Political elites are people who “exercise disproportionately high influence on the outcome of events or policies” (Pierce Reference Pierce2008, 119). Elites frequently have access to information that is not available from official sources (Al’Abri et al. Reference Al’Abri, Hogan, Lingard, Sellar, In Stacey and Mockler2024; Lilleker Reference Lilleker2003). Interviewing elites is therefore valuable to political scientists who study phenomena that take place behind closed doors. Elite interviews can also help researchers move from description towards interpretation (Fujii Reference Fujii2017) and understanding (May Reference May2011).
A number of articles that explore various facets of elite interviewing have appeared in this journal and elsewhere (Berry Reference Berry2002; Glas Reference Glas2021; Goldstein Reference Goldstein2002; Porisky and Glas Reference Porisky and Glas2023; Soedirgo and Glas Reference Soedirgo and Glas2020). Recurring themes include how “insider” and “outsider” status influence positionality and the limitations of this binary framework as well as the challenges associated with elite interviews and how these interact with positionality. This article explores these themes from a new perspective. I begin by explaining what I term “in-betweener status,” showing how this builds on existing scholarship. I then discuss four previously documented methodological considerations for researchers conducting elite interviews, reflecting on the opportunities my in-betweener status afforded in relation to these—along with several drawbacks. I suggest that these insights have relevance to researchers at all different points on the insider–outsider continuum.
I draw on my experience conducting 22 interviews with education policy makers in England while researching a book titled How Policy Happens (Menzies, Reference Menziesforthcoming.) Interviewees included former ministers, civil servants, advisors, and think-tank staff. I completed structured reflections (Al’Abri et al. Reference Al’Abri, Hogan, Lingard, Sellar, In Stacey and Mockler2024; Berger Reference Berger2015) focused on central methodological themes within a day-or-two of 20 interviews. Additionally, when analyzing the transcripts, I coded approximately 70 excerpts that exemplified the methodological considerations. I also draw comparisons with fieldwork conducted in other jurisdictions where my positionality differed considerably.
INSIDERS, OUTSIDERS, AND IN-BETWEENERS
Scholars frequently assume a “categorical and consequential distinction” (Glas Reference Glas2021, 438) between “elite” and “nonelite” as well as insiders and outsiders (Porisky and Glas Reference Porisky and Glas2023). In political research, these categories often overlap such that researchers are frequently characterized as both nonelite and outsiders. Yet neither binary is “monolithic,” (Glas Reference Glas2021, 438) “discrete,” or “static” (Porisky and Glas Reference Porisky and Glas2023, 52). Scholars have described how status can change and trust accumulate over successive interviews without the researcher necessarily acquiring insider or elite status (Fujii Reference Fujii2017; Porisky and Glas Reference Porisky and Glas2023). However, Savage warns that over time, political researchers can become “networked into the very elite policy networks they seek to understand and critique” (Savage et al. Reference Savage, Gerrard, Gale and Molla2021, 312). Meanwhile Berger (Reference Berger2015) describes moving from outsider to insider status as her personal circumstances changed. Thus, few researchers will be “true” insiders or outsiders “from beginning to end” (Fujii Reference Fujii2017, 19).
My experiences as a former think-tank leader who now both researches and participates in policy making exemplify the shortcomings of a rigid, binary view of positionality. I use the term “in-betweener” to capture this for two reasons. First, my research continually takes me back and forth along the insider–outsider continuum. In some cases, I interview elite actors with whom I have worked for over a decade or whom I know personally. In such cases—as this article makes clear—I often bear the hallmarks of an elite insider. In other cases, I interview elite figures who were active before I joined the sector or in countries where I have not worked. Although I still benefit from some residual access and familiarity, my experiences can be more akin to those of the archetypal outsider. Second, I am aware from my experience that within government, “insiderness” can operate as a series of concentric and multidimensional circles, each with varying degrees of power and influence. Although my policy work has drawn me into some of these circles, I know that there are inner circles from which I am excluded. Because many of my interviewees are at the center of such rings, I do not always feel like—nor am I seen as—a “true insider.” This is exacerbated by the fact that I have never held the official Westminster roles that are often considered the currency of insider status.
