be no setting; for insrance, Rosanna Hertz (1995,1997b, 1997c) focused on locating women in a historic moment rather than in a place. Additionally, in their study of single mothets, Hertz and Ferguson (1997) interviewed women who did not know each other, who were not part of a single group or village. Ar times, informants are not readily accessible or identifiable, but anyone the researcher meets may become a valuable source of information. Hertz and Ferguson relied on tradespeople and friends to identify single mothers for their study. Fontana and Smith (1989) found that respondents are not always teadily identifiable. In studying Alzheimer's disease patients, they discovered it was o f re n possible to confuse caregivers and patients in the early stages of the disease. Also, in Fontana's (1977) teseatch on poor elderly, he had no fixed setting at all; he simply wandered from bench ro bench in the park where the old folks were sitting, talking to any disheveled old person who would talk back.
Spradley (1979) aptly differentiates among vaiious types of intetviewing. He desctibes the following interviewet-respondent inreraction, which would be unthinkable in traditional sociological circles yet is the very essence of unsrructured interviewing—the establishment of a human-to-human relation with the respondent and the desire to understand father than to explain:
Presently she smiled, pressed her hand to her chest, and said: "Tsetchwe."It was her name. "Elizabeth," I said, pointing to myself. "Nisabe," she answered. . . . Then, having surely suspected that I was a woman, she put her hand on my breast gravely, and, finding out that I was, she touched her own breast. Many Bushmen do this; to them all Europeans look alike. "Tasu si" (women), she said. Then after a moment's pause Tsetchwe began to teach me. (pp. 3-4)
/
Spradley goes on to discuss all the things an interviewer learns from the natives about them, theit culture, their language, their ways of life. Although each and every study is different, these are some of the basic
75
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
elements of unstructured intetviewing. These elements have been discussed in details alteady, and we need not elaborate upon them too much (for detailed accounts of unstructured interviewing, see, among others, Adams & Preiss, 1960; Denzin, 1989b; Lofland, 1971; Spradley, 1979). Here we provide brief synopses. Please remember that these are presented only as heuristic devices; evety study uses slightly different elements and often in different combinations.
Later in this chapter, in discussing new trends, we will deconstruct these * notions as we frame the interview as an active, emeigent process. We con-' tend that our interview society gives people instructions on how to comply with these heuristics (see Silverman, 1993, 1997a, 1997b). Similarly, James Scheurich (1997} is openly critical of both positivistic and interpretive intetviewing, as they are both based on modernist assumptions. Rather than being a process "by the numbers," for Scheurich, interviewing (and its language) are "persistently slippery, unstable, and ambiguous from person to person, from situation to situation, from time to time" (p. 62).
Accessing the setting. How do we "get in"? That, of course, varies according to the group one is attempting to study. One may have to disrobe and casually stroll in the nude if doing a study of nude beaches (Douglas Sc Rasmussen, 1977), or one may have to buy a huge motorbike and frequent seedy bars in certain locations if attempting to befriend and study the Hell's Angels (Thompson, 1985). The different ways and attempts to "get in" vary tremendously, but they all share the common goal of gaining access to the setting. Sometimes there is no setting per se, as when Fontana (1977) attempted to study poot elderly on the streets and had to gain access anew with each and every interviewee.
Understanding the language and culture of the respondents. Rosalie Wax (1960) gives perhaps the most poignant description of learning the language and culture of the respondents in her study of "disloyal" Japanese in concentration camps in the United States between 1943 and 1945. Wax had to overcome a number of language and cultural problems in her study. Although respondenrs may be fluent in the language of the interviewer, there are different ways of saying things and, indeed, certain things that should not be said at all, linking language and cultural manifestations. Wax makes this point:
76
tyt
The Interview: From Structured Questions to Negotiated Text
I remarked that í would like to see the letter. The silence chat fell on the chatting group was almost palpable, and the embarrassment of the hosts was painful to see. The faux pas was not asking to see a Letter, for ietrers were passed about rather freely. It rested on the fact that one did not give a Caucasian a letter in which the "disloyal" statement of a friend might be expressed, (p. 172)
Some researchers, especially in anthropological interviews, tend to rely on interpreters, and thus become vulnerable to added layers of meanings, biases, and interpretations, which may lead to disastrous misunderstandings (Freeman, 1983). At times, specific jargon, such as the medical metalanguage of physicians, may be a code that is hard for nonmimbers to understand.
