140 ♦ FORMS OF INTERVIEWING Spradley, J. 1979. The Ethnographic Interview. New York: Holt, Rinehart & Winston. Stewart, A.J. 1994. "The Women's Movement and Women's Lives: Linking Individual Development and Social Events." Pp. 230-50 in The Narrative Study of Lives, Vol. 2, Exploring Identity and Gender: The Narrative Study of Lives, edited by A. Lieblich and R. Josselson. Thousand Oaks, CA: Sage. Tillich, P 1957. Dynamics of Faith. New York: Harper & Row. Titon, J. 1980. "The Life Story, "journal of American Folklore 93:276-92. van Gennep, A. 1960. The Rites of Passage. Chicago: University of Chicago Press. White, M. and D. Epston. 1990. Narrative Means to Therapeutic Ends. New York: Norton. Widdershoven, G. A. M. 1993. "The Story of Life: Hermeneutic Perspectives on the Relationship between Narrative and Life History." Pp. 1-20 inThe Narrative Study of Lives, Vol. 1, edited by R. Josselson and A. Lieblich. Newbury Park, CA: Sage. Witherell, C. and N. Noddings, eds. 1991. Stories Lives Tell: Narrative and Dialogue in Education. New York: Teachers College Press. H FOCUS GROUP INTERVIEWING A dozen or so years ago, few social scientists bad even heard of focus groups. Now focus group interviewing is a widely accepted research method. The focus group interview can be defined as "a research technique that collects data through group interaction on a topic determined by the researcher" (Morgan 1996:130). This is a broad definition that includes most forms of group interviews, with the exception of observing naturally occurring conversations in ongoing interaction. Although some researchers do differentiate among varied kinds of group interviews (Frey and Fontana 1991; Kahn and Manderson 1992), the more common practice is to treat the focus group as a wide-ranging method in which the researcher has a variety of options for conducting the actual interviews. The fact that focus group interviewing is such a flexible data gathering technique is undoubtedly one of the reasons for its popularity. In this chapter I examine the rapid growth of the use of focus groups in the social sciences, both to explore why researchers conduct focus groups in the way they do and to stimulate thinking about different ways to conduct them. The chapter falis into three major sections. The first addresses the growing popularity of focus groups in the social sciences, especially in terms of how developments in marketing have become increasingly attractive to social researchers. Turning next to focus group methodology, in the second section I consider various approaches to moderating, in particular the issue of how structured moderating should be. In the third section I compare individual and group interviews in relation to questions of valid-it}' and the interchange ab i lity of these forms of interviewing as research procedures. The chapter concludes with a discussion of future directions for focus group interviewing. ♦ "41 142 ♦ FORMS OF INTERVIEWING ♦ The Popularity of Focus Groups The growth in social scientists' use of focus groups has been phenomenal. A search of both Sociological Abstracts and Psychological Abstracts reveals only a handful of publications about research that used focus groups during the 1980s, followed by a rapid upswing starting around 1990. By the end of the 1990s, the number of articles based on focus group research was well over 200 per year. Allowing for the lag between when a research project enters the field and when it gets published, the rising number of articles in the early 1990s points to an active experimentation with focus groups during the middle of the preceding decade. This corresponds to the first appearance of both methodological articles (e.g., Basch 1987; Morgan and Spanish 1984) and textbooks (e.g., Krueger 1988; Morgan 1988) that presented the method for a social science audience. It is, of course, easier to describe the growing popularity of focus groups than it is to explain why this occurred. Social scientists have only recently begun to pay careful attention to the history of research in their disciplines (e.g., Converse 1987; Piatt 1996; see also Piatt, Chapter 2, this volume), hence it is challenging to answer the question of why there was such a dramatic increase in the use of a particular method. One way of understanding the evolution of research methods is to consider their strengths and weaknesses. For example, researchers during the 1950s were especially fascinated by the possibility of representing a whole country's beliefs through just a thousand or so survey interviews (Converse 1987). Similarly, the 1970s saw an increasing interest in the possibility of bringing about social change through the evaluation of experimental social service programs (Campbell 1969). These strengths ostensibly explain the methods' respective popularity. If the strengths of a method are the primary force that guides its usage, then an examination of trends from roughly 1985 to 1990 in the social sciences, as well as in the larger society, should reveal an increasing interest in a set of research goals that called for the use of focus groups. By most accounts, the single most compelling purpose that focus groups served was to bridge social and cultural differences. Thus Richard A. Krueger and I urged our colleagues to "consider focus groups when you need a friendly research method that is respectful and not condescending to your target audience" (Morgan and Krueger 1993:18). This was matched by a broader interest in recognizing and understanding diversity, so it is not surprising that focus groups have become a prominent tool for quite literally giving voice to those outside the mainstream of society. (For summaries as well as critiques of this argument, see Johnson 1996; Cunningham-Burley, Kerr, and Pavis 1999.) It may be an overstatement, but there is undoubtedly a grain of truth to the contrast between using surveys to summarize the views of the entire nation during the 1950s and using focus groups to get closer to the thoughts and experiences of smaller and more specific segments of society in the 1990s. THE MIGRATION OF FOCUS GROUPS FROM MARKETING A different way of accounting for the shift toward focus groups is to trace the actual process through which they migrated from marketing and made their appearance in the social sciences. This approach replaces the seeming inevitability of focus groups with a more complex historical account of their rising popularity. Most important, it points to a set of circumstances that still are exerting an influence on how social science researchers use focus groups. There is indeed a consensus that the current use of focus groups in the social sciences arose through contacts with market- Focus Group Interviewing ♦ 143 ing, where it had been a popular technique since the 1950s (Johnson 1996; Krueger 1994; Morgan 1998a). This is not the whole story of the origins of focus groups, of course, given that some of the most important early work was done by the well-known social scientists Robert Mer-ton and Paul Lazarsfeld (see the introductory chapter in Merton, Fiske, and Kendall 1990). By the 1980s, however, that early work had been largely forgotten, and nearly all of the usage in the social sciences applied versions of focus groups that originated in the field of marketing. Throughout most of their history in marketing, focus groups have been an applied technique that fell outside the boundaries of academic market research. Because market researchers in academic settings had ignored qualitative methods such as focus groups, there was very little in the way of research-based guidelines for the use of focus groups in marketing practice. Similarly, because focus groups were not included in the formal curriculum, most marketers learned how to do focus groups through informal, on-the-job training. Indeed, a review of my 1988 book Focus Groups as Qualitative Research in a marketing journal (McQuarrie 1990) suggested that if social scientists were paying more attention to focus groups, marketing researchers might also want to take them more seriously. At first glance, this absence of academic