Qualitative Interviewing The Art Of Hearing Data Herbert J. Rubin Irene S. Rubin <§ %\ SAG E Publications International Educational and Professional Publisher Thousand Oaks London New Delhi 76 QUALITATIVE INTERVIEWING type of legislation, then what you have found may be generalizable. You would probably try several other areas of legislation before concluding that your results were in fact generalizable. To summarize, to extend results from qualitative interviews, choose interviewees who can provide grounded and accurate information, and talk to enough of them to get a complete picture of the research arena and feel confident that you understand it. Keep adding interviewees until what you are hearing from the interviewees begins to repeat. Then you can extend what you have learned beyond the original setting and the original interviewees by a logic of comparison. DESIGNING INTERVIEWS FOR DEPTH, DETAIL, VIVIDNESS, AND NUANCE One of the key differences between qualitative interviewing and survey interviewing is that the surveyors are trying to generalize relatively simple information, such as who are you going to vote for, whereas the qualitative interviewers are trying to learn about complex phenomena. Qualitative interviewers don't try to simplify, but instead try to capture some of the richness and complexity of their subject matter and explain it in a comprehensible way. But the richness from qualitative interviewing doesn't happen by itself; it needs to be designed into the pattern of questioning. One of the goals of interview design is to ensure that the results are deep, detailed, vivid, and nuanced. Depth^ms&ns getting a thoughtful answer based on considerable evidence as well as getting full consideration of a topic from diverse points of view. In early interviews you may have to encourage conversational partners to provide depth. In the introduction to the interview, you explain what you are interested in and why you are interested in it, so the interviewee will feel comfortable talking in depth on that topic. And you word questions to suggest the desired level of thoughtfulness or depth of the answer. "How do you like being a college senior?" conveys little information about the expected depth of answer, but, "Tell me how being a college senior differs from being a junior" invites the interviewee to think about the differences and explain with examples. People are more willing to talk in depth if they conclude that you are familiar with and sympathetic to their world. When possible, before the More Issues in Design tj interview begins, immerse yourself in the setting to learn what you can about the issues and start to understand the vocabulary that the interviewees use. When you use this vocabulary, you are telling the interviewees you are familiar with their world. To indicate sympathy, you might mention past experiences you have had in a similar agency, or an experience that you had that woke you up to the problems the group is facing. Or you can show your sympathy fay describing the goal of your research in terms of getting more attention for their group. Follow-up questions are a major way to get more depth. Suppose the initial question to a school administrator was, "What are some of the key problems you confront?" You may get a quick answer such as, "Oh, political problems are troublesome." To get more depth, you have to ask follow-up questions, such as, "What do you mean, 'political problems'?" or "Could you give me an example?" Then you could ask something like, "Are the problems getting worse or better?" and ask for a new set of examples. You can follow up again with a request for further elaboration by asking, "Why is it getting worse; what is changing?" or "How are you dealing with these pressures?" or "What are the impacts of these pressures?" You have series of questions, each asking for more depth. Another way to obtain more depth on a subject is to come back to it later and ask for clarification. You can schedule a second interview with the same person and say, "I read over the interview that we did last time, and there were some places where I did not fully understand what you told me. Could I ask you a few more questions about that?" Or you can pick an event that occurred between the first and second interview that illustrated what you had discussed in the first interview and ask for elaboration: In the first discussion, you may have talked about controversies on the school board. Then, after the first interview and before the second one, there may have been a brouhaha at a school board meeting when one board member demanded that the students be taught creationism instead of evolution in biology class. You can ask the interviewee's reaction to that event as a way of obtaining more depth on a matter you have already discussed. You can also go for depth by asking questions that unravel issues backward in time. When Irene was studying the effect of President Reagan's budget cuts on federal agencies, she understood why people were upset—they could lose their jobs—but did not understand what 7B QUALITATIVE INTERVIEWING seemed like paranoia about the way the cuts would be administered. Maybe some past event had taught professional staff to fear the politicians. So she began to ask interviewees about their prior experiences with cutbacks. She found that President Reagan's attacks on the bureaucracy reminded many of the agency staffers of the partisan, personal, and anti-Semitic attacks on the bureaucracy of President Nixon's administration a decade earlier. Asking about the past elicited the depth that made the present understandable (I. S. Rubin, 1985). You ask for detail by requesting particulars. Life is lived in details; the evidence for the generalizations you draw is in the specifics. You ask how hoboes find a place to sleep (Spradley, 1979), how hospital employees recognize and deal with patients' pain, or how elected officials maintain contact with their constituents (Fenno, 1978). You learn about each step of how a budget is balanced or how funds are put together from many sources to help finance housing for the poor. Having technical knowledge of the field makes it easier to ask questions that will elicit detailed answers. For