Deciding on how to present oneself. Do we present ourselves as representatives from academia studying medical students (Becker, 1956)? Do we approach the interview as a woman-to-woman discussion (Spradley, 1979)? Do we "dress down" to look like the respondents (Fontana, 1977; Thompson, 1985)? Do we represent the colonial culture (Malinowsky, 1922), or do we humbly present ourselves as "learners" (Wax, I960)? This decision is very important, because once the interviewer's presentational self is "cast," it leaves a ptofound impression on the respondents and has great influence over the success (or lack of it) of the study. Sometimes, inadvertently, the researcher's presentational self may be misrepresented, as John Johnson (1976) discovered in studying a welfare office, when some of the employees assumed he was a "spy" for management despite his best efforts to the contrary.
Locating an informant. The researcher must find an insider, a member of the group studied, who is willing to be an informant and act as a guide and a translator of cultural mores and, at times, jargon or language. Although the researcher can conduct interviews without an informant, he or she can save much time and avoid mistakes if a good informant becomes available. The "classic" sociological informant is Doc in William Foote Whyte's Street Corner Society (1943). Without Doc's help and guidance, it is doubtful that Whyte would have been able to access his subjects at the level he did. Very instructive is Paul Rabinow's (1977) discussion of bis relationship with his main infotmant, Abd al-Malik ben Lahcen. Malik acted as a
77
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
»HO' translator but also provided Rabinow with access to the cultural ways of
the subjects, and by his actions provided Rabinow with insights into the
vast differences between a University of Chicago researcher and a native
Moroccan.
Gaining trust. Survey researchers asking respondents whether they would or would not favor the establishment of a nuclear dump in their state (Frey, 1993) do not have too much work to do in the way of gaining trust; J respondenrs have opinions about nuclear dumps and are very willing to ■t: express them, sometimes forcefully. But it is clearly a different story if one wants to ask about a person's frequency of sexual intercourse or preferred method of birth control. The intetviewet needs to establish some trust with the respondents (Gcourel, 1974). Paul Rasmussen (1989) had to spend months as a "wallflower" in the waiting room of a massage parlor before any of the masseuses gained enough trust in him to divulge to him, in unsttuctured interviews, the nature of their "massage" relations with clients. Gaining trust is essential to the success of the interviews and, once gained, trust can still be very fragile. Any faux pas by the researcher may destroy days, weeks, or months of painfully gained trust.
Establishing rapport. Because the goal of unstructured interviewing is understanding, it is paramounr rhat the researcher establish rapport with respondents; that is, the researcher must be able to take the role of the respondents and attempt to see the situation from their viewpoint, rather than superimpose his or her world of academia and preconceptions upon them. Although a close rapport with the respondents opens the doors to more informed research, it may create problems as the researcher may become a spokesperson for the group studied, losing his or her distance and objectivity, or may "go native" and become a member of the group and forgo his or her academic role. At times, what the researcher may feel is good rapport turns out not to be, as Thompson (1985) found out in a nightmarish way when he was subjected to a brutal beating by the Hell's Angels just as his study of them was coming to a close. At the other end of the spectrum, some researchers may never feel they have established rapport with their subjects. Malinowski (1967/1989), for example, always mistrusted the motives of the natives and at times was ttoubled by their brutish sensuality or angered by their outright lying or deceptions: "After lunch I [carried] yellow calico and spoke abour the baloma. I made a small sagali, Navavile. I was fed up with the niggers" (p. 154).
78
The interview: From Structured Questions to Negotiated Text ■■•
Collecting empirical materials. Being out in the field does not affotd researchers the luxury of video cameras, soundproof rooms, and high-quality recording equipment. Lofland (1971) provides derailed information on doing and writing up intetviews and on the types of field notes researchers ought to take and how to organize them. Yet field-workers often must make do; their "tales" of their methods range from holding a miniature tape tecorder as inconspicuously as possible to taking mental notes and then rushing to the privacy of a bathroom to jot notes down, on toilet papers at times. We agree with Lofland that regardless of the circumstances, researchers ought to (a) take notes regularly and promptly: (b) write everything down, no matter how unimportant it may seem at the time; (c) try to be as inconspicuous as possible in note taking; and (d) analyze their nores frequently.