attention to focus groups in marketing seems similar to the dominance of quantitative methods in the social sciences during this period. The key distinction, however, is that focus groups were widely used in applied marketing practice even if they were largely ignored in academic settings. This situation posed a dilemma for social scientists who were becoming interested in focus groups. Although there clearly was a substantial knowledge base about focus groups, this knowledge was not available through textbooks or the usual research literature. One especially striking feature of the empirical reports of marketing focus groups prior to 1990 is their vague descriptions of methodology. From a social science point of view, this would be unacceptable. For the marketers who wrote these articles, however, the details of their methods were the product that they were selling to their clients. Whereas social scientists might be rewarded for publishing methodological articles that provided guidance about when and how to do focus groups, commercial marketers who did so would be giving away their "stock-in-trade." So how did social scientists acquire the knowledge they needed to use focus groups? One important pathway was through the field of social marketing. Starting in the 1970s, a group of marketers sought to apply their techniques to "social problems," not just to the marketing of goods and services (Andreason 1995). This led to a partnership between the marketing firm of Potter-Novelli and a group of applied demographers who used focus groups to study fertility and contraception in Mexico (Folch-Lyon, de la Macorra, and Schearer 1981). This early linkage between social marketing and demography is a nice illustration of the importance of a particular context in the development of methods, because demographers, despite the quantitative dominance in that speciality, have continued to be an important influence on focus groups within academic research (e.g., Knodel 1993, 1995). More recently, social marketing techniques, including focus groups (Basch 1987), have been pursued by both academic and applied researchers in the field of public health, targeting a wide variety of health-related behaviors. Another point of contact was the connection between survey researchers and pollsters. Ironically, Lazarsfeld was one of the most important sources for this continuing exchange between the more academic and more applied sides of survey research (Converse 1987). Within marketing, focus 144 ♦ FORMS OF INTERVIEWING groups were often treated as the first stage in a research process that would be followed by surveys (Hayes and Tatham 1989). Indeed, many marketers asserted that the lack of statistical generalizability for focus group required confirmation by survey research. In practice, this meant that focus groups came to be treated as a valuable tool for creating survey questionnaires. As academic researchers became more aware of this practice, they began to experiment with the use of focus groups in the development of their own survey instruments. One illustration of this practice is the work of a group of researchers at the University of Michigan's Institute for Social Research who used focus groups to develop a survey on the sexual practices of gay men in one of the first social science studies on AIDS (Joseph et al. 1984). In that case, the inclusion of a qualitative method in the creation of a survey instrument was easily justified by the need to do exploratory work on a new topic in an understudied population. Since that time, focus groups have become a routine option in the development of survey instruments (Fowler 1995). For both social marketing and survey research, there were existing pathways between marketing and applied research within the social sciences. Social scientists also gained exposure to focus groups through a number of less formal mechanisms. In some cases, social scientists who did consulting work added focus groups to the services they offered (Richard Zeller, personal communication). In other cases, conducting marketing focus groups provided employment for future graduate students in the social sciences (Robin Jarrett, personal communication). In my own case, I was designing a project to use group discussions as a source of qualitative data when a friend with exposure to marketing asked me why I was going to so much effort to reinvent focus groups (see the acknowledgments in Morgan 1988). These were just a few of the points of connection that encouraged the crossover of focus groups from marketing to the social sciences. By the end of the 1980s, this process was occurring in any number of different places. For some fields, the nature of the connection was obvious, such as the use of focus groups in election campaigns and the subsequent interest among political scientists. In other fields, such as nursing, the source of the original connection remains less clear. What is clear is that the movement of focus groups into the social sciences was not something that happened just once. Similarly, it is probably not accurate to treat it as something that was championed by just a few advocates. Instead, the 1980s produced a growing interest in focus groups across a number of social science fields. THE ACCEPTANCE OF FOCUS GROUPS IN THE SOCIAL SCIENCES If the above discussion addresses some of the factors that affected how focus groups moved from marketing, what can be said about the context they moved into, namely, the social sciences during the 1980s? First of all, it is important to recognize that social scientists at that time considered marketing, especially applied marketing, to be well beneath them in terms of methodological rigor. This low-status origin may help to explain why few of the early social science articles on focus groups made more than passing mention of their debt to marketing (for an exception, see Morgan and Spanish 1984). Most descriptions of the social sciences in this period (e.g., Denzin and Lincoln 1994) would show a quantitative dominance but would also mention an increasing interest in qualitative methods. This growing interest in qualitative methods was certainly a favorable influence, which helps to explain why focus groups reemerged in the social sciences at this point, after a lapse of some 30 years. Another relevant factor Focus Group Interviewing ♦ 145 was the major reduction in funding for social science research in the United States that began during the Reagan administration. Without external funding, researchers needed to pursue smaller-scale projects in which they themselves did much of the work. Focus groups were well suited to this constraint. The emphasis on evaluation research during this period also played an important role. Both the development of programs through "formative evaluation" and the assessment of programs through "summative evaluation" could benefit from in-depth knowledge about specific situations and specific client populations that focus groups provided (Krueger 1988, 1994). It was relatively easy to adapt focus groups to these tasks, because the existing procedures for generating discussions of commercial products could also fit discussions of social services. Although there was a strong movement toward qualitative evaluation throughout this period (e.g., Patton 1990), focus groups frequently functioned as supplementary studies within projects based on quasi-experimental designs. This arrangement was beneficial in many ways, because it gave focus groups legitimacy as a research method at the same time it provided qualitative researchers with access to funding. Still, these "partnerships" were typically organized around the assumption that the quantitative aspects of the project were the most important. The same could be said about the uses of focus groups in survey research. They provided quantitative researchers with the opportunity to benefit from a supplementary qualitative method while largely remaining within the boundaries of their traditional approach. All of this emphasis on the use of focus groups within quantitative research projects raises obvious questions about