Blurring the Line Between Researcher and Researched in Interview Studies

Abstract

There is a rich history of feminist critique of research methods in the social sciences, with feminist scholars noting an androcentric bias and a focus on the values and concerns of the dominant group (Harding, 1987; Reinharz, 1992). The experiences and concerns of women and girls have historically been ignored in traditional social science research or studied only in relation to men (Harding, 1987). As a result, feminists have expressed the need for methods that respect and empower women and that accept women’s experiences as legitimate sources of knowledge (Campbell & Wasco, 2000). Qualitative research has been discussed as a feminist research method because women’s voices can be heard on their own terms and the traditional hierarchy between researcher and participant can be flattened (Campbell & Wasco, 2000; Oakley, 1988; Sprague & Kobrynowicz, 2003).

This commentary is intended to serve as a companion to our original research article (Chmielewski & Yost, 2013) and to contribute to the conversation on feminist research methods. In it, we provide additional details about our research project, as well as theorize those details in terms of feminist values within social science research. We utilized semistructured interviewing, a method that can embody feminist ideals (Garko, 1999; Sprague, 2005), although interviewing has received less feminist theoretical attention than some other methods (e.g., ethnography; Stacey, 1991). The impetus for this commentary came from one potentially critical detail of our research process: second author Megan served as the first participant in the study, interviewed by first author Jennifer; the analysis and writing that followed was primarily conducted by Jennifer under Megan’s supervision.¹

In this commentary, we detail the evolution of our thinking on the dual roles played by Megan as researcher and researched; we explain the dialogue that took place between us and editors of Psychology of Women Quarterly (PWQ); and we conclude by considering the ways in which our practice, which blurred the line between researcher and participant, may have been consistent with some feminist goals, but perhaps not with others. We, along with PWQ’s editors, hope that this essay will spark further conversation about interview methods among feminist social scientists.

Our Research Process

We began our research project by firmly grounding it in feminist values and a constructionist, phenomenological epistemology. We were particularly compelled by the delineation of several guiding themes of feminist psychological research provided by Worell and Etaugh (1994). By undertaking a qualitative project from a constructionist paradigm with participants beyond college students, we implicitly challenged the tenets of traditional, “objective,” scientific inquiry (Worell & Etaugh’s Theme 1). By studying women’s perceptions of the influences on their body image—a topic of relevance to women’s lives—without comparison to a sample of men, we focused on the experiences and lives of women as a group worthy of study in their own right (Worell & Etaugh’s Theme 2). By opening the conversation to sources of positive body image, rather than focusing only on factors relating to poor body image, we promoted women’s empowerment (Worell & Etaugh’s Theme 3) and envisioned social change (Worell & Etaugh’s Theme 6). Additionally, by inviting women to speak as experts on their own experience, we hoped to embody empowerment in the process of data collection. Given critiques of the researcher–researched relationship as potentially exploitative (Campbell & Wasco, 2000; O’Shaugnessy & Krogman, 2012; Sprague, 2005), we used interviews and drew heavily on quotes from our participants as a means of recognizing their expertise (Ralph, 1988).

With these feminist research ideals in mind, Jennifer proposed the project as part of her undergraduate honors thesis, which was supervised by Megan. Jennifer came to the topic of her honors project—influences on lesbian and bisexual women’s body image—as a result of personal experience (body image issues as a bisexual adolescent, growing up in a town with few “out” sexual minority women) and a deep desire to understand that experience. As Jennifer refined her research questions, she decided that she would focus exclusively on the experiences of sexual minority women, about which there are many open questions. Thus, we determined that a qualitative, grounded method (with minimal a priori hypotheses) best suited her research questions.

By the end of the first semester, Jennifer had developed an interview protocol. As part of the development process, Jennifer informally discussed her interview questions with several peers in the LGBT (lesbian, gay, bisexual, and transgender) community on campus (many of whom were psychology or sociology majors). She also conducted a “pilot” interview in which she ran through the protocol with a friend who role-played responses as if she were a bisexual woman. Jennifer found this process helpful in revising the interview protocol.

As the first semester came to a close, Jennifer had advertised her research project to faculty and staff on our campus and to the mailing list of a regional LGBT resource center. Volunteers were coming forward, and we decided that one additional run-through would be beneficial. Role-play, of course, is not necessarily a good proxy for a true interview, and so we discussed the possibility of a true-to-life interview experience. We decided that Megan would serve as an interviewee. This decision provided several benefits. First, Jennifer would have the opportunity to interview an adult rather than a college student (paralleling her intended sample of nonstudent adults). Second, she would interview a person who truly met the inclusion criteria rather than someone pretending to do so. Finally, we thought that it would be less intimidating for Jennifer to have her first interview with someone she knew well, rather than with a stranger. Although cognizant of disparities in the usual student–teacher relationship, we had established a great deal of familiarity with each other: Jennifer had been a student in four of Megan’s classes, had served as a teaching assistant in one of Megan’s classes, had volunteered in Megan’s lab for several semesters, and was also an executive board member of a feminist student organization that Megan advised.