♦ Other Types of Unstructured Interviewing
We consider the issues of interpreting and repotting empirical material later in this chapter In this section, we briefly outline some different rypes of unstructured interviews.
Oral History
The oral history differs from other unstructured interviews in purpose, but not methodologically. The oral collection of historical materials goes back to ancient times, but its modein-day formal organization can be traced to 1948, when Allan Nevins began the Oral History Project at Columbia University (Starr, 1984, p. 4). Oral history captures a variety of forms of life, from common folks talking about their jobs in Studs Terkel's Working (1975) to the historical recollections of presidenr Harry Truman in Merle Miller's Plain Speaking (1974; see Stan, 1984). Often, oral history transcripts are not published, but many may be found in libraries, silent memoirs waiting for someone to rummage through them and bring their testimony to life. Recently, oral history has found great popularity among feminists (Gluck Sc Patai, 1991), who see it as away to understand and bring forth the history of women in a culrure that has traditionally relied on masculine interpretation: "Refusing to be tendered historically voiceless any longer, women are creating a new history—using our own voices and experiences" (Gluck, 1984, p. 222).
79
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
•Tlí!'
Relevant to the study of oral history (and, in fa«, to all interviewing) is the srudy of memory and its relation to recall. For instance, Barry Schwartz (1999) has examined the ages at which we recall critical episodes in our lives; he concludes rhat "biographical memory ... is better understood as a social process" and that "as we look back, we find ourselves remembering our lives in terms of our experience with others" (p. 15; see also Schwartz, 1996). Carolyn Ellis (1991) has resorted to the use of S sociological introspection" to reconstruct biographical episodes of her past life. Notable among Ellis's works in this genre is her reconstruction of her 9-year relationship with her partner, Gene Weinstein, in which she describes the emotional negotiations the two of them went through as they coped with his downward-spiraling health, until the final negotiation with death (Ellis, 1995).
Creative Interviewing
Close to oral history, but used more conventionally as a sociological tool, is Jack Douglas's (1985) "creative interviewing." Douglas argues against "how-to" guides to conducting interviews because unstructured interviews take place in the largely situational everyday worlds of members of society. Thus interviewing and interviewers musr necessarily be creative, forget how-to rules, and adapt themselves to the ever-changing situations they face. Similar to oral historians, Douglas sees interviewing as collecting otal reports from the members of society. In creative interviewing, these reports go well beyond the length of conventional unstructured interviews and may become "life histories," with interviewing taking place in multiple sessions over many days wirh the subject(s).
Postmodern Interviewing
Douglas's concern with the imporrant role played by the interviewer qua human being, which is also shared by feminist oral historians, became a paramount element in the interviewing approaches of postmodern anthropologists and sociologists in the mid-1980s. Marcus and Fischer ($86) address ethnography at large, but their discussion is germane to unstructured intetviewing because, as we have seen, such interviewing constitutes the major way of collecting data in fieldwork. Marcus and Fischer voice reflexive concerns abour the ways in which the researcher influences the study, both in the merhods of data collection and in the
m
The Interview: From Structured Questions to Negotiated Text
techniques of reporting findings; this concern leads to new ways to conduct intemews, in rhe hope of minimizing, if not eliminating, interviewer influence. One such way is polyphontc inrerviewing, in which the voices of the subjects are recorded with minimal influence from the researcher and are nor collapsed together and reported as one, through the interpretation of the researcher. Instead, the multiple perspectives of the various subjects are reported and differences and problems encounrered are discussed rather than glossed over (see Krieger, 1983). Interpretive interactionism follows in the footsteps of creative and polyphonic interviewing, but, borrowing from James Joyce, adds a new element, that of epiphanies, which Denzm (1989a) describes as "those interacrional moments that leave marks on people's lives [and] have the potential for creating transformational experiences for the person" (p. 15). Thus the topic of inquiry becomes dramatized by rhe focus on existential moments in people's lives, hopefully ptoducing richer and more meaningful dara. Finally, as postmodernism seek new ways of undersranding and reporting data, we wish to note the concept of oratysis, which refers "to the ways in which oral forms, derived from everyday life, are, with the recording powers of video, applied to the analytical tasks associated with literate forms" (Ulmer, 1989, p. xi). In oralysis, the traditional product of interviewing, talk, is coupled with the visual, providing, according to Ulmer (1989), a product consonant with a society that is dominated by the medium of television.