the reception of focus group interviewing among researchers who had been trained in the use of other qualitative methods. This is an area that I personally have addressed ever since my earliest work on focus groups (Morgan and Spanish 1984), and it is explicitly embodied in the title of my 1988 book, Focus Groups as Qualitative Research. In my conclusions to that brief volume, I stated that I was "optimistic about the ability of focus groups to establish a unique position within the existing array of methods for gathering qualitative data," and that "if we are to expand our horizons to include focus groups as a routine option, it will most likely happen through their adoption by those who already have a solid background in qualitative research" (pp. 76-77). Although focus groups have indeed established their position as a qualitative method, the practitioners of other qualitative methods have had little to do with this success. In a few cases, qualitative researchers have encouraged the growth of the field, such as through sponsorship of a special issue on focus groups in the journal Qualitative Health Research (Carey 1995). In other cases, they have discounted focus groups, as in Michael Agar and James Mac-Donald's (1995) unfavorable comparison of focus group interviewing to ethnographic interviewing. By and large, however, established researchers with expertise in qualitative methodology have simply ignored focus groups. Because those with expertise in other qualitative methods have paid little attention to them, these researchers also have had little influence on the development of focus groups within the social sciences. The movement of focus groups into the social sciences thus presents a contrast between rapid acceptance in a variety of applied fields and a more tepid reception from established qualitative researchers. The most obvious change that occurred through this contact was the use of focus groups as a qualitative method within fields that had traditionally relied exclusively on quantitative methods. The focus group method itself, however, did not undergo a great deal of change in its migration from marketing to the social sciences. 146 ♦ FORMS OF INTERVIEWING ♦ Approaches to Moderating Most treatments of focus group methodology emphasize the need to keep the discussion on topic while encouraging the group to interact freely. It is the moderator's job to walk this tightrope. Arguably, there are many possible ways to balance the demands of both keeping a focus group discussion on topic and allowing the participants to express their own interests. Yet most research projects rely on only a narrow range of moderating strategies. To understand what moderators do, it is important to distinguish between the larger role that moderators frequently play and the specific activity of moderating the focus group discussion. Although it is common to think of the moderator's role solely in terms of what happens during group discussion, moderators almost always do more than that. In most focus group projects, the individuals who act as moderators also design and oversee the recruitment process that brings the participants to the groups. They write the questions that will guide the discussions. Following group discussion, it is usually the moderator's job to do the analysis and prepare the research reports. Elsewhere, I have argued that good recruitment, question writing, and analysis are just as essential to focus group research as good moderating, yet all of these activities are far less visible (Morgan 1995). Like an iceberg, the most obvious aspect of moderating is only part of the larger reality. THE VALUE OF MORE STRUCTURED APPROACHES IN MARKETING To understand why the marketing approach to focus groups often uses a more structured strategy, it helps to understand the broader role that the moderator serves in that field. Moderators typically perform a boundary-spanning role that connects what their clients want to know with what the participants say in the focus groups. One of the most interesting aspects of this effort to connect clients and participants occurs when the clients, behind a one-way mirror, watch a moderator lead a focus group (for fuller discussion of this arrangement, see Morgan 1998b). Even when marketing researchers do not have clients watching from behind mirrors, they routinely give videotapes of group proceedings to their clients, so that clients have a record of what the moderator did or did not do. Because moderating skills are one of the most costly things that clients are purchasing, being watched creates a strong incentive for moderators to prove their skills by taking a visible and active role in directing the group discussion. This need to perform before a client who is paying the bills may be the single biggest difference between what moderators do in marketing and what they do in the social sciences. 1 had the opportunity to observe a group of marketers and social scientists as they came to grips with this issue during a panel at the annual conference of the American Sociological Association in 1998. The session was organized by a group of marketers who had university affiliations, whereas the audience consisted mostly of academic researchers. When the presenters discussed the routine (for them) aspects of renting professional facilities with one-way mirrors and dealing with the clients who were observing the proceedings from the "back room," it created quite a stir in the audience. Eventually, someone stood up and, in disbelief, asked the panel something like, "You mean to say that the people who pay you to do the research actually watch the groups from behind a mirror? What human subjects committee ever approved that?" The rest of the session amounted to an exercise in virtual cross-cultutal communication as the marketers and the sociologists attempted to sort out their different assumptions about how to do focus groups. For current purposes, the difference that matters most is that marketers tend to use a more structured approach whereas social Focus Group Interviewing ♦ 147 Table 7.1 COMPARISON OF MORE AND LESS STRUCTURED APPROACHES ; TO FOCUS GROUPS ; More Structured Approaches Coal: Answer researchers' questions. Researchers' interests are dominant. Questions set the agenda for discussion. Larger number of more specific questions. Specific amounts of time per question. Moderator directs discussion. Moderator "refocuses" off-topic remarks. Participants address the moderator. scientists frequently use a less structured one. This difference has a long history, going back at least to Merton's initial exposure to Lazarsfeld's use of group interviews in radio research. After observing his first group, Merton voiced his opinion that the hired moderator was "inadvertently guiding the responses" and "not eliciting spontaneous expressions," so Lazarsfeld had Merton demonstrate his preferred style of interviewing by moderating the next group (Merton et al. 1990:xv-xvii; Rogers 1994). Table 7.1 summarizes the differences between more and less structured approaches to focus groups across the board, regardless of who is sponsoring the research. The left-hand column of the table lists the characteristics of more structured approaches, which center on the researchers' interests. The moderator's influence on the degree of structure begins well before the interview itself. The right-hand column lists the equivalent characteristics of a less structured approach, showing that participants' interests have a much greater impact on the course of the discussion. Taken together, the characteristics listed in the table indicate that the degree of structure has as much to do with the kinds of questions asked as with the way the moderator conducts the discussion of those questions. Less Structured Approaches Goal: Understand participants' thinking. Participants' interests are dominant. Questions guide discussion. Fewer, more general questions. Flexible allocation of time. Moderator facilitates interaction. Moderator can explore new directions. Participants talk to each other. red In fairness, the "need to perform" for