Prior to the interview, we discussed the possibility of using the data, and we thought we might do so if we were unable to find enough participants, and also because we recognized that Megan met all the inclusion criteria for the study. During the interview, which took place in the research suite (as did most other interviews), Jennifer proceeded through the protocol in the same fashion as she would with all later participants, and Megan fully and authentically participated, responding openly and honestly to the questions.

Jennifer moved on to the next interviewee soon after the interview with Megan took place. Jennifer transcribed these two interviews and began the analysis (Interpretative Phenomenological Analysis [IPA]; Smith & Eatough, 2007) over the winter break. In the second semester of the project, Jennifer continued interviewing and analyzing data until we felt we had achieved a sufficient sample of participants and no new themes were identified. At this stage, we decided to retain Megan’s interview as part of the data set because of both reasons noted above: indeed, it had proved difficult to find bisexual participants and Megan’s interview fit well with the rest of the sample. Because this project was her honors thesis, Jennifer took primary responsibility for the IPA analysis, doing the line-by-line coding and developing the initial thematic structure for each interview on her own. Megan served as a second reader, discussed the themes, and checked to be sure that all themes were represented in the data. Given that Jennifer was taking the lead in data collection, analysis, and writing and that Megan’s transcript was not handled any differently from the other participants, we did not include the detail about Megan’s participation in the original article.

The Dual-Role Question

Jennifer successfully defended her thesis near the end of the academic year, and we moved forward with revising the thesis for publication. It was nearing the end of the publication process (at the final copyediting stage) when both of us were struck by the concern that our research method had not been fully disclosed. Specifically, when reading Sage’s standard disclaimer regarding “conflicting interests” that accompanies page proofs, we wondered whether the dual role played by Megan in this project should be noted as a potential conflict of interest, or should be described in the body of the article. Megan contacted Jan Yoder, PWQ’s editor, to explain the situation and to seek her advice in moving forward.

Jan first recruited two of PWQ’s associate editors, Heather Bullock and Maria Gurevich, who have qualitative expertise, to discuss the issue. The conversations among Jennifer, Megan, and this editorial team pushed our thinking from the practical issue (i.e., is it a conflict of interest problem for a researcher to be part of the data set?) to a theoretical discussion about feminist research methods. In the remainder of this essay, we describe those conversations, and then we end with reflections on feminist methodology that arose over the course of these discussions.

Editorial Conversations: Pilot Data?

The editorial team initially asked whether Megan’s interview ought to be considered pilot data and thus removed from the article. Megan and Jennifer considered this idea, and we tentatively concluded that Megan’s interview was not pilot data. First, pilot work is generally exploratory, attempting to inform the upcoming full study. For example, a qualitative researcher might conduct a pilot to gain a more in-depth focus of the theory to inform the study (Frankland & Bloor, 1999; Maxwell, 1998). Additionally, pilot research is often undertaken with the explicit intention of making changes to the protocol before moving forward with the full study. The interview that Jennifer conducted with Megan did not fit these definitions of “piloting.” In our study, this interview did not inform the upcoming study any more than each individual participant informed the subsequent interviews. That is, after each and every interview, Jennifer had more information and deeper insight derived from the previous participant and made minor changes to the protocol. From a grounded, constructionist perspective, this flexibility and responsiveness by the researcher is simply an expected part of the process (Grafanaki, 1996).

While discussing what constitutes pilot work, the editorial team suggested that one empirical way to address this question would be to remove Megan’s transcript from the data set to see what would, and would not, change in the analysis and the article. We did so, to the extent that we could, given that it was part of the original development of our data set, and we determined that the analysis would largely remain the same. For all subthemes to which Megan contributed, there was at least one other participant who also stated that theme. Thus, removing Megan’s data would not affect the overarching thematic structure of the article, and all superordinate themes would have remained intact. Because of the manner in which IPA is conducted (researchers using IPA initially analyze each individual transcript on its own, and only in later stages draw upon the results of one transcript’s analysis to analyze other transcripts), we felt confident that the thematic structure that was developed with Megan’s data would also have been developed without her data.

However, Megan’s comments in theme 4.2, which refers to body surveillance when in relationships with women, made this determination more complicated (Chmielewski & Yost, 2013, p. 235). This subtheme includes information from two participants, one of whom is Megan, but the details of their responses are quite different from one another. One participant discussed the detrimental effects that her partner’s negative body image had on her own body image. In contrast, Megan discussed issues relating to butch–femme differences and highlighted the negative impact that having a partner with a more masculine gender presentation (and a slimmer, less curvaceous body shape) had on her own body image. Both participants differed from the rest of the sample, who all generally felt better about their bodies when with women rather than with men, which is why they both encompass theme 4.2. If Megan’s data were removed from the article, we would still have been able to report that one woman felt better about her body with men than with women, but we would have lost the ability to theorize about bisexual women’s body image in terms of butch–femme gender presentations.

Continued Conversations: Researcher and Researched?