♦ Gendered Interviews
The housewife goes into a well-stocked store to look for a frying pan. Her thinking probably does not proceed exactly this way, but it is helpful to think of the many possible two-way choices she might make: Cast iron or aluminum? Thick or thin? Metal or wooden handle? Covered or not? Deep or shallow? Large or small? This brand or that? Reasonable or too high in price? To buy or not? Cash or charge? Have it delivered or carry it. . . . The two-way question is simplicity itself when it comes to recording answers and tabulating them. (Payne, 1951, pp. 5S-S6)
i
The above quote represents the prevalent paternalisric attitude toward women in interviewing (see Oakley, 1981, p. 39) as well as the paradigmatic concern with coding answers and therefore with presenting limited,
81
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
dichotomous choices. Apart from a tendency to be condescending to women, the traditional interview paradigm does not account for gendered differences. In fact, Babbie's classic text The Practice of Social Research (1992) briefly references gender only three times and says nothing about the influence of gender on interviews. As Ann Oakiey (1981) cogently points out, both the interviewers and the respondents are considered faceless and invisible, and they must be if the paradigmatic assumption of gath-eting value-free data is to be maintained. Yet, as Denzin (1989a, p. 116) * tells us, "gender filters knowledge"; that is, the sex of the interviewer and '.'■■ that of the respondent do make a difference, as the intetview takes place within the cultural boundaries of a paternalistic social system in which masculine identities are differentiated ftom feminine ones.
In the typical interview there exists a hierarchical relation, with the respondenr being in the subordinate position. The interviewer is instructed to be courteous, friendly, and pleasant:
The interviewer's manner should be friendly, courteous, conversational and unbiased. He should be neither too grim nor too effusive; neither too talkative nor too timid. The idea should be to put the respondent at ease, so that be will talk freely and fully. (Selltiz, Jahoda, Deutsch, Sc Cook, 1965, p. 576; emphasis added)
Yet, as the last above-quoted line shows, this demeanor is a ruse to gain the trust and confidence of the respondent without reciprocating those feelings in any way. Interviewers are not to give their own opinions and are to evade direct questions. What seems to be a conversation is really a oneway psendoconversation, raising the ethical dilemma (Fine, 1983-1984) inherent in the study of people for oppoitunistic teasons. When the respondent is female, the interview presents added pioblems, because the pteestablished format directed at information relevant fot the study tends both to ignoie the tespondenťs own concerns and to curtail any attempts to digress and elaborate. This format also stymies any revelation of personal feelings and emotions.
Warren (1988) discusses problems of gender in both anthropological and sociological fieldwotk, and many of these are found as well in the ethnographic interview. Some of these problems are the ttaditional ones of entree and trust, which may be heightened by the sex of the interviewer, especially in highly sex-segregated societies: "I never witnessed any ceremonies that were barred to women. Whenever I visited compounds I sat
82
The Interview: From Structured Questions to Negotiated Text
with the women while the men gathered in the parlors or in front of the compound.... I nevet entered any of the places where men sat around to drink beer or palm wine and to chat" (Sudarkasa, 1986; quoted in Warren, 1988, p. 16).
Solutions to the problem have been to view the female anthropologist as androgyne ot to grant her honorary male status for the duration of her research, Warren (1988) also points to some advantages of a researcher's being female and therefore seen as harmless or invisible. Other problems are associated with the researcher's status and race and with the context of the interview; again, these problems are magnified for female researchers in a patemalisdc world. Female interviewers ar times face the added burden of sexual overtures or covert sexual hassle (Warren, 1988, p., 33).
Feminist researchers have suggested ways to circumvent the traditional interviewing paradigm. Oakley (1981) notes that interviewing is a masculine paradigm, embedded in a masculine culture and stressing masculine traits while ar rhe same time excluding traits such as sensitivity, emotionality, and others that are cultutally viewed as feminine trairs. There is, however, a growing reluctance, especially among female tesearchers (Oakley, 1981; Reinharz, 1992; Smith, 1987), to continue interviewing women as "objects," with little or no regard for them as individuals. Although this reluctance stems from moral and ethical reasons, it is also relevant methodologically. As Oakley (1981) points out, in interviewing there is "no intimacy without reciprocity" (p. 49). Thus the emphasis is shifting to allow the development of a closer relation between interviewer and respondent; researchers are attempting to minimize status differences and are doing away with the traditional hierarchical siruarion in inrerviewing. Interviewers can show their human side and answer questions and express feelings. Methodologically, this new approach provides a greater spectrum of responses and greater insight into the lives of respondents—or "participants," to avoid the hierarchical pitfall (Reinharz, 1992, p.22)— because it encourages them to control the sequencing and the language of the interview and also allows them the freedom of open-ended responses (Oakley, 1981; Reinharz, 1992; Smith, 1987). To wit: "Women were always . . . encouraged to 'digtess' into details of their personal histories and ro recount anecdotes of their working lives. Much important infor-mation was gathered in this way" (YSindle, 1984; quoted in Reinharz, 1992, p. 25).