go- the client is only one reason marketers tend po- to use a more structured approach. In addi- ;ws tion, marketing focus groups often involve irst members of the general public who have the only weak attachment to the discussion iid- topic. For many marketing topics, consid- on- erable effort must be expended not only to nad get the participants engaged initially but to : of keep them on topic during the subsequent lup discussion. It thus makes sense to organize 4). the discussion around a well-defined set of be- concrete issues, such as what participants hes like or dislike about a product or service, less how they would compare it to available al- rhe ternatives, and what they might do to im- tar- prove it. les, There is thus a good fit between market-sts. ers' more structured approach to focus - of groups and their need to work with partici-iew pants who have a low level of involvement the with the research topic. The larger lesson is uc- that this approach to moderating is driven tits' by a specific set of needs and goals. When the social scientists are operating in similar drier, cumstances, they too can benefit from the idi- strengths of a more structured approach to as moderating. When the circumstances are ons different, however, they need to step back on- and consider whether this approach still meets their needs. 148 ♦ FORMS OF INTERVIEWING QUESTIONING THE VALUE OF STRUCTURED APPROACHES IN THE SOCIAL SCIENCES Social scientists have a wider array of options for how they do focus groups because they pursue a broader range of goals across a number of different disciplines. This often leads them to work on topics that are quite different from those addressed in marketing research. One of the most important contrasts is that social scientists are more likely to work with participants who are closely connected to the research topic—such as my own work with recent widows (Morgan 1989) and with caregivers for family members with Alzheimer's disease (Morgan and Zhao 1993). When the participants in a focus group have a high level of personal commitment or emotional involvement with the topic, it is easier for them to start and maintain a discussion, so a less structured approach is a realistic option. Yet social scientists have continued to emphasize a relatively structured approach to focus groups. This is particularly problematic in situations where more structured focus groups run the risk of limiting the discussion to the topics the researchers want to hear about rather than revealing the participants' own perspectives. If social scientists' reliance on more structured approaches to focus groups were simply an unquestioned inheritance from earlier practices in marketing, this would indicate a serious lack of critical thinking. The reality, of course, is more complex. As I have noted in the earlier historical discussion, social marketing, survey research, and evaluation research were some of the most important crossover points for the migration of focus groups from marketing into the social sciences. Like marketing, each of these fields typically works with participants who have a low level of involvement with the topic. Because the needs of this type of social science research matched the more structured approach that was already prevalent in marketing, there was little need to modify the existing techniques. Further, as these fields supplied many of the first uses of focus groups in the social sciences, their reliance on a relatively structured approach to interviewing served as a model that influenced later applications of focus groups. This historical emphasis on structured approaches to interviewing is also present in nearly all of the books available about focus groups, which place a great deal of emphasis on the things the moderator should do to lead the group. There is considerable discussion of how to control difficult participants and how to get shy people talking, how to control overly talkative groups and breathe life into flat discussions, and the like. This kind of instruction makes it easy to conclude that focus groups will fail without the active direction of a highly skilled moderator. It is not surprising, then, that most novice moderators begin with assumptions that emphasize a more structured approach, so they are likely to perpetuate past practices unless they are exposed to alternatives. One way that I have tried to offer that alternative—in addition to my published descriptions of less structured focus groups (Morgan 1997:39-42, 1998b:43-.53)—is through training sessions that describe what I personally consider to be my "ideal focus group," which is based on a less structured approach. The ideal group would start with an opening question that was designed to capture the participants' interest, so that they themselves would explore nearly all of the issues that a moderator might have probed. Then, just as the allocated amount of time for that question was running out, one of the participants in the ideal group would spontaneously direct the others' attention to the topic for the second question by saying something like, "You know what really strikes me is how many of the things we're saying are connected to____" Anyone who has done much moderating has experienced this magic moment, as the group goes right where you want it to, without any help from you. When that hap- Focus Group Interviewing ♦ 149 pens, a less structured approach to moderating can keep the discussion going with little more than a smile and a nod. In my version of an ideal focus group, that kind of minimal response from the moderator would be all that was ever necessary, because the group itself would work through all the topics of the interview guide. Finally, 5 or 10 minutes before the session was supposed to be over, the discussion would begin to wind down, and the moderator could move toward closure with a typical wrap-up request, such as, "This has really been wonderful, and I'd like to finish by having each one of you summarize . . . ." In this ideal version of a less structured group, the moderator would have to ask only the first and the last questions. Beyond that, the group itself would cover every topic on the interview guide. Although this may sound like a fantasy, I have come close to it on several occasions. The trick is to remember that there is much more to moderating than just what the moderator does during the group. A less structured approach works best when the participants themselves are just as interested in the topic as the researcher is, so the first step is a recruitment process that carefully matches the participants to the research topic. Then the moderator has to write a guide in which the first question not only gets the discussion flowing but opens up a number of other topics that the participants will be eager to explore. So creating the possibility for a less structured focus group depends on a great deal more than the things that a moderator does or doesn't do during the discussion itself. Indeed, I like to say that the reason I have yet to moderate a group that fully matches my ideal has as much to do with my abilities at recruitment and question writing as with my moderating skills. My version of this ideal may be appropriate for many social science research projects (including the kind of work that I myself do), but it is far less likely to work when the participants have a low level of involvement with the research topic. And imagine what would happen if someone used this moderating strategy for a consulting contract where the clients were watching from behind a one-way mirror. Wouldn't those clients have to wonder why on earth they were paying the moderator so much money to do "nothing"? Of course, the real work would have been done before group discussion got under way, in the recruitment process and the writing of the interview guide, but none of that would show up during the group itself. Neither a more structured nor a less structured approach to focus groups is appropriate in every circumstance. Instead, researchers need to make well-informed decisions between these options. If my tracing of