Upon agreeing that the interview in question ought not to be considered pilot data, the editorial team considered why Sage’s conflict of interest disclosure might have raised a red flag for us as authors. They questioned whether the heart of the matter lay in the potential conflict between the roles of researcher and participant. In addition, they wondered whether interview designs explicitly separate researcher from participant in ways that other methods (e.g., ethnography) do not. They asked: If interview methods draw a line between researcher and researched, then is it ever appropriate to include a researcher’s interview in the analysis?

We found this question very intriguing. First, we considered the nature of the potential conflict between participant and researcher. It seems that statements of conflicting interests generally refer to financial conflicts, such as in drug trials: If the researcher stands to financially gain from a certain outcome, then the objectivity of the design and analysis could be compromised. That kind of financial issue was not relevant here, so instead we tried to think through the “interests” that a participant and a researcher each have in an interview study.

A participant’s interests are potentially multiple: to contribute to science and/or society, personal gain (e.g., research credit in a course, payment, and institutional change) and to have one’s voice heard. A researcher’s interests are also potentially multiple: to contribute to a body of literature in one’s field, academic gain (such as achieving publication), and other intangible gains (e.g., recognition from colleagues and feeling a sense of accomplishment).

We tentatively concluded that these two sets of interests do not necessarily conflict with one another. The only interests that seem potentially in conflict are between achieving publication and the desire to tell one’s story accurately. However, it seemed to us that this would be of concern in a study that had a priori hypotheses that could be confirmed by the data, but not in a study using a constructionist paradigm. Qualitative studies, particularly those grounded in data, do not have predetermined hypotheses that are being “tested” by the data (Corbin & Strauss, 2008) and that was the case in our study. What did remain as a concern, though, was the extent to which Megan’s story differed due to her position as a coinvestigator rather than a naïve participant; we see this as a different issue than a “conflict of interest,” though, so we explore this later in the present article.

The process of data analysis and writing would be another instance where potential conflicts of interest could arise. Upon reviewing our procedures, we believe that our analysis process protected against potential conflicts becoming actual conflicts. Because this project was Jennifer’s honors thesis, she conducted the analysis, following the IPA procedures laid out by Smith and Eatough (2007). Megan, as the supervising faculty member, served as a second reader, discussed Jennifer’s thematic structure with her, and checked to be sure that all themes actually appeared in the data. The prescriptive nature of IPA and the fact that Jennifer was responsible for the analysis lead us to conclude that no potential conflicts between the roles of the researcher and the participant were actualized.

Although Megan did not have conflicting roles according to our determinations, she certainly did have multiple roles. Thus, a final issue we considered was that of researcher bias. Did Megan’s involvement in the study from its conceptualization, followed by her participation as a research participant and then her position as a supervisor of the analysis and writing, unduly influence the course of the project?

As feminist researchers, we believe that all research is selective and “biased,” conducted to answer particular questions that the researcher deems important, using specific types of evidence drawn from methods that the researcher chooses (Mays & Pope, 1995). It is always the researcher’s job to create and shape studies, and Megan did this through her role as supervisor of the project. But we do not think that Megan’s participation in our study gave her influence over participants in any direct way: Jennifer was the sole interviewer and Megan had no contact with participants.

Where Megan did have indirect influence, however, was as the first interviewee. Jennifer made slight modifications to the interview protocol with each and every participant. Because this was a semistructured interview study, each interview informed subsequent interviews (Dilley, 2000; Grafanaki, 1996; Holloway, 1997; Maxwell, 1998; Morrow, 2005) as Jennifer made changes based on what she had learned with previous participants. Also, each interview varied in terms of wording, order, and follow-up questions, with Jennifer allowing participants to bring up topics of importance to them at different times. As the first interviewee, Megan did indirectly shape the rest of the interviews in this way. After our interview, Jennifer smoothed out transitions, eliminated three questions that she decided were superfluous, and increased self-disclosure and commenting to build rapport with participants with whom she did not have an established relationship.

Jennifer continued making changes based on interview data following the second and third interviews as well; these two interviews informed the interview protocol just as Megan’s interview had. For example, after the second interview, Jennifer eliminated introductory question asking participants to describe themselves in general because she found that it was irrelevant and perhaps an overwhelming question. Because both Megan and the second interviewee discussed feelings of not being accepted and being sexualized, Jennifer added a question about these types of discrimination for the third and subsequent interviews.

Jennifer’s continued modification of the interview questions meant that each participant informed subsequent interviews (Dilley, 2000; Holloway, 1997). As the first participant, Megan did have greater (indirect) influence in shaping the interview protocol; however, due to the nature of semistructured interviewing (Grafanaki, 1996; Morrow, 2005) and our commitment to a constructionist paradigm, we are not sure that this was undue influence.

A final issue of “researcher bias” remained. During our exchanges, the editorial team pointed out the possibility that what Megan reported in her interview may have been somehow “more” than what she truly experienced, because she had actively thought about her experiences in the context of this research study (something that other participants could not do). This may have occurred, but we are hard-pressed as to how we could empirically demonstrate whether it did or did not actually occur. Instead, we would again draw on a constructionist approach in order to address this issue.