“Insiderness” can operate as a series of concentric and multidimensional circles, each with varying degrees of power and influence.
In the field of cultural research, Kipnis, Bebek, and Brőckerhoff (Reference Kipnis, Bebek and Brőckerhoff2021) compiled terms that describe intricate facets of insiderness and outsiderness, but none fully capture the status I am describing. One of these is “boundary spanner” (Ball and Junemann Reference Ball and Junemann2012), a term used to describe the dual identity (Berger Reference Berger2015) of researchers who straddle policy and academic worlds. However, this does not capture the fluid and situation-dependent gradations of insiderness described in the present article. Meanwhile, the term “returning insider” implies leaving and returning to a status rather than occupying multiple positions simultaneously. The term “partial insider” is closer to my experience but is problematically ambiguous because it is easily conflated with the (albeit related) concept of partiality, or bias. It also obscures the fact that a researcher may be a core insider at times but peripheral at others. Finally, Porisky and Glas propose the methodologically useful term “credible visitor,” founded on humility, reflexivity, competence, and transparency. Scholars at all points on the insider–outsider continuum would be well advised to honor these “commitments” (2023, 53), but the term itself is less appropriate for a researcher conducting interviews on their home turf for whom honoring such commitments may look rather different. The term “in-betweener” therefore seems well suited to capturing the liminal status I have experienced. Moreover, given that all researchers of elites are to some extent in-betweeners, the reflections set out below have the potential to aid the pursuit of reflexivity for a range of scholars.
FOUR CHALLENGES TO ELITE INTERVIEWING
Lillleker (2003) summarizes widely documented challenges in elite interviewing under four categories: access, securing answers, veracity, and dealing with the data that are gathered. In the following sections, I reflect on how my positionality influenced my experiences in each of these areas.
Access
The literature is replete with advice on overcoming barriers to access (Goldstein Reference Goldstein2002; Lilleker Reference Lilleker2003; Pierce Reference Pierce2008). However, as an in-betweener, relative proximity to some potential interviewees and distance from others led me to take a more personalized approach. Rather than defaulting to formality, I tailored my language and approach to the individual, often making contact via direct messages on Twitter or Whatsapp. I suspect that this informal style set the tone for open and relaxed interviews. Although being close to a target interviewee and having name recognition made this strategy easier, I also tried to anticipate which approach individuals would respond to best, sometimes successfully applying the same tactics with people I knew less well.
Most of the time, when I did not know an individual I requested introductions from existing contacts so that my primary contact could advocate for my trustworthiness (McHugh Reference McHugh and Walford1994). However, whereas Beckman and Hall (2013) recommend starting with “less important” interviewees and building up, I preferred to seek introductions from previously secured, high-status interviewees who could signal that participation was worthwhile. Sometimes, high-status contacts from the period in which I have been active in the policy world opened the door to those from earlier periods or in other countries. Researchers with potential access to at least some high-status interviewees may benefit from securing some of these individuals early.
Despite the importance of trust and informal ties, these can be trumped by institutional constraints and personal preferences. I was lucky that most individuals I approached quickly agreed to be interviewed, but one former secretary of state—who had initially said they would be “delighted” to participate—cancelled following a return to government, as the interview was no longer compatible with their new role. Another refused my request, despite us having an established relationship, explaining they had received too many similar requests lately and that their reflections were already available in their published diaries. A third secretary of state (whom I did successfully interview) offered some insight into this, saying, “I get fed up of having to listen to myself.” Multiple factors, beyond defensiveness and time constraints, can therefore be barriers to interviewing elites, and researchers should be wary of assuming that their position on the insider–outsider continuum will translate directly into ease or difficulty of access.