Rosanna Hertz (1997a) makes the self of the researcher visible and suggests that it is only one of many selves the researcher takes to the field. She
83
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
asserts that interviewers need to be reflexive; that is, they need "to have an ongoing conversation about experience while simultaneously living in the moment" (p. viii). By doing so, they will heighten the understanding of differences of ideologies, cultures, and politics between interviewers and interviewees.
Hertz also underscores the importance of "voices"—how we, as authors, express and write our stories, which data we include and which we exclude, whose voices we choose to represent and which we do not. The concern with voices is also found, very powerfully, in a volume edited by Kim Marie Vaz titled Oral Narrative Research With Black Women (1997). One of the contributors, Christine Obbo (1997), states:
This chapter is a modest exercise in giving expression to women's voices and in rescuing their perceptions and experiences from being mere murmurs or backdrop to political, social and cultural happenings. Women's voices have been devalued by male chronicles of cultural history even when the men acknowledge female informants; they are overshadowed by the voice of male authority and ascendance in society, (pp. 42-43)
This commitment to maintaining the integrity of the phenomena and preserving the viewpoint of the subjects, as exptessed in their everyday language, is akin to the stand raken by phenomenological and existential sociologies (Douglas Sc Johnson, 1977; Korarba & Fontana, 1984) and also reflects the concerns of postmodern ethnographers (Marcus &c Fischer, 1986). The differences are (a) the heightened moral concern for subjects/participants, (b) the attempt to redress the male/female hierarchy and existing paternalistic power structure, and (c) the patamount importance placed upon membership, because the effectiveness of male researchers in interviewing female subjects has been largely discredited.
Ruth Behar (1996) addresses the ambiguous nature of the enterprise of interviewing by asking: Where do we locate the researcher in the field? How much do we reveal about ourselves? How do we reconcile our different roles and positions? Behar makes us see that interviewer, wtiter, respondent, and interview are not clearly distinct entities; rather, they are jntertwined in a deeply problematic way.
Some feminist sociologists have gone beyond concerns with interviewing or fieldwork in itself. Laurel Richardson (1992a) is striving for new forms of expression to report her findings and has presented some of her fieldwork in the form of poetry. Patricia Clough (1998) questions the
84
The Interview: From Structured Questions to Negotiated Text   '"-
whole enteiprise of fieldwork under the current paradigm and calls for a reassessment of the whole sociological enterprise and for a rereading of existing sociological texts in a light that is not marred by a paternalistic bias. Their voices echo the concern of Dorothy Smith (1987), who eloquently states:
The problem [of a research project] and its particular solution are analogous to those by which fresco painters solved the problems of representing the different temporal moments of a story in the singular space of the wall. The problem is to produce in a two-dimensional space framed as a wall a world of action and movement in time, (p, 281)
A growing number of researchers feel that we cannot isolate gender from other important elements that also "filter knowledge." Among others, Patticia Hill Collins (1990) has wtitten eloquently about the filtering of knowledge through memberships—of being black and female in American culture, in her case. Kath Weston (1998) makes just as powerful a case for sexuality, which, she contends, should not be tteated as a compartmentalized subspecialty, because it underlies and is integral to the whole of the social sciences. Clearly, gender, sexuality, and race cannot be considered in isolation; race, class, hierarchy, status, and age (Seidman, 1991) are all part of the complex, yet often ignored, elements that shape interviewing.
♦ Framing and Interpreting Interviews
Aside from the problem of framing real-life events in a two-dimensional space, we face the added problems of how the framing is being done and who is doing the framing. In sociological terms, this means that the type of interviewing selected, the techniques used, and the ways of recording information all come to bear on the results of the study. Additionally, data must be interpreted, and the researcher has a great deal of influence on what part of the data will be reported and how it will be reported.