From a realist perspective, the researcher takes the participant at his or her word, assuming that the words that are spoken reflect reality. From a constructionist perspective, however, the researcher takes the words of the participant as indicative of a reality—a reality that is true for the interviewee at the time of the interview with that researcher, but that might not be her reality when talking with a different interviewer, or when talking with her friends, or at some other time in life (Burr, 2003). A constructionist perspective would view interview data as a text that is coconstructed by the researcher and the interviewee; it is always partial information, it is always contextual, and it is never ultimate truth. An interview involves a participant making meaning from her own experiences, interpreting her experience in light of prior knowledge and the current context.

If Megan had actively thought about her experiences in the context of our research study, and if that thinking influenced her to discuss her experiences in a particular way in the interview, she still would have been reporting her reality at the time of the interview—just as all participants did. The foreknowledge did make Megan different from the other participants on the whole; Megan’s knowledge of most of the interview questions meant that when she was asked a question, it was something she had heard before (or even had a hand in writing). This may have allowed her to think about her experiences in a different way than she would otherwise have done had she entered into the interview “cold.” But, we think it is important to note that every participant differed from the others in many meaningful ways (in terms of experience, knowledge, background, and other social identities), and each was constructing her interpretation of her experiences at the time of the interview. We remain unsure as to how important it is that Megan differed from the other participants in this way, given the constructionist paradigm undergirding this study—a perspective that does not view an interview as revealing objectively true and verifiable information.

Blurring the Line Between Researcher and Researched

These rich conversations we were having with Jan, Heather, and Maria made us even more certain that Megan’s dual roles should be acknowledged in order to fully inform the reader of all procedures that were followed. We did not decide that Megan would serve as the initial participant until Jennifer was about to begin data collection; thus, the original design of the study did not include the anticipation of Megan’s participation, and this was an important detail that we and the editorial team concluded must be shared with readers. However, rather than rewrite the original research article, the editorial team proposed the present series of commentaries to explore the issue of the boundaries between researchers and participants, and whether there is value for feminist scholars in blurring that boundary.

Although this conversation began with our own concerns as authors over potential “conflicting interests,” in the end, it opened the door for a more theoretical consideration of feminist research methods. Interview methods traditionally do draw a line between the researcher and the researched (Sprague, 2005), and the interviewer is traditionally considered an objective collector of data. The process of our conversations, though, caused us to question whether feminist scholars might be inclined to purposefully blur that line and instead incorporate the subjectivity of the interviewer.

Feminist scholars already discuss blurring the line between researcher and researched in other grounded, constructionist methods, such as ethnography and participatory action research (Cammarota & Fine, 2008; England, 1994; Gilbert, 1994; McIntyre, 2007; Zavos & Biglia, 2009). However, semistructured interviews differ from ethnography and participatory research in several key ways, most notably in the level of engagement of the researcher with the community under study. Ethnographic and participatory studies typically involve a researcher becoming a part of the community, speaking with multiple informants within that community, and the analysis draws on a range of data (e.g., observation notes, individual and sometimes group interviews, and personal reflections); interview studies typically do not involve this level of engagement with a community, and the analysis is limited to the speech in the interviews themselves. Despite the differences, in both instances, the research involves participants speaking to a researcher about the issue under study.

With that conversational and relational component in mind, we could think of several reasons why feminist scholars might choose to blur the line between researcher and researched in interview studies. All of these reasons center on the broad feminist goal of breaking down hierarchies between people (Campbell & Wasco, 2000; Oakley, 1988; Sprague & Kobrynowicz, 2003; Worell & Etaugh, 1994); in contrast to this feminist goal, research studies have traditionally proceeded with the researcher holding and exercising power over participants (Wolf, 1996). Although the reduction of power and hierarchies in the research setting was certainly not the driving force behind our decision to include Megan’s data in our analysis, upon reflecting on our research study in the context of the discussion we were having with the editorial team, it occurred to us that our method could speak to this feminist goal.

The Problem of Power and Hierarchies

As we mentioned in the beginning of our reflection, psychologists Worell and Etaugh (1994) claimed that feminist research should challenge traditional scientific methods. They urged feminist scholars to “emphasiz[e] the researcher as an individual who interacts with participants in meaningful ways that enrich both the observer and observed” (p. 446). Feminist scholars have discussed the commitment to developing transparency in more open and less hierarchical relationships with participants to emphasize the importance of a mutually meaningful relationship between researchers and participants in the cocreation of knowledge (Garko, 1999; Shields & Dervin, 1993).

In traditional conceptualizations of research, a hierarchy exists between researcher and participant where the researcher is the “expert” who has control over the study and analyses whereas the participant does not (Campbell & Wasco, 2000; Oakley, 1988). The researcher may also have other privileged identities in relation to the participant, as is the case when men study women, or White researchers study People of Color. The typical dynamic between the researcher and participant does little to facilitate trust in a relationship because the interviewee is expected to disclose personal information without any reciprocation from the researcher, and ultimately, this personal information becomes data for the researcher’s use.