the historical development of focus groups in the social sciences is accurate, fields such as evaluation, social marketing, and survey research do indeed match some of the circumstances that lead marketers to use a more structured approach. In other areas, however, social scientists' reliance on more structured focus groups may be due to little more than their greater familiarity with that approach. Viewing social science approaches to moderating in historical perspective leads to the conclusion that the established practices deserve to be continued in some cases and questioned in others. If researchers are going to question their established procedures, they need both a sense of what their options are and a set of guidelines for deciding when one choice would be preferred over others. Fortunately, the existing methodological knowledge about more and less structured approaches to moderating now provides a starting point for such decisions. ♦ Comparing Individual and Group Interviews The existing expertise that social scientists have developed in individual interviewing might have served as a considerable re- 150 ♦ FORMS OF INTERVIEWING source for the development of focus groups as well. Instead, even the limited contact between these two seemingly similar methods has been relatively hostile. Much of this hostility has taken the form of questions about whether the data from focus groups are as "natural" or "valid" as the data from individual interviews. Further, the fear that focus groups not only produce poor data but can also be done more quickly than intensive interviewing has led to a belief, in some corners, that they pose a threat to "real" qualitative research. This section addresses both charges about the adequacy of the data from group interviews and concerns that focus groups might be used as a substitute for more in-depth approaches to qualitative research. Responding to these concerns is only one goal. More important, researchers need to learn from these disputes and move past them. An emphasis on the mutually relevant aspects of individual and group interviews will benefit both methods. ARE FOCUS GROUPS LESS NATURAL OR LESS VALID THAN INDIVIDUAL INTERVIEWS? In both one-on-one and group interviews, the interviewer and the research participant^) work together to create their conversation, but it is the interviewer who initiates the contact, determines the content of the conversation, asks the questions, and serves as the audience for the responses to those questions. From this perspective, Yvonna Lincoln and Egon Guba (1985) are right to speak of "naturalistic" inquiry, rather than treating any form of interviewing as truly natural. At the same time, however, Lincoln and Guba clearly claim that qualitative interviewing is more natural than survey interviewing; thus, even if all methods are at best naturalistic, some are apparently more natural than others. Unfortunately, the idea of a hierarchy of naturalness has also carried over to the idea that even if focus groups are qualitative interviews, they are more artificial than individual, open-ended interviews. I had a particularly memorable encounter with the idea that focus groups are less natural than individual interviews when I taught a workshop for several professors from the former Soviet Union, to help them study the transition to democracy in their home countries. My week of teaching about focus groups was preceded by a similar unit on autobiographical interviews taught by an anthropologist. On my second day of class, I was confronted with the opinion that focus groups are a contrived way of talking to people, at least in comparison to the techniques that had been presented the week before. In response, I asked the students how "natural" it is to have a complete stranger spend several hours talking about just the portions of his or her life that involved politics. In contrast, I asked whether it would be possible to bring together a group of neighbors to discuss how the politics in their country had changed since independence. They responded enthusiastically with remarks like, "You'd never be able to get them to go home!" By the end of the week, the students were engaged in a lively debate about the relative merits of individual and group interviews for the projects they were planning to do. This simple example illustrates how even a small amount of prior familiarity with a technique can make it into a de facto standard for assessing what one encounters later. Although I personally believe that seniority is the main reason some people think of group interviews as less natural than individual interviews, there are several seemingly substantive reasons behind this claim, and it is instructive to examine them. One possible source of the sense that group interviews are more artificial is the effort involved in bringing together a number of people for a focus group, which is often more overt than the work that it takes to conduct a series of individual interviews. Thus a focus group appears to be more Focus Group Interviewing ♦ 151 "staged" in comparison to each of the separate meetings in a set of individual interviews, even though it is the researcher who creates both kinds of conversations. The highly visible role of the moderator in focus groups is another source for the claim that they are less natural than individual interviews. This matches a belief that group dynamics are more complex than the dynamics in one-on-one interviews, so the skillful management of the interview is more important to the success of focus groups. The problem with this claim is that it is based on an appeal to common sense rather than actual evidence. In fact, it is just as easy to argue the opposite—that group interviewing is easier. As I have illustrated in the previous section, a group of participants who are interested in a topic can keep a discussion going with very little direction from the interviewer. Of course, that approach requires participants who have a relatively high level of involvement with the topic. But what about interviewing people who have little interest in a topic? Would you really want to spend an hour interviewing someone about bar soap or car seats? Yet marketers routinely get groups of people to share and compare their thoughts about such mundane topics. Another way to question commonsense assumptions about the naturalness of individual versus group interviews is to pose a counterfactual argument. What if social scientists had begun with group interviews and become interested in individual interviewing only at a later point—would that routine experience with focus groups lead them to think of focus groups as more natural and individual interviews as more problematic? Interestingly, this reversal of fortune is actually the situation in marketing research. There, focus groups are the better-established tool, whereas "one-on-ones" and "customer visits" (McQuarrie 1996) suffer by comparison to the well-established procedures and well-known value of focus groups. Thus the sense that one kind of interview is more natural than another may be a simple reflection of which of the two is more familiar. From a claim that focus groups are less natural than individual interviews it is only a short step to the more serious assertion that they are less valid. According to this argument, people are more likely to say what they "really" think in individual interviews because the presence of other participants during a focus group will influence what everyone says. It certainly is true that the same people might say different things in individual interviews than they would in a group discussion, but that does not mean that one set of statements is distorted and the other is not. Instead, if people say different things in different contexts, that is an interesting fact that may well be worthy of study in its own right. Carefully designed studies in which the same people are asked about the same topics in both individual and group interviews are rare, but Daniel Wight (1994) provides a particularly useful example in his work on adolescent boys' behavior toward girls their age. When