Framing Interviews
/
Numerous volumes have been published on the techniques of structured interviewing (see, among others, Babbie, 1992; Bradburn et al., 1979; Gorden, 1980; Kahn 6c Cannel, 1957). There is also a voluminous
85
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
j im literature on group interviewing, especially in marketing and survey research (for a comprehensive review of literature in this area, see Stewarr Sc Shamdasani, 1990). The uses of group interviewing have also been linked to qualitative sociology (Morgan, 1988). Unstructured interviewing techniques have been covered thoroughly (Denzin, 1989b; Lofland, 1971; Lofland Sc Lofland, 1984; Spradley, 1979).
As we have noted, unstructured interviews vary widely, given their informal nature and depending on the nature of the setting, and some eschew the use of any preestablished set of techniques (Douglas, 1985). Yefthere are Techniques involved in interviewing whether the interviewer is just being "a nice person" or is following a format. Techniques can be varied to meet various situations, and varying one's techniques is known as using tactics. Traditionally, the researcher is involved in an informal : conversation with the respondent, thus he or she must maintain a tone of "friendly" chat while trying to remain close to the guidelines of the topic of inquiry he or she has in mind. The researcher begins by "breaking the ice" with general questions and gradually moves on to more specific ones, while also—as inconspicuously as possible—asking questions intended to check the veracity of the respondent's statements. The researcher should avoid gerting involved in a "real" conversation in which he or she answers questions asked by the respondent or provides personal opinions on the matters discussed. A researcher can avoid "getting trapped" by shrugging off the relevance of his or her opinions ("It doesn't matter how I feel, it's your opinion that's important") or by feigning ignorance ("I really don't know enough about this to say anything; you're the expert"). Of course, as we have seen in the case of gendered interviewing, the researcher may reject these techniques and "come down" to the level of the respondent to engage in a "real" conversation, with give-and-take and shared empathic understanding.
The use of language, particularly the use of specific terms, is important in the creation of a "sharedness of meanings" in which both interviewer and respondent understand the contextual nature of specific referents. For instance, in studying nude beaches, Douglas and Rasmussen (1977) dis-coverecUhat the term nude beach vtrgin had nothing to do with chastity; father, it referred to the fact that a person's buttocks were whire, thus indicating to others that he or she was a newcomer to the nude beach. Language is also important in delineating the type of question (broad, narrow, leading, instructive, and so on).
BÓ
The Interview: From Structured Questions to Negotiated Text
Nonverbal techniques are also important in interviewing. There are four basic modes of nonverbal communication:
f Toxemic communication is the use of interpersonal space to communicate attitudes, chronemics communication is the use of pacing of speech and length of silence in conversation, kinesic communication includes any body movements or postures, and paral'mguistic communication includes all the variations in volume, pitch and quality of voice. (Gorden, 1980, p. 335)
All four of these modes represent important Techniques for the researcher; in addition, rhe researcher should carefully note and record respondents' uses of these modes, for inrerview dara are more than verbal records and should include, as much as possible, nonverbal features of the interaction. Finally, techniques vary with the groups being interviewed; for insrance, interviewing a group of children requires a different approach from the one an interviewer might use when interviewing a group of elderly widows (Lopata, 1980).
Interpreting Interviews
Many studies using unstructured interviews are not reflexive enough about the interpreting process; common platitudes proclaim that the data speak for themselves, that the researcher is neurral, unbiased, and "invisible." The data reported tend to flow nicely, there are no contradictory data and no mention of what data were excluded and/or why. Improprieties never happen and the main concern seems to be the proper, if unreflexive, filing, analyzing, and reporting of events. But anyone who has engaged in fieldwork knows better; no matter how organized the researcher may be, he or she slowly becomes buried under an increasing mountain of field notes, transcripts, newspaper clippings, and audiotapes. Traditionally, readers were presented with the researcher's interpretation of the data, cleaned and streamlined and collapsed in rational, non-conrradictory accounts. More recently, sociologists have come to grips with the reflexive, problematic, and, at times, contradictory nature of data and with the tremendous, if unspoken, influence of the researcher as author. "What Van Maanen (1988) calls "confessional style" began in earnest in the 1970s (see Johnson, 1976) and continues unabated to our day, in a soul cleansing by researchers of problematic feelings and sticky situations in the field. Although perhaps somewhar overdone at times, these
87
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
iftO'
"confessions" are very valuable, as they make the readers aware of the complex and cumbersome nature of interviewing people in their natural settings and lend a tone of realism and veracity to studies. For example: "Yestetday I slept vety late. Got up around 10. The day before I had engaged Omaga, Koupa, and a few others. They didn't come. Again I fell into a rage" (Malinowski, 1967/1989, p. 67).