As Jennifer’s supervisor for the study, Megan participated in the study with similar knowledge of the background literature, research questions, and goals as Jennifer. This mutual engagement in the research process served to reduce the inequity during the interview itself and allowed Megan and Jennifer to interact in more meaningful ways throughout the research study. Unlike the traditional hierarchy between researcher and participant, Megan and Jennifer had a shared knowledge of and connection to the study. Megan was able to feel like an active collaborator and benefited not only from sharing her own experiences with the research topics but also through reading and discussing the analyses of her own and others’ experiences with Jennifer.

Megan was, in this respect, different from a typical interviewee who has limited knowledge of the goals of the research study. However, in the present research, we are uncertain about the extent to which Megan was different; all participants were in fact very knowledgeable about the background literature surrounding gender issues and body image. For instance, the other participant who held a PhD also taught courses on women’s experiences, including issues related to sexual orientation and to body image. One of the participants with a bachelor’s degree was heavily involved in feminist activism, including issues of sexual orientation and identity, and another of the participants with a master’s degree had a long history of LGBT-centered activism. Based on their interview responses, we would say that all but one participant indicated a history of intellectual engagement with the scholarly and activist literatures that guided our study. Finally, all participants easily and quickly responded to questions about feminism, gender identity, sexual orientation, and their bodies, which Jennifer took as evidence that all had thought about (and probably read about) these issues in the past.

Although we believe that our preexisting relationship contributed to our researcher–interviewee interaction being less hierarchical than were Jennifer’s other interviews, we also recognize that Megan’s dual position in our study unintentionally created other (undesirable) hierarchies between participants. Despite the fact that many participants were highly familiar with the topic area of the study, it is still certainly the case that Megan entered into her interview with a history of involvement in the development of the research project. Some participants prefaced stories at times with comments like “I don’t know if this is relevant to what you are interested in,” or asked Jennifer questions about her own identities and her plans for the study, but Megan participated without these concerns, knowing who Jennifer was and what questions were to come in the interview. This did make her different from other participants who did not have the privilege of knowing Jennifer well or knowing about the plans for dissemination of the research, but looking back, Megan’s prior involvement may have provided the foundation for a more open and transparent interview.

Reducing Inequality Through Disclosure

Some feminist scholars have addressed the issue of the hierarchy in research by restructuring power through disclosure, where both the participant and the researcher share experiences and information about their identities or experiences (Aitken & Burman, 1999; Oakley, 1988). This approach is considered a feminist strategy because a researcher’s disclosure and openness in interviews is thought to diminish participants’ vulnerability (England, 1994; Oakley, 1988). However, this procedure does introduce the danger that the focus of the interview would shift from the participant to the researcher, which would simply perpetuate the historic silencing of women’s voices (Reinharz & Chase, 2002). Other feminists have argued that increased disclosure and intimacy during the research relationship may be exploitative because the researcher ultimately takes the participants’ stories for use as data under the guise of a more informal conversation (Stacey, 1988). Furthermore, the researcher tends to have more power than the participant over which identities to disclose and which to keep hidden, and thus she or he can continue to hide behind the professional researcher identity (Aitken & Burman, 1999).

In our research, Jennifer was committed to the idea of breaking down power structures inherent in the researcher–participant relationship with all participants. However, with most participants, she found it difficult to do so while balancing these theoretical points. In the interview with Megan, the preexisting relationship as well as Megan’s role in the project helped to achieve this balance seamlessly. We believe that our interview more easily achieved the feminist goal of reducing power disparities between researcher and participant because of Megan’s dual role.

Although Jennifer was asking the questions and had the role of “researcher” during the interview, Megan was a psychologist and researcher with a PhD, whereas Jennifer was an undergraduate student researcher. Also, Megan was Jennifer’s supervisor and professor, responsible for grading this very project. All of this contributed to our interview tipping the traditional power structure between researcher and participant (Aléx & Hammarström, 2008). Jennifer could not “hide” behind her researcher identity in the interview and was instead interviewing someone who held greater institutional power than she did.

Megan and Jennifer’s shared history and personal knowledge of each others’ identities and values also eliminated a need to balance disclosure during their interview. During the interviews in general, Jennifer chose not to disclose upfront her sexual identity or feminist identity, although she did answer honestly when asked (which some participants did). She discovered that many participants seemed to assume similarities in terms of values, but these were often not explicitly stated, which may have affected how participants chose to respond. Megan, however, knew Jennifer on a personal and a professional level. This familiarity eliminated the need for Jennifer to negotiate disclosure during the interview, where Megan was comfortable focusing on her own experiences rather than questioning Jennifer’s history.

Thus, the interview with Megan did proceed differently because she did not have to question Jennifer about her intentions, whereas the other participants often felt compelled to do so. We wonder whether participants who did not have the privilege of involvement in the research topic and design, and who were not in the position to evaluate Jennifer’s work as her supervisor, may have been more inhibited in their responses because of uncertainty about the aims of the project. This remains an open question.