example, Irene once began an interview with a high official in the Department of Housing and Urban Development by asking what had happened to the "O" regs. Her question indicated her awareness of the ongoing debate over this technical matter and suggested to the official that she was interested in detail. To solicit detail, you design interviews in ways that ask for specifics early on. If the interviewee mentions several cases, and you follow up on each one, you are signaling interest in detail. Each time you ask for an illustration or a step-by-step description of what happened, you are asking for detail. Once the interviewees understand that you want this detail, they will usually continue to provide specifics without being prodded. Another way to suggest that you want to hear about things in detail is to ask people how they carry out specific tasks for which they are responsible. The following excerpt from interview notes taken in Tampa, Florida, illustrates how well this can work: Irene: How do you put the management system, the SLA [service level analysis], and the budget together? Delilah: We start at the lowest part of the chain, the micro level, working with the management reports.... For example, hole patch- Uort Issues in Design 79 ing, you put the number and location, estimating by recent history, or how many lane miles you will pave. It's accomplishment oriented. It specifies the level of service we want to be at, for example, mow everything twice a year. Or will we only have a certain amount of money, and will only be able to afford 11/2 times a year? We do the ideal oné first, based on standards. At the average productivity rate, how many could we do if everything is normal? Translate that into crew days. We may not have that number of employees, so we have to revise it. Then, with a staffing estimate, we produce costs at standard rates. We find the total dollar value of labor, equipment, and materials. It's like the budget, but the budget is more detailed. But we can see how close we are to the budget target we have been given. Each activity has different levels of priority, a, b, and c. If we don't get 1.2 [million?] we will start with [cutting] the c's, and go up to the a's. We know how to play with it. If we reduce mowing from 2 times to 1 1/2, will that create idle crews? Where would they go? What will it cost? Can they go to another c activity? If so, that might be okay. This answer is wonderfully detailed and it stimulated the conversational partner to go into depth on related subjects in later answers. Although the detail seemed overwhelming at first, it opened up the theme of how bureaucrats respond to political demands, how they comply with demands to cut the budget, and how those demands translate into reduced services. Asking for a mess of details no one could love turned out to be the means of obtaining answers to a much larger question about how technical systems and political systems interact. By asking detailed questions on technical points that are of interest to the interviewee, you convey interest in learning about the details. For example, toward the beginning of Herb's interviews, when his interviewees talked about "tax credits," Herb asked what "internal interest rates" they received, a rather specific concern, but of great importance to the developers, who know the amount of funds they receive depends on this rate. By asking this question, Herb signaled his concern in hearing about details. Depth and detail differ but complement each other. To illustrate the difference between going after depth and going after detail, suppose you 80 QUALITATIVE INTERVIEWING are doing a life history with a middle-aged woman and she tells you that she recently invited her mother to move to her town. Going for depth, you ask questions that require a thoughtful response, such as, "Was encouraging, your mother's move a difficult decision for you?" To suggest your concern with details, you can ask, "What was happening in your mother's life when she decided to move?" "Is she more dependent on you now?" "Can you give me some examples of her increased dependence?" Clearly, depth and detail are related, and much of the time you want both. Detail adds solidity, clarity, evidence, and example; depth adds layers of meaning, different angles on the subject, and understanding. You get details by asking for examples, and then you explore the examples, step by step. In the interview with the woman whose mother came to town, she might say, "Yes, my mother depends on me much more than she did. For instance she expects me to take her to the dentist and to her doctor's appointments." You can then ask for details, "What happens when you take her to an appointment?" She answers, "You know, I pick her up and bring her to the clinic and sit with her while she waits for the doctor, then I take her home." You are getting detail, but still lacking depth. You continue looking for depth, asking, "Why don't you drop her off and come back for her?" She answers, "I can't go anywhere else and come back for her, she gets too nervous." You would begin to understand what this dependence meant and how it was felt. You have the depth of the interviewee's reasoning and the details to back it up. In addition to depth and detail, the interviewer should also design the questions to evoke vivid responses that will convey the range of feeling of the interviewee to those who read your report later on. To be vivid, an anecdote or example need not be dramatic or extreme, but it should create a strong vicarious experience for the audience. One way to encourage vivid answers is to ask for firsthand descriptions of events and not interrupt them when they occur, even if, for the moment, they seem off the topic. Consider this example that Jim Thomas shared with us from his prison research: [This guard would] mess with guys, and then he'd say, "Go ahead and beat me." And one day, somebody came up to him and hit him with a pipe, he's m More Issues in Design 8 j got a plate in his head now, [laughs] and he was lying on the floor, and blood was gushing, [laughs] and he was crying like a baby: "Please don't kill me, please don't kill me." He thought he was going to die right there. '/ The story is