an adult male interviewed these boys individually, they displayed relatively sensitive understanding of what girls their age expect in a relationship. Yet when this same interviewer led the boys in group interviews, they exhibited noticeably more macho attitudes, bragging about how they could get girls to do anything that they wanted (see also Eder and Fingerson, Chapter 9, this volume). Wight explains this difference by arguing that being around other boys brought a whole set of male-oriented norms into play for his respondents. Group interviews among adolescent boys are thus likely to invoke a particular aspect of their peer culture, but whether this is good or bad depends on what the research is about. For some purposes, observing the kind of macho behavior that boys bring out in each other might be crucial. For other purposes, this predictable group dynamic might get in the way. Rather than claiming that one set of results is more valid than the other, it 152 ♦ FORMS OF INTERVIEWING makes more sense to treat each method as more useful for some purposes and less useful for others. There is no denying the fact that the types of interviews that researchers conduct can shape the things they hear in interview conversations. Unfortunately, there is a gap in the basic knowledge that would predict when a researcher might hear one thing in group interviews and another thing from individuals. At this point, little more than generalities are available about the greater depth and detail to be gained from individual interviews and the possibility of observing social norms at work in group discussions. One way to improve the understanding of what each method can and cannot do would be to do more studies like Wight's that provide systematic comparisons of group and individual interviews. This would certainly be more productive than further debates about the relative naturalness and validity of the two methods, ARE FOCUS GROUPS A SHORTCUT THAT CAN REPLACE INDIVIDUAL INTERVIEWS? The most prominent claim that focus groups amount to a stripped-down, shortcut approach to qualitative research is undoubtedly Agar and MacDonald's {1995} critique, which appeared in the journal Human Organization. I agree that the typical practice of conducting three or four highly structured focus groups can never replace a series of in-depth, ethnographic interviews. What Agar and MacDonald ignore is the fact that focus groups that are done in this fashion typically serve very applied purposes that are quite different from the purposes of most ethnographic research. Even if most focus groups are designed to serve other purposes, this still leaves the question of whether focus groups are interchangeable with individual interviews in that they produce data that have the same degree of depth and detail as those produced by individual interviews. I would ar- gue that it is quite possible to use focus groups for such purposes. One example is the set of 53 focus groups that Jenny Kitzinger (1994) conducted to understand how people think about AIDS and why they think that way. Kitzinger worked with a variety of groups in which the participants were already acquainted with one another (including members of a retirement club, mothers from a preschool, prison officers, and male prostitutes). Using preexisting groups allowed her to hear how the participants used elements of their shared experiences to discuss a controversial topic. This kind of focus group research can indeed produce in-depth information about cultural understandings, but it is also quite different from small studies that serve narrow, applied purposes. These broad responses address the basic concern that originally motivated Agar and MacDonald's critique, but they do not speak to the substantive conclusions that those authors draw from their experience with group and individual interviews. It is instructive to consider three of the major points that Agar and MacDonald raise: ♦ Because the group interview format requires the moderator to direct the discussion, focus groups produce a limited form of interaction that is as much like a meeting as it is like a conversation. ♦ Individual interviews encourage informants to explain their "folk knowledge" to the interviewer, whereas the participants in a group conversation simply use their "indexed knowledge" without any need to explain themselves to one another. 4 Because of the limited kinds of interaction in focus groups, one must make a line-by-line analysis of detailed transcripts in order to understand the process that produced the data. Agar and MacDonald's (1995) first point concerns their need to fall back on a Focus Group Interviewing ♦ 153 directive style of moderating, despite their efforts to conduct a less structured group. With the benefit of hindsight, I submit that this problem was actually due to a flawed recruitment process, which Agar and MacDonald themselves characterize as a "comedy of errors" {p. 79). Although Agar and MacDonald sought seven or eight LSD users to discuss that drug, they ended up with just four respondents, only one of whom had extensive experience with LSD. As a result, they had trouble conducting an unstructured group among participants who had little experience with the topic they were supposed to discuss. By comparison, if the authors had intended to do an ethnographic interview, then locating only one experienced LSD user would have been sufficient. Even though Agar and MacDonald cite several focus group texts that emphasize the importance of recruitment, they seem to have followed a set of procedures that were more appropriate for individual interviewing. A pair of expert ethnographic interviewers managed to make one of the most common mistakes of novice focus group researchers. Even though Agar and MacDonald's failure to follow the standard advice on recruitment did have a predictable effect on group dynamics, their violation of the received wisdom on focus group recruitment also produced some interesting insights. Because the participants had a wide range of backgrounds with respect to the core topic, their discussion showed that teenagers who lack experience with LSD share the adult community's simplistic notions about the drug. If Agar and MacDonald had successfully followed the textbook advice and created a homogeneous group of LSD users, this finding would have been "designed away." This use of a homogeneous group to create more manageable group dynamics is another procedure that the social sciences have borrowed in a relatively unquestioned fashion from marketing. As a consequence, focus group researchers have done little to investigate the possible advantages of more mixed groups. Thus one benefit of Agar and MacDonald's inadvertent experiment with mixed groups is that it provides a reason to question traditional procedures in focus group research. Agar and MacDonald's second point, that the conversations in focus groups tend to rely on shared knowledge that does not need to be explained to an interviewer, also arises from their failure to follow well-established procedures in focus group research. The problem here is that their focus group participants were already acquainted through their involvement in the same drug treatment program. Agar and MacDonald complain against the "rule" requiring strangers in focus groups, but they fail to grasp that focus group participants who do not know each other will make their "folk knowledge" explicit as they explore the range of opinions and experiences the other participants bring to the group. Focus group researchers have well-developed procedures for dealing with "folk knowledge" and "indexed talk," such as encouraging participants to explore the degree of consensus and diversity in the group by first "sharing" and then "comparing" their opinions and experiences (Morgan 1997). What is lacking is a conceptual framework for discussing these issues, so Agar and MacDonald's ability to pinpoint and examine issues related to folk knowledge is quite useful. This seems to be a case where individual and group interviewing have different ways of reaching the same goal, but research on individual interviews has gone farther in developing concepts that could be beneficial in focus groups as well. The third point that Agar and MacDonald make concerns the need for detailed analysis of transcripts that capture some of the basic features of interaction in focus groups, such as pauses, overlapping speech, and "back-channel" responses (hmm, unh-hunh, yeah, and so on). I do not dispute the idea that this kind of analysis can offer certain advantages. 