We also wonder whether Megan’s position as Jennifer’s research supervisor may have inhibited Megan’s responses and whether Jennifer was uncomfortable because she was questioning one of her professors. Harkess and Warren (1993) suggest that both too much and too little closeness between interviewer and interviewee can inhibit disclosure and bias interpretation. Although Megan reports that she responded truthfully and openly, perhaps for some participants it is easier to discuss personal topics with a researcher whom one does not know. Jennifer recalls that, in the interview with Megan, she felt self-conscious at times, worrying that there could be discomfort in breaking down their professor–student relationship in a personal interview; with other participants, her concerns were instead to establish her own credibility and trustworthiness.

We wonder whether these different concerns and expectations held by Jennifer and by participants (i.e., Megan vs. other Dickinson College faculty and staff whom Jennifer knew vs. those interviewees who were complete strangers) entered into the interview process in ways that were not recognized at the time nor in hindsight. We imagine that Jennifer’s various relationships with all the participants may have affected the course of the interview and participants’ disclosure choices.

Reducing Inequity Through Collaboration

Feminist researchers have already discussed, as a feminist intervention, the need for member checks or member validations, which allow participants to comment on their transcript and clarify information prior to publication (Sprague, 2005). In our project, all participants were invited to provide this validation by reading an initial draft of Jennifer’s entire Results section (and two participants actually did so). Megan, however, was able to provide those types of comments earlier in the research process, that is, when reading drafts of individual sections of the Results, not just the final product. Similar to the other participants who agreed to read the analysis, Megan did not indicate that any of her comments in the interview had been misinterpreted. The collaborative process for which Worell and Etaugh (1994, p. 446) argue (i.e., “‘subjects’ should not become objects to be manipulated by the researcher but collaborators in the process”) was well achieved in our study because of Megan’s continued active participation. We recognize, again, that Megan was the only participant who afforded this privilege.

We also question how entitled participants felt to provide feedback when invited. Three participants did not respond at all when invited to read and comment on the draft of the Results and Discussion sections of the article. Those who did respond (including Megan) stated that they felt the analysis was truthful to their experiences. We wonder if participants who did not agree with our interpretation did not feel they had the power to say so. Megan, of course, was a more empowered participant in the process, because as the faculty supervisor of the project, she would have been able to disagree with, and modify, Jennifer’s interpretation. Thus, Megan’s position in analysis (as both a supervisor teaching her student about qualitative research and as a participant in the study) allowed her greater freedom than other participants had to disagree and to know that her disagreement would be heard.

Reflections, Implications, and Conclusions

We recognized from the outset that traditional research methods treat researchers and participants very differently and that researchers designing a traditional interview study would likely create and enforce a clear divide between the researcher and the participant. Our research project deviated from that design in ways that may be uncomfortable (it was, of course, our own discomfort over the conflicts of interest statement that prompted these discussions). In the end, we returned to the question that prompted this discussion—Is it a problem for a researcher to be part of the data set?—and mulled over for a final time why one might argue that it is. Indeed, why were we ourselves ambivalent about our practice, at times fully confident that our research and our methods were sound, and at other times concerned that our process was questionable?

We believe our ambivalent discomfort came down to the value placed on objectivity. We were both primarily trained in the research paradigm that dominates U.S. psychology, that is, quantitative research grounded in logical positivism or post-positivism. Objectivity (referring to the researcher’s distance from the research topic that is thought to allow relatively unbiased, and therefore credible, study) is a hallowed goal of this type of research. Yet, we both had stepped outside this paradigm (albeit not for the first time) while working on this research project. We imagine that a large part of our ambivalence was due to our feeling that we ourselves were straddling two (research) worlds. It was also likely due to our recognition that the majority of our peers (i.e., other U.S.-based psychological researchers) might find our choice in this research study problematic.

Of course, our ambivalence must also have been due to the fact that even within the qualitative research literature, in interview studies, and in research working from a constructionist rather than a positivist paradigm, it is uncommon for psychological researchers to be part of their own study. There are exceptions both in the psychological literature—ethnography and participatory action research, as we have noted throughout this commentary—and in other social science disciplines (e.g., autoethnography: Ellis, Adams, & Bochner, 2010; Trotter, Brogatzki, Duggan, Foster, & Levie, 2006; scholarly personal narrative: Defenbaugh, 2008; Nash, 2004; Ryan, 2006), but our study was not situated as any of these. A large part of our discomfort, then, was due to our choice to use one researcher’s interview as part of a data set in a study that, although feminist and constructionist, was not explicitly designed to include the researchers’ voice. In providing Jan with the full details of our research study, we wished to be fully transparent about our process and explain how it deviated from traditional interview studies.

Despite our discomfort and our ambivalence (or perhaps because of it), we were pushed to consider whether a strict divide between researcher and researched serves feminist goals. Just as some scholars have celebrated focus groups (Wilkinson, 2004), ethnography (Stacey, 1991), and participatory action research (Cammarota & Fine, 2008) as feminist methods because of their potential to reduce the exploitation of participants by the researcher, we ended this process of discussion and reflection with the belief that blurring the line between researcher and participant within an interview study could also be a feminist strategy serving the same purpose.