vivid. It could be used in conjunction with others similar . io it to provide a convincing description of prison violence. Talking about violence in the abstract just doesn't reach the reader the same way .that episodes like this do. Qualitative interviewing seeks a realistic description of a situation or cultural pattern. Realistic description includes shades of gray, not just -black and white. Nuance means precision in description, not blue, but cornflower blue, not just love, but love with energy and joy. It explores subtlety of meaning. What exactly does it mean to say you love someone? Is there a little fear mixed in with the love? Or a little dominance? How is the idea shaded or toned? You design questioning to elicit nuance and to reject unshaded answers. Some people's speech is naturally nuanced, but many people talk with bold distinctions, requiring you to ask for shades of meaning. First, you indicate you want nuance by presenting questions that show you expect something other than black-and-white or yes-or-no type responses. "Did you want to go to Thailand" asks for a yes-or-no answer. "How did you feel about the offer of a trip to Thailand" encourages a more nuanced answer. You elicit nuanced responses by asking follow-up questions, such as, "Does it always happen like that?" or implying that there must be another side or that something in the answer was too simple to be complete. The following excerpt from Irene's interview notes illustrates how to imply that an answer is too simple: Irene: I am curious about the effect of changing from at-large to partly at-large, partly district elections. Budget Director: [We] mix three at-large [councilmanic seats] and 4 district. If they were all district, we would have ward politics. We would have to locate 7 things in 7 wards. You want segments to be represented—it is more like a house and senate. Irene: There must be some pressure from the districts. How do you accommodate to that? 82 QUALITATIVE INTERVIEWING Budget Director: This question reminds me of my PA days, worrying about ethics, how will you distribute resources. Equity may not be enough, you may have to combine equity with other distributional criteria. You can't just do the downtown either, the neighborhoods need some too. . . . Part is equity, part is a vision of what the community needs. It is also influenced by other sources of money; partnerships direct money where it wouldn't be if there weren't matching funds. The answer to the follow-up question was much more nuanced in its implications that there are multiple competing interests that affect financial allocations and that the interests all get juggled so that everyone ends up getting something. You can also follow up on the nuances that conversational partners provide in their answers, to show that you are interested and want to hear more. In this interview, the budget director was explaining the advantages of a new budget system called target-based budgeting, and Irene asked for nuance by offering a simple summary of the interviewee's answer, a summary so simple the interviewee reacted against it and gave a more complete and subtle answer: Budget Director: It makes the departments and coordinators understand, and makes them prioritize and do the reasonable [thing]. It's not me or the mayor that says, "I don't want to hear that." They [the department heads] have to define what is important. Early, people are being creative, [asking] what should we be doing. As they go from thought process, [and] put pencil to paper, [and then] put detail behind it, some of them [the ideas] drop out. As the department looks at their [staffs] request and talks to the coordinators, things start to sift out, departments adjust the request list to those that are more appropriate. It's not top down, it's not "I only want to see 4 projects instead of the 10 you have." Irene: ... it forces the departments to winnow? .. . Budget Director: We force the departments to manage, they are new and innovative, there is incentive to do things. There is self-investigation, self-inspection, [they say] "can we do this better," they More issues in Design S3 talk of things. It's not done by the internal auditor, that creates hostility. The departments get involved and question how they do things. They do preliminary staffing [analysis], they meet with accountants and permit people to find out if things are possible, they don't just say, "mayor, let's go with it." Another way of getting nuances is to listen for signs of ambivalence or apparent contradictions and then inquire into them. In Herb's early interviews, community developers seem to vacillatexin how much they still believed in protest actions, at times saying they were terrific, and later on indicating that they were highly problematic. Herb designed follow-up questions to explore these mixed feelings and obtained thoughtful and nuanced responses on when protest is appropriate and when it causes more harm than good. PUTTING THE INTERVIEWS TOGETHER One major goal of design is to ensure that you get the depth, detail, and nuance that are the strength of qualitative research. Another purpose is to help you put together the separate pieces of information that you gather. An interview project is more than a single interview; you may interview the same person several times, you may interview many people in one site, and you may do interviewing in several sites. All this information has to come together at the end, and unless you build into the design some means of putting the pieces together, you may be very frustrated when it comes time to write your report. Combining examples, descriptions, and illustrations from different interviews to form a narrative is much easier if the individual interviews are designed originally in ways that help them fit together. We describe in later chapters how to draw an outline for yourself that serves as a guide across the interviews, ensuring that the pieces you collect fit together. The outline allows you to ask different people different questions and still have a place for each piece of information you get. You also have to weigh the versions of a variety of people. The interviewees know different things, the strength of their memories may