1 do, however, disagree with the assertion that only focus groups would benefit from the analysis of interaction. In fact, Agar and MacDonald 154 ♦ FORMS OF INTERVIEWING {1995) present only the bare text in their examples from individual interviews and justify this practice on the grounds that they "are no longer analyzing interaction" (p. 83). Even if group and individual interviews differ in their basic forms of interaction, there is still much to be learned from a comparative analysis of the kinds of interaction that they do share—such as starting a conversation, encouraging some responses and not others, and managing shifts in topic (see Schaeffer and Maynard, Chapter 28, this volume). Although Agar and MacDon-ald treat the need to analyze interaction as something that divides individual and group interviewing, I believe that it creates an opportunity for more dialogue between the two methods. Overall, considering Agar and MacDon-ald's claims both reinforces some of the well-established principles behind why social science researchers do focus groups the way they do and calls some of those procedures into question. On the one hand, the mistakes that Agar and MacDonald made provide a cautionary tale against importing the assumptions associated with individual interviews into focus groups. On the other hand, some of their "mistakes" produced interesting data, and their efforts to understand the differences between the methods demonstrate a set of conceptual frameworks and analytic techniques that will be new to most focus group researchers. Viewed from this perspective, this encounter between group and individual interviewing does indeed serve the larger purpose of making social scientists aware of the need for both change and continuity in the ways they use their methods. As I noted at the beginning of this chapter, the experience that social scientists already had with individual interviewing could have helped them in developing more innovative approaches to focus groups. Sadly, this possibility has been blocked by the need to address the kinds of issues that dominate this section. Further, the need to answer charges about the value of group interviewing has produced a de- gree of defensiveness in focus group researchers. The need to justify the reasons for doing focus groups not only diverts energy that could have gone into innovation, it reinforces a reliance on tried-and-true procedures for doing group interviews. Instead of arguing about the supposed superiority or inferiority of either method, social science researchers would do better to channel their energies into understanding both the differences and the similarities between these two forms of interviewing. Underlying several of the specific points discussed in this section has been the claim that many of the differences between focus groups and individual interviews in the social sciences arise from the different purposes that guide their use. It may be that these two methods will maintain their current division of labor, with more reliance on focus groups for relatively limited, applied projects and greater use of individual interviewing for studies that require more depth and detail. Even if these different roles do persist, researchers need to avoid the conclusion that this difference arises solely from the inherent strengths or weaknesses of either method. Instead, it is important to recognize that both methods can be adapted to serve a wide variety of purposes, and that the dominant direction that each has taken may be largely a result of historical circumstances. ♦ Finding New Directions Paying serious attention to the argument that methods are shaped by the social and historical contexts in which they are used creates the possibility of revising and reinventing research methods. But doing so requires a reflexive awareness of why researchers use their methods in the ways that they do. As I noted at the outset, the very newness of focus groups in the social sciences presents an unusual opportunity in this regard, because this field has not yet felt the full weight of tradition. Hence I will Focus Group Interviewing ♦ 155 devote my concluding comments to ways that future researchers can both build on existing knowledge about focus groups and create new ways of using focus groups. CREATING CONTINUITY Continuity does not just exist, it has to be created. The self-evident manner in which this occurs is through the unquestioned acceptance of existing assumptions. The most important way in which social scientists justify their procedures is through methodological research, so that they can offer evidence rather than tradition as the basis for their work. So it is not surprising that a group of social scientists who considered "future directions for focus groups" (Morgan 1993) came to the conclusion that the field needs a program of research on focus groups. The goal of that research would be to produce a better understanding of the difference it makes to do focus groups one way rather than another. This kind of methodological work is beginning to appear (see the review in Morgan 1996), but it remains scattered and idiosyncratic. One major factor that limits such work is the difficulty of assessing outcomes. It is easy to design focus groups that vary in size or moderating style or mixed versus homogeneous composition, but what is the best way to characterize the differences that might result from these variations in research design? One promising set of tools for addressing this issue is based on earlier work in discourse analysis (Potter and Wetherell 1987). I refer to this as an "expanded version of discourse analysis," because most of the work that has applied this approach to focus groups has been relatively eclectic. For example, Greg Myers (1998) describes his method of analysis as a blend of conversation analysis, linguistically based discourse analysis, and the study of argument. What all of these approaches have in common is the same kind of close analysis of transcribed conversations that Agar and MacDonald (1995) recommend. Consider three articles that illustrate this approach. Myers (1998) investigated both the ways that moderators handle the work of closing off versus extending exchanges among participants and the ways that participants can express disagreement without threatening group dynamics. Claudia Puchta and Jonathan Potter (1999) examined the perennial challenge of keeping a group on track while simultaneously encouraging open conversation, and their analysis shows how moderators managed this problem by elaborating on basic questions in ways that gave participants multiple options about how to respond. Finally, Myers and Phil Macnaughten (1999) researched the ways that participants created transitions between topics as well as the ways that the moderator's back-channel utterances influenced the group interaction. It should be evident how this kind of work could help us to assess the potential effects of different ways of doing focus groups. For example, earlier in this chapter I mentioned my own preference for a moderating style that encourages a process of "sharing and comparing" among focus group participants (Morgan 1997:20, 1998a: 12). In my own moderating, I have a series of things that I do to encourage the process of sharing and comparing, including the opening instructions that I give, the kinds of questions 1 use, and the way that I manage my own interaction with the participants. Through an expanded version of discourse analysis, it should be possible to find out which of these things really work. In fact, I believe that all interviewers have at least implicit explanations