Showing the human side of the researcher and the ptoblematics of
unstructured interviewing has taken new forms in deconstructionism
*(Derrida, 1976). Here the influence of the authoi is brought under scru-
-'tiny. Thus the text created by the researcher's rendirion of events is
"deconstructed"; the authot's biases and taken-for-gr anted notions are
exposed, and, at times, alternative ways to look at the data are introduced
(Clough, 1998).                                                                                       i.s-i
Postmodern social teseatchers, as we have seen, attempt to expose and
minimize the role of the researcher qua field-worker and qua author.
Thus,  for  instance,  Ciapanzano  (1980)  reports Tuhami's  accounts,
whethet they be soci o historical renditions, dreams, or outrighr lies,
because they all constitute a part of this Morrocan Arab s.ubjecr's sense of
self and personal history. In interviewing Tuhami, Crapanzano learns not
only about his subject but about himself:
As Tuhami's interlocutor, I became an active participant in his life history, even though I rarely appear directly in his recitations. Not only did my presence, and my questions, prepare him for the text he was to produce, but they produced what I read as a change of consciousness in him. They produced a change of consciousness in me too. We were both jostled from our assumptions about the nature of the everyday world and ourselves and groped for common reference points within this limbo of interchange, (p. 11)
No longer pretending to be faceless subject and invisible researcher, Tuhami and Crapanzano are portrayed as individual human beings with theit own personal histories and idiosyncrasies, and we, the readers, learn about two people and two cultutes.
on ♦ Eth ical Considerations
Because the objects of inquity in intetviewing are human beings, researchers must take extreme cat e to avoid any hatm to them. Traditionally, ethical
88
Trie Interview: "from Structured Questions to Negotiated Text
concerns have revolved around the topics of informed consent (receiving consent by the subject after having carefully and truthfully informed him or her about the research), right to privacy (protecting the identity of the subject), and protection from barm (physical, emotional, or any othei kind).
No sociologist or other social scientist would dismiss these three ethical concerns. Yet, there are other concerns that are less unanimously upheld. The controversy concerning overt/covert fieldwotk is more germane to participant observation, but could include the surrepririous use of tape-recording devices. Warwick (1973) and Douglas (1985) argue for the use of covett methods, because they mirror the deceitfulness of everyday-life realiry, whereas others, including Kai Erikson (1967), are vehemently opposed to the study of uninformed subjects.
Another problematic issue stems from the researcher's degree of involvement with the group undet study. Whyte (1943) was asked to vote more than once during the same local elections (i.e., to vote illegally) by membeis of the group he had gained access to, and befriended, gaining their ttust. He used "situational ethics," judging the legal infraction to be minor in comparison to the loss of his fieldwork if he refused to vote. Thompson (1985) was faced with a more serious possible legal breach. He was terrified of having to witness one of the alleged tapes fot which the Hell's Angels had become nototious, but, as he reports, none rook place during his research. The most famous, and widely discussed, case of questionable ethics in qualitative sociology took place during Laud Humphreys's research for Tearoom Trade (1970). Humphreys studied homosexual encounters in public resrrooms in parks ("tearooms") by acting as a lookout ("watchqueen"). Although this fact in itself may be seen as ethically incorrect, it is the following one that has raised many academic eyebrows, Humphreys, unable to interview the men in the "tearoom," tecoided their cars' license-plate numbets, which led him to find their tesidences with the help of police files. He then interviewed many of the men in theit homes without being recognized as having been their "watchqueen."
Another ethical problem is raised by the veracity of the teports made by researchers. For example, "Whyte's (1943) famous study of Italian stteet corner men in Boston has come undet severe scrutiny (Boelen, 1992), as some have alleged that Whyte portrays the men in demeaning ways that did not reflect their visions of themselves. Whyte's case is still unresolved, but it does illustrate the delicate issue of ethical decisions in the field and in reporting field notes, even mote than 50 years later (Richatdson, 1992b).