Researchers conducting interview studies from a feminist perspective might choose to provide to participants detailed information about the study prior to the interview taking place: the goals of the research project, some background literature, the interview questions themselves, and the plans for reporting and publishing. This would allow scholars to reap the benefits of a more open relationship between researcher and researched, but without the pitfall that our study included (with only one of the six participants being involved in this way).

Along with detailed information about the research project, feminist interviewers might disclose relevant personal information about themselves to build rapport before an interview (Oakley, 1988). This would be beneficial in establishing trustworthiness and credibility, and it could facilitate greater comfort and reduced power disparities in the interview—so long as the focus of the interview remained on the participant’s experiences (Campbell & Wasco, 2000).

Finally, feminist interviewers could choose to more actively involve participants throughout the research process (Creswell & Miller, 2000). Although we shared drafts with all participants, only one participant was privileged as an active collaborator throughout the entire research process. Participants could be asked for member checks at multiple stages of data analysis as well during the write-up of the results. Increasing and maintaining collaboration for all participants throughout the process of data analysis and publication would provide a more egalitarian and empowering experience for participants to have their voices heard (Karnieli-Miller, Strier, & Pessach, 2009).

We refrain from recommending that feminist researchers allow themselves to be interviewed in their own study, because (as we hope this commentary has conveyed) our thinking on this practice is evolving. However, we do think it is fair to say that many research studies are initiated because the researcher has personal experience related to the question at hand. Whereas scholars in some disciplines may deride this as “me-search,” implying that good research cannot be personal, those in other disciplines (feminist psychology included) value the richness, the deep desire to understand a phenomenon, and the passion that a personal connection with one’s research brings (DuBois, 1983). This leads us to ask: What points of view are left out of research reports when researchers do not contribute their voice to the data? What perspectives and theoretical advances are lost when researchers refrain from allowing their voices to explicitly inform the final manuscript, as data or otherwise? For instance, if we had systematically included both of our stories in the data set, Jennifer’s story (an adolescence marked by body image issues, objectification, and a developing bisexual identity, followed by college years characterized by positive and affirming feminist coursework and supportive experiences in feminist and LGBT student organizations) might have further enhanced our understanding of bisexual women’s body image.

When psychological researchers are amenable to the idea that knowledge is constructed rather than objectively discovered, perhaps blurring the line between researcher and participant would actually enrich the empirical literature. Rather than bias data, acknowledging researchers’ own experiences by explicitly including their perspectives in the research process may serve to further ground feminist research in lived experience. Our hope is that our willingness to bring these issues to the forefront and to seek additional reflections from other feminist scholars will help further our discipline's thinking about our own roles in the research process.

Footnotes

Acknowledgments

We would like to thank Jan Yoder, Maria Gurevich, and Heather Bullock for raising many of the questions addressed in this commentary.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Note

References

Aitken

Burman

(1999). Keeping and crossing professional and racialized boundaries: Implications for feminist practice. Psychology of Women Quarterly, 23, 277–297.

Aléx

Hammarström

(2008). Shift in power during an interview situation: Methodological reflections inspired by Foucault and Bourdieu. Nursing Inquiry, 15, 169–176.

Burr

(2003). Social constructionism. New York, NY: Routledge.

Cammarota

Fine

(2008). Revolutionizing education: Youth participatory action research in motion. New York, NY: Routledge.

Campbell

Wasco

S. M.

(2000). Feminist approaches to social science: Epistemological and methodological tenets. American Journal of Community Psychology, 28, 773–791.

Chmielewski

J. F.

Yost

M. R.

(2013). Psychosocial influences on bisexual women’s body image: Negotiating gender and sexuality. Psychology of Women Quarterly, 37, 242–250.

Corbin

Strauss

(2008). Basics of qualitative research (3rd ed.). Thousand Oaks, CA: Sage.

Creswell

J. W.

Miller

D. L.

(2000). Determining validity in qualitative inquiry. Theory into Practice, 39, 124–130.

Defenbaugh

N. L.

(2008). Under erasure: The absent ill body in doctor-patient dialogue. Qualitative Inquiry, 14, 1402–1424.

10.

Dilley

(2000). Conducting successful interviews: Tips for intrepid research. Theory into Practice, 39, 131–137.

11.

DuBois

(1983). Passionate scholarship: Notes on values, knowing and method in feminist social sciences. In Bowles

Klein

R. D.

(Eds.), Theories of women’s studies (pp. 105–117). London, England: Routledge and Kegan Paul.

12.

Ellis

Adams

T. E.

Bochner

A. P.

(2010). Autoethnography: An overview. Forum Qualitative Sozialforschung [Forum: Qualitative Social Research], 12. Retrieved from http://nbn-resolving.de/urn:nbn:de:0114-fqs1101108

13.

England

K. V. L.

(1994). Getting personal: Reflexivity, positionality, and feminist research. Professional Geographer, 46, 80–89.

14.