SECURING ANSWERS
Elite interviewees are “accomplished definers of reality” (Whitty and Edwards Reference Whitty, Edwards and Walford1994, 18), and power dynamics can allow them to avoid interviewers’ questions (Lilleker Reference Lilleker2003). Therefore, conversations can be dominated by interviewees’ promoting their achievements and attempting to shape their legacy (Glas Reference Glas2021).
Some scholars argue that interviewees are more open with insiders, lending “greater depth to the data” (Dwyer and Buckle Reference Dwyer and Buckle2009, 58). I certainly found that many of my interviewees were comfortable discussing past errors and sometimes enquired into my own recollections and interpretations. For example, despite already having overrun our allotted time, one interviewee asked, “have I said anything that surprised you at all?” and then “any other sort of feedback or thoughts from your perspective on what I have said?” This prompted an extended casual conversation in which we compared personal recollections and discussed similarities and differences in her account compared with that of other interviewees. This led to new insights into why she was able to exercise a particularly high degree of control over the policy agenda. It is not unusual for researchers to remark on the value of informal conversations at the end of structured interviews, but in this case the door to a fruitful discussion was opened by our existing relationship and my interviewee’s interest in my recollections and interpretation.
Some interviewees with whom I was less close appeared stuck in “broadcast” mode. I reflected in my postinterview diary that one policy maker from an earlier period “had notes of stuff he wanted to flag about ‘good policy making.’” Similarly, one of his colleagues concentrated on explaining “how the policy world works” rather than responding to my questions. My notes recall that “I had to be quite directive, which wasn’t ideal.” On the other hand, an interview is always the product of the interaction that emerges between two (or more) individuals (Fujii Reference Fujii2017) and status does not always “map neatly onto elite interactions” (Glas Reference Glas2021, 440). There was a stark contrast between my two broadcast-style interviews and my discussion with a former colleague of these two individuals despite the dynamic being the same “on paper.” In my notes from that third interview, I remarked that the interviewee was “obviously very humble about things that went wrong—that’s their overall approach, so big lack of defensiveness.” The contrary can also apply. The interviewee with whom I had the closest relationship was noticeably reticent during our discussion, and I remarked in my notes that he seemed to be in “a different role to how I know him, ‘playing the civil servant.’” Helpfully, after the interview he was able to explain his sense of having been restricted by the civil service code, underlining the fact that institutional rules can outweigh personal connections. But rules always require interpretation, and, shortly after, the two of us met at a party along with another interviewee and the three of us discussed the interviews. The second individual explained they had not felt constrained—despite us being less close and them being subject to the same rules. Their view was that it had been possible to mentally set aside certain topics and speak freely outside of that. On another occasion, a third individual reported feeling completely comfortable because they trusted me to handle the data sensitively. Finally, a fourth, whom I had known for many years, declined the interview after consulting a more experienced colleague, fearing an interview might violate the Official Secrets Act. These four cases demonstrate that it is not just interviewers’ (and interviewees’) status or role that determines how interactions unfold but individuals’ personality and their interpretation of institutional rules. As Al’Abri et al. note, learning how elites navigate these dynamics can sometimes provide as much insight “as the actual content of the interview” (2024, 236).
Institutional rules can outweigh personal connections. But rules always require interpretation
Elites can be reluctant to discuss certain topics, and it is easy to overlook their vulnerability (Neal and McLaughlin Reference Neal and McLaughlin2009). One approach to securing answers is to provide outline questions in advance (Lilleker Reference Lilleker2003). My strategy was to ask interviewees to propose a policy that they were involved in to discuss in detail. They did this while completing a consent form in advance. Handing over some control ensured that participants were happy to discuss the policy and anchored the discussion in a specific case (Beckmann and Hall Reference Beckmann and Hall2013) while giving me the flexibility to lead discussion where I wanted. There was a risk that they might choose unrepresentative examples, but most selected more than one policy—often a typical and an unusual case or one that they deemed successful and another that they deemed unsuccessful. Atypical cases proved especially useful because they prompted discussion of how things normally happen. Unfortunately, some interviewees did not find time to inform me of their choice in advance and others changed their mind at the last minute. This made preparation difficult, but it was not normally a problem when discussing policies from my period in England given my background knowledge. It was more challenging in less-familiar periods and contexts. Nonetheless I have used the same technique in other countries and simply played the role of “clueless foreigner,” asking interviewees to explain things from first principles, a process which can itself prove insightful (Fujii Reference Fujii2017, 66).