89
METHODS OF COLLECTING AND ANALYZING EMPIRICAL MATERIALS
A growing number of scholars, as we have seen (Oakley, 1981), feel that most of traditional in-depth interviewing is unethical, whether wittingly or unwittingly. The techniques and tactics of interviewing, they say, are really ways of manipulating the respondents while treating them as objects or numbers rathet than individual human beings. Should the quest for objectivity supersede the human side of those we study? Consider the following:
*         One day while doing research at the convalescent center, I was talking to
one of the aides while she was beginning to change the bedding of one of the patients who had urinated and soaked the bed. He was the old, blind, ex-wrestler confined in the emergency room. Suddenly, the wrestler decided he was not going to cooperate with the aide and began striking violently at the air about him, fortunately missing the aide. Since nobody else was around; T had no choice but to hold the patient pinned down to the bed while the aide proceeded to change the bedding. It was not pleasant: The patient was squirming and yelling horrible threats at the top of his voice; the acid smell of urine was nauseating; I was slowly loosing my grip on the much stronger patient, while all along feeling horribly like Chief Broroden when he suffocates the loboromized Mac Murphy in Ken Kesey's novel But there was no choice, one just couldnotsit back and take notes while the patient tore apart the aide. {Fontana, 1977, p. 187; emphasis added)
Clearly, as we move forward wirh sociology, we cannot, to paraphrase what Herbei t Blumer said so many years ago, let the methods dictate our images of human beings. As Punch (1986) suggests, as field-workers we need to exercise common sense and responsibility, and, we would like to add, to our subjects fitst, to the study next, and to ourselves last.
♦ New Trends in Interviewing
The latest trends in inrerviewing have come some distance from sttuctured questions; we have reached the point of inteiview as negotiated text. Ethnographers have realized for quite some time that researchers are not invisible, neutral entities; tather, they are part of the interactions they seek to study and influence those interactions. At last, inteiviewing is being brought in line with ethnography. There is a growing realization that intet-viewers are not the mythical, neutral tools envisioned by survey research. Interviewers are increasingly seen as active participants in interactions with
90
The Interview: From Structured Questions to Negotiated Text  ->
tespondents, and interviews are seen as negoriated accomplishments of both interviewers and respondents that ate shaped by the contexts and situations in which they take place. As Schwandt (1997) notes, "It has become increasingly common in qualitative studies to view the interview as a form of discourse between two or more speakers or as a linguistic event in which the meanings of questions and tesponses are contextually grounded and jointly constructed by interviewer and respondent" (p. 79). We are beginning to realize that we cannot lift the results of interviews out of the contexts in which they wete gathered and claim them as objective data with no sitings attached.
Interview as Negotiated Accomplishment
Let us briefly recap the two traditional approaches to the interview, following Holstein and Gubrium (1995, 1997). These authors use Jean Converse and Howard Schuman's Conversations at Random (1974) as an exemplar of the interview as used in survey research. In this context the interviewer is carefully instiucted to remain as passive as possible, so as to t educe his or her influence—the scope of the interviewer's function is to access respondents' answers. This is a rational type of inteiviewing; it assumes that theie is an objective knowledge out there and that if one is skilled enough one can access it, just as a skilled surgeon can remove a kidney from a donor and use it in a diffeient context (e.g., for a parient awaiting transplant).
Holstein and Gubrium (1995, 1997) regard Jack Douglas's (1985) creative interviewing as a lomanticist type of inteiviewing. Creative interviewing is based on feelings; it assumes that researchers, qua inteiviewers, need to "get to know" respondents beneath their rational facades, and that researchers can reach respondents' deep wells of emotion by engaging them, by sharing feelings and thoughts with them. Douglas's interviewer is certainly more active and far less neutral than Converse and Schuman's, but the assumptions are still the same: that it is the skills of interviewers that will provide access to knowledge and that theie is a core knowledge that leseaichers can access.
Holstein and Gubrium finally consider the new type of interviewing— well, "new" isn't exactly accurate, given that their reference for this is the woik of Ithiel de Sola Pool, published in 1957. To wit: "Every interview ... is an interpersonal drama with a developing plot" (Pool, 1957, p. 193; quoted in Holstein & Gubrium, 1995, p. 14). Holstein and
91