for why they conduct their interviews in the ways that they do, so discourse analysis is a way to find out whether the things that interviewers do truly make a difference. I want to be clear, however, that I am advocating these tools primarily as ways to assess what happens during focus groups. Some of the articles cited above make the case for this expanded version of discourse analysis as a general-purpose approach to analyzing focus groups and other forms of 156 ♦ FORMS OF INTERVIEWING interviews, but I believe that this remains an open question. Most of these authors readily admit that this sort of detailed analysis takes a great deal of time, and some of those who have tried it have found that the substantive returns simply did not justify the effort required (e.g., Gamson 1992). As with any method, researchers will have to decide on a case-by-case basis whether discourse analysis meets their needs. There is already, however, a strong case that these tools can help determine how specific research design decisions affect what actually happens in focus groups (see Schaeffer and Maynard, Chapter 28, this volume). Discourse analysis is certainly not a panacea when it comes to researching the things that happen in focus groups. It does, however, offer researchers a way to get started on a program of research about focus groups. Perhaps the results will show that marketing researchers were right all along—that the traditions social scientists borrowed from that field really are the procedures that work most consistently. More likely, social science researchers will discover that the established techniques represent good "typical case" solutions, but there are also a variety of circumstances in which other ways of doing things are more useful. ENCOURAGING INNOVATION It should be obvious that the same tools that make it possible to examine existing procedures in focus groups can be just as useful for the assessment of proposed innovations. But it is one thing to assess the value of innovative ways to do focus groups and quite another to find those new approaches. Fortunately, there is a great deal of naturally occurring variation in the use of focus groups, so what amounts to standard practice in one part of the field can seem quite innovative elsewhere. One promising example of this naturally occurring variation is the recent appearance of several articles describing the use of focus groups in participatory action research, most notably in a collection edited by Rosaline Barbour and Jenny Kitzinger (1999). There are many ways in which participatory action research differs from the typical project where the research team collects data from the participants and then departs; for present purposes, however, the most important difference lies in the relationships that unite the researchers and the participants. This basic modification of the research setting has implications for nearly every aspect of focus group research. For example, what does "recruitment" mean in this context, and does it even make sense to use "questions" as the basis for discussions that encourage empowerment? Certainly the moderating style in participatory action focus groups must be different from leading either structured or unstructured discussions for other purposes. It would be nice to have answers to these questions, but at this point nearly all of the published articles on participatory action focus groups are devoted to the general merits of this approach. This lack of methodological detail is understandable in a form of research that is driven by its action orientation. Yet once the value of focus groups for participatory action research is well established, that should motivate the appearance of "how-to" articles specific to that field. When that happens, focus group researchers outside the field will also be able to benefit from these innovations. It may be the case that some of the procedures in participatory action research focus groups are context-bound and thus unlikely to find uses outside that field, but, with any luck, some of these procedures will be useful across a variety of contexts. Participatory action research is, of course, just one example of how the field as a whole can benefit by borrowing procedures that were developed elsewhere. Cross-cultural variations in focus groups are another source of procedures that are well established in one setting but innovative in another. A British student in one of my focus group workshops taught me a Focus Group Interviewing ♦ 157 valuable lesson in this regard. After I drew a diagram of a "typical" focus group setup, with the moderator and participants seated around a table, she wanted to know, "What's all this business about a table? I've done lots of focus groups, and we never use a table." True enough, it seems that focus groups in Britain emerged from a media-oriented form of marketing that relied on "living room discussions," where a group of neighbors would watch a television program together and then discuss it afterward (Lunt and Livingstone 1996). Consequently, British focus groups are more likely to be conducted in homes or homelike settings (for a comparable example from the United States, see Gamson 1992). Cross-cultural variations on focus groups are not limited to the use of tables or viewing rooms with one-way mirrors. Over the years, I have heard any number of anecdotes about how focus group procedures have been adapted to both Western and non-Western cultures. What is lacking, however, is any systematic investigation of these naturally occurring innovations. Beyond simple descriptions of the variations, there needs to be an explanation of why things were done one way rather than another. Once other researchers understand the purpose that a particular proce- dure is supposed to serve, this can stimulate their methodological imaginations to think of other ways to use that innovation. I want to close with a particularly radical suggestion. Social scientists need to go back to marketing and find out more about why they do the things they do. Although social scientists borrowed many of their current approaches to focus groups from marketing, there has been remarkably little continuing contact between the two fields. This creates a substantial opportunity for innovation. In making renewed contacts with marketing approaches to focus groups, social scientists should expect to have their own assumptions challenged, but that kind of learning is one of the great advantages of "cross-cultural contact." As this reference to marketing should make clear, these recommendations for developments in group interviewing return to the original spirit of innovation that accompanied the migration of focus groups into the social sciences. Just as social scientists in the 1980s responded to the basic appeal of focus groups, current practitioners should be equally enthusiastic about locating new ways to do group interviews. The goal should be not only to use this method, but to develop it as well. References Agar, M. H. and J. MacDonald. 1995. "Focus Groups and Ethnography." Human Organization 54:78-86. Andreason, A. 1995. Marketing Social Change: Changing Behaviors to Promote Health, Social Development, and the Environment. San Francisco: Jossey-Bass. Barbour, R. S. and J. Kitzinger, eds. 1999. Developing Focus Group Research: Politics, Theory and Practice. London: Sage. Basch, C. E. 1987. "Focus Group Interview: An Underutilized Research Technique for Improving Theory and Practice in Health Education," Health Education Quarterly 14:411-48. Campbell, D. T. 1969. "Reforms as Experiments." American Psychologist 24:409-29 Carey, M. A. 1995. "Issues and Applications of Focus Groups: Introduction." Qualitative Health Research 5:413. Converse, J. M. 1987. Survey Research in the United States: Roots and Emergence, 1890-1960. Berkeley: University of California Press. Cunningham-Burley, S., A. Kerr, and S. Pavis. 1999. "Theorizing Subjects and Subject Matter in Focus Group Research," Pp. 186-99 m Developing Focus Group Research: Politics, Theory and Practice, edited by R. S. Barbour and J. Kitzinger. London: Sage.