Veracity
The significance and nature of veracity challenges in elite interviews depend on the researcher’s underlying ontology and epistemology, particularly whether they are seeking to establish objective facts in the positivist tradition or are taking a more interpretivist approach (Fujii Reference Fujii2017; Savage Reference Savage2020). In line with a more positivist approach, Lilleker emphasizes the need to cross-reference and triangulate, drawing on “good knowledge of the facts” (2003, 212).
When operating at the proximate end of the insider–outsider continuum, I had two advantages in seeking veracity. First, preestablished trust and status ensured that I was comfortable with robustly challenging interviewees’ accounts. Second, I was able to draw on extensive prior knowledge to construct counterfactuals and point to alternative or contradictory facts. The case of what one interviewee described as a “fuck up” illustrates both. My interviewee began by blaming two individuals for a policy blunder, and I followed up by asking why an alternative course of action had not been followed. The interviewee avoided responding, instead justifying his actions. I then sought further reflection by sharing first-hand recollections of the event’s context. At this point he conceded, “the way we … the way NAME 1 [a minister] announced it, was a fuck up,” self-correcting the initial “we” to focus the blame on a politician before going on to blame the prime minister’s office. I then elaborated on my recollections, stating, “I remember being in a room with NAME 2 [another minister] and NAME 1 around then and this kind of mood that there was that was just like “obviously there’s so much waste happening”—so it was almost a jubilation at the fact that it was possible to make these cuts.”
My experiences as a former think-tank leader who now both researches and participates in policy making exemplify the shortcomings of a rigid, binary view of positionality
At this point, he blamed the shadow secretary of state and speculated about a possible insider conspiracy. My postinterview reflections record that “[we] began by both saying it was good to finally sit down together after years in each other’s orbit. I felt aware that [NAME] is a clever spinnerFootnote 1 so was conscious of the presentation of things to me… . Because we were talking about policies that I’d spent years critiquing, I drilled into quite specific critiques of the policy [and] wondered if I’d gone too far down this line, but I also know that NAME likes a good hearty debate.”
My experiences as a former think-tank leader who now both researches and participates in policy making exemplify the shortcomings of a rigid, binary view of positionality
With extensive preparation, a credible visitor might have secured access to the insider knowledge needed to challenge the proffered account. Some researchers are also comfortable summoning the courage to question elites in a robust manner. However, in the game of this interaction my status clearly afforded privileges and shortcuts. Although we had never met before and I had orbited in a more peripheral concentric circle, it was relatively easy for me to tap into shared connections and evocative examples. Previous, similar interactions with his entourage also lent me additional confidence. It is worth noting though that the exchange did not yield the definitive version of events that a positivist researcher might seek. Instead, it yielded insights into how an advisor in this role might make sense of and present events.
Proximity can also present challenges when pursuing veracity. A reflexive approach is therefore essential. Savage et al note that studies of elite networks can constitute “research of elites, by elites, for elites” (2021, 307), reproducing group-think and exacerbating social desirability bias. In-betweeners might be particularly tempted to appear “on the same side” as interviewees (Whitty and Edwards Reference Whitty, Edwards and Walford1994, 22) because their status can be less secure than that of true insiders. Moreover, status will have particularly significant, material consequences for those who are still professionally active in the policy field. Additionally, researchers who study events with which they were personally involved may share the same desire as the elites they interview to shape their legacy. Therefore, I have had to be particularly open about my positionality when researching topics that concern me personally, like the influence of experts in recent education policy formation (Menzies Reference Menzies2024). I have also found that interpretivism sometimes better recognizes—and values—the meaning-making opportunities afforded by my unusual positionality.