Frankland

Bloor

(1999). Some issues arising in the systematic analysis of focus group materials. In Barbour

R. S.

Kitzinger

(Eds.), Developing focus group research (pp. 144–155). London, England: Sage.

15.

Garko

M. G.

(1999). Existential phenomenology and feminist research: The exploration and exposition of women’s lived experiences. Psychology of Women Quarterly, 23, 167–175.

16.

Gilbert

M. R.

(1994). The politics of location: Doing feminist research at “home.” Professional Geographer, 46, 90–96.

17.

Grafanaki

(1996). How research can change the researcher: The need for sensitivity, flexibility and ethical boundaries in conducting qualitative research in counseling/psychotherapy. British Journal of Guidance and Counseling, 24, 329–338.

18.

Harding

(1987). Feminism and methodology. Bloomington, IN: Indiana University Press.

19.

Harkess

Warren

C. A. B.

(1993). The social relations of intensive interviewing: Constellations of strangeness and science. Sociological Methods and Research, 21, 317–339.

20.

Holloway

(1997). Basic concepts for qualitative research. Oxford, England: Blackwell Science.

21.

Karnieli-Miller

Strier

Pessach

(2009). Power relations in qualitative research. Qualitative Health Research, 19, 279–289.

22.

Maxwell

J. A.

(1998). Designing a qualitative study. In Bickman

Rog

D. S.

(Eds.), Handbook of applied social research methods (pp. 69–100). Thousand Oaks, CA: Sage.

23.

Mays

Pope

(1995). Rigour and qualitative research. British Medical Journal, 311, 109–112.

24.

McIntyre

(2007). Participatory action research. Thousand Oaks, CA: Sage.

25.

Morrow

S. L.

(2005). Quality and trustworthiness in qualitative research in counseling psychology. Journal of Counseling Psychology, 52, 250–260.

26.

Nash

(2004). Liberating scholarly writing: The power of personal narrative. New York, NY: Teachers College Press.

27.

Oakley

(1988). Interviewing women: A contradiction in terms. In Roberts

(Ed.), Doing feminist research (pp. 30–61). New York, NY: Routledge.

28.

O’Shaugnessy

Krogman

N. T.

(2012). A revolution reconsidered? Examining the practice of qualitative research in feminist scholarship. Signs: Journal of Women in Culture and Society, 37, 493–520.

29.

Ralph

(1988). Researching from the bottom: Lesson of participatory research for feminists. In Currie

(Ed.), From the margins to the centre: Selected essays in women’s studies research (pp. 134–141). Saskatchewan, Canada: The Women’s Studies Research Unit, University of Saskatchewan.

30.

Reinharz

(1992). Feminist methods in social research. New York, NY: Oxford University Press.

31.

Reinharz

Chase

(2002). Interviewing women. In Holstein

Gubrium

(Eds.), The handbook of interview research: Context and method (pp. 221–238). Thousand Oaks, CA: Sage.

32.

Ryan

(2006). Researching Irish gay male lives: Reflections on disclosure and intellectual autobiography in the production of personal narratives. Qualitative Research, 6, 151–168.

33.

Shields

V. R.

Dervin

(1993). Sense-making in feminist social science research: A call to enlarge the methodological options of feminist studies. Women’s Studies Forum, 16, 65–81.

34.

Smith

J. A.

Eatough

(2007). Interpretative phenomenological analysis. In Lyons

Coyle

(Eds.), Analysing qualitative data in psychology (pp. 35–50). Thousand Oaks, CA: Sage.

35.

Sprague

(2005). Feminist methodologies for critical researchers: Bridging differences. New York, NY: AltaMira Press.

36.

Sprague

Kobrynowicz

(2003). A feminist epistemology. In Hesse-Biber

S. N.

Yaiser

M. L.

(Eds.), Feminist perspectives on social research (pp. 78–98). New York, NY: Oxford University Press.

37.

Stacey

(1988). Can there be a feminist ethnography? Women’s Studies International Forum, 11, 21–27.

38.

Stacey

(1991). Can there be a feminist ethnography? In Gluck

S. B.

Patai

(Eds.), Women’s words: The feminist practice of oral history (pp. 111–119). New York, NY: Routledge.

39.

Trotter

Brogatzki

Duggan

Foster

Levie

(2006) Revealing disagreement and discomfort through auto-ethnography and personal narrative. Qualitative Social Work, 5, 369–388.

40.

Wilkinson

(2004). Focus groups: A feminist method. In Hesse-Biber

S. N.

Yaiser

M. L.

(Eds.), Feminist perspectives on social research (pp. 271–295). New York, NY: Oxford University Press.

41.

Wolf

(1996). Situating feminist dilemmas in fieldwork. In Wolf

(Ed.), Feminist dilemmas in fieldwork (pp. 1–55). Boulder, CO: Westview.

42.

Worell

Etaugh

(1994). Transforming theory and research with women. Psychology of Women Quarterly, 18, 443–450.

43.

Zavos

Biglia

(2009). Embodying feminist research: Learning from action research, political practices, diffractions, and collective knowledge. Qualitative Research in Psychology, 6, 153–172.