I conducted six interviews with members of the planetary science group and four interviews with members of the molecular biology laboratory. In the planetary science group, I interviewed all core group members that were willing to participate, that is, four senior scientists and a post-doctoral researcher. To learn more about the perspective of junior scientists, I additionally interviewed a PhD student. Except for the post-doctoral researcher, all interviewees were male. In the molecular biology laboratory, however, it was impossible for me to interview all or most group members. Therefore, I chose to interview at least one researcher from each level of hierarchy and to include at least one female perspective, eventually conducting four interviews with a male PhD student, another male, a former PhD student who had just obtained his degree, a female associate professor who in the group’s hierarchy formed part of the mid-level, and the group leader. All interviews lasted between 40 and 90 minutes and took place in interviewees’ offices or, if available, the group’s dedicated meeting room. Only the two PhD students from the molecular biology laboratory were interviewed in a university canteen.
To prepare for interviewing, I conducted two pilot interviews outside the two selected research groups and which provided a testing ground for my questions. When I later had familiarized myself with the selected research groups through extended observation, I began to interview group members. The shift from observing to interviewing was significant, since it implied a shift from a rather passive role in which I could keep my observations to myself to a role in which I structured the exchange between myself and the scientists more openly.
My approach can be described as “theory-generating expert interviewing” (Bogner & Menz, 2009, p. 47f.). Nevertheless, I do not perceive of my interviewing as the kind of expert interviewing that is geared towards particularly informed, exclusive or elite perspectives on scientific practice (as, for example, suggested by Zuckerman, 1972). Instead, I approached every interviewee as being “expert” in his or her individual daily professional practice, notwithstanding age or reputation. I employed a semi-structured question format for the interviews (Fontana & Frey, 2000, p. 653). Each time I used an interview guide with about 15 questions on it, which, over time, I slightly but not substantially adapted. Sometimes I asked the questions I had prepared literally; other times I reformulated them ad hoc so that they would not interrupt the conversational flow. Often, I would change their order by skipping questions that were brought up by interviewees themselves or that did not apply. When, for example, a PhD student told me that he had not yet written a journal article, I would not ask him about collaborative writing processes. My role in interviewing was moderately directive, allowing for detailed descriptions and digressions but guiding interviewees toward certain topics and questions. As I had observed most interviewees at their work beforehand, we could relate to specific persons, incidents or articles in our interview in order to illuminate more general points.
Interviewing is a method of data generation with an in-built verbal bias, privileging explicit reflection. On the one hand, the inexplicable and tacit can be difficult to access in interviews; on the other hand, the expectation to explicate themselves and the conversational dynamics between interviewer and interviewee may result in over-verbalizing and creating interview artifacts and ad hoc interpretations that tell more about the interview than about its subject. In my work, I have addressed these challenges by conducting multiple interviews, bringing each distinct individual perspective to the fore, and combining interview data with data gathered through observation.
From my experience, I conceive of interviewing as a dialog, a situated co-construction between interviewer and interviewee (King & Horrocks,
2010, p. 134), a discursive activity that forces individuals to face the “otherness" of their interlocutor in a process of dialectical negotiation. As Schostak points out:
The interview [...] is a place where views may clash, deceive, seduce, enchant. It is as much about seeing a world—mine, yours, ours, theirs—as about hearing accounts, opinions, arguments, reasons, declarations: words with views into different worlds. (Schostak, 2006, p. 1)
At the same time, for both interviewed scientist and interviewing philosopher, the interview is a compressed confrontation with their own work: for the scientist, because she is asked about her work; for the interviewer, because it is a test of the fruitfulness of her research question. Interviewing mobilizes intellectual resources of both interviewee and interviewer. Both have their own knowledge about the phenomenon in question, be it previously gained theoretical knowledge or practical experience, and both kinds of knowledge are necessary to establish an interview relationship between them. Hence, interviewing can be understood as mediation between different stocks of knowledge and experience. While the philosopher-interviewer draws on his or her conceptual discourses, the interviewed scientists will be asked to mobilize their professional experience in answering them. Interviews are thus a space where theory and experience, the abstract and concrete, meet through actual dialog.
The concept of co-construction enables one to make sense of the fundamental asymmetries involved in interviewing: the interview is jointly, but asymmetrically, co-constructed. Any understanding of what an interview actually is should take into account the fact that interview data originate from a particular context. Interviews are unusual situations. They are not part of everyday life, but interrupt daily routines and thus offer a niche for reflections that otherwise might not occur. Both interviewer and interviewee contribute to the interview, but they do so in very different ways. Neither during nor after the interview can interview partners be equal peers. Both of them are observers. During the interview, they observe one another and interpret facial expression, gesture, tone ofvoice, and implicit and explicit messages. But whereas the interviewer-analyst is primarily listening during the actual interview, the interviewee remains silent during the analysis process. The interpretation that is pivotal for a philosophical study is the one made by the interviewing philosopher, and in the end he or she comes to represent the co-constructed interview.
Interviewing has been a pleasant experience. Most of my questions were met with goodwill. The challenge for interviewing with philosophical intentions lies, however, before the actual occasion. Preparing good questions is a time-consuming task. Useful questions have to be both meaningful to the interviewee and meaningful with regard to the philosophical issues at hand. In my experience, asking practicing scientists to take up philosophical theorizing does not yield results of the quality desired. Interviewing is a co-construction, but its grounds have to be carefully prepared by the investigator.
-  Yet, in contrast to, e.g., Hasu and Miettinen (2006) and Ellis and Berger (2003), my dialogicalapproach carries no “interventionist" motivation. The concept of co-construction should not leadthe reader to believe that I have transcended the form of traditional interviewing—I certainlyhave not. I have restrained myself to asking questions, elaborating on these questions and offeringreformulations. In single instances I have explained in simple terms how some philosophers wouldthink about the issue in question. I have not, however, confronted interviewees with an elaboratedescription of my tentative, theoretically informed perspective.
-  The issue of representation has been dealt with comprehensively in anthropology, see, e.g., Denzin(1994, p. 503). Interviewing philosophers represent the people who are participating in their studyand they should deal with this burden carefully, see Fine, Weis, Weseen, and Wong (2000).