Dealing with the Data Gathered from Interviews
Lilleker’s fourth and final challenge is “how to reference data gained from interviews” (2003, 212). On on hand, knowing who said something is critical when seeking to understand elite testimony, but naming can heighten the risk of self-censorship (Walford Reference Walford and Walford1994).
The consent forms my interviewees completed confirmed that their comments would not be publicly linked to their name but also gave them the option of appearing in a compiled list of interviewees. Most chose to appear on this list and one former minister told me they appreciated this opportunity for their legacy to be acknowledged. Although the consent form set the formal parameters for how data would be dealt with, interviewees held additional expectations for how their accounts would be used. Ball reports similar experiences, noting that “one or two people expressed the hope that I would treat their more indiscreet contributions sensitively” (Reference Ball and Walford1994, 98). Trust is a precious commodity, and expectations around careful handling and sensitive judgement may be heightened when a researcher has tapped into reservoirs of trust. Therefore, I have often checked quotations with interviewees before using them—even in anonymous form—judging that going above and beyond is the ethical thing to do. One interviewee thanked me for doing so, remarking “it’s very decent of you,” whereas another specifically asked for the quotation to be attributed.
Conversely, I sometimes feel able to go beyond formal constraints by requesting explicit permission to attribute uncontentious quotes. Considerations in doing so include my understanding of how comments are likely to be interpreted within the interviewees’ elite community, my sense of how the individual wants to be perceived, the likelihood that an informed insider might identify them, and the extent to which the interviewee would mind this. Clearly, ethical conduct demands that all assurances given in a consent form be honored, but more intimate knowledge of an interviewee’s context can afford opportunities for more fine-grained judgement. In contrast, when I am less familiar with an interviewee and their context, I perhaps wrongly tend to take a more standardized approach. Legacy is so important to many elites that taking time to deal with interviewees’ data in a personalized consultative way—guided by more than just risk aversion—can constitute a form of reciprocity between researchers and their research subject.
Legacy is so important to many elites that taking time to deal with interviewees’ data in a personalized consultative way—guided by more than just risk aversion—can constitute a form of reciprocity between researchers and their research subject.
CONCLUSION
I began this article by arguing—in line with Glas and colleagues—that a binary view of elite and nonelite is logically and practically problematic (Glas Reference Glas2021; Porisky and Glas Reference Porisky and Glas2023; Soedirgo and Glas Reference Soedirgo and Glas2020). I reviewed various terms for researchers’ status and suggested that in-betweener captures something important about the insider–outsider continuum that is particularly relevant to my unusually liminal positionality. I then reviewed a series of challenges faced by researchers when interviewing elites, reflecting on how these challenges manifested in the context of my in-betweener status.
As this article shows, proximity does not always play out in straightforward or predictable ways. Interviewees’ institutional positions, preferences, personality, and interpretation of rules can trump relational proximity. Proximity can also bring challenges: although personal involvement provided me with the insight and confidence to probe veracity robustly, it undoubtedly shaped my interpretation, demanding heightened reflexivity. Finally, although trust can encourage frankness, it also imposes additional unspoken obligations on the interviewer. However, regardless of a researchers’ status, careful judgement about data handling—within ethical bounds—can be an opportunity for respectful reciprocity.
ACKNOWLEDGMENTS
The international fieldwork referenced in this article was generously funded through two grants from the Charitable Foundation for Education Development.
CONFLICTS OF INTEREST
This article describes a series of tensions and loyalties that could arguably be considered nonfinancial conflicts of interest.
Open access
