Interviews with ‘policy elites’ constitute crucial sources of information in policy research. To explain how policies are developed and to unpack decision-making, we need to open up the ‘black box’ and look inside the closed-off negotiation processes that very few individuals are privy to. There are some unique challenges related to ‘researching up’ that sets policy elite interviews apart from much social science research. The policy research literature highlights some of the main challenges (e.g., Desmond 2004; Lancaster 2017; Natow 2019; Ross 2001). These challenges include getting access to the policy elites (access), the risk of the elite presenting biased information (biased information) and the power asymmetry between researcher and elite (positionality). The quality of the data is affected by these issues and particularly when researching highly politicised topics, it can be tremendously difficult to collect reliable information which is sometimes even purposely hidden from the public domain. Interviews with policy elites can be the only source of information; however, because elites have stakes in how these processes are presented and disseminated, the information they share might be biased. Furthermore, they are capable of communicating facts in a favourable manner because they are trained to be strategic in how they portray situations and in what information to share. The findings from this paper, while being lessons drawn from conducting elite interviews, have relevance beyond. Elite interviews can be considered an ‘extreme case’ regarding these three challenges, nevertheless such challenges emerge when conducting interviews more generally, such as during expert interviews.

This paper specifically addresses the three challenges identified in the elite interview literature, and it considers ‘policy elites’ to be actors who have disproportionately high levels of influence on policymaking processes (see Henriksen and Seabrook 2021). Even though there are several serious issues relating to the data quality of policy elite interviews, they remain a highly relevant method and are widely practiced. Surprisingly, no-one has so far discussed the effects of conducting elite interviews in an online format, through video-chatting apps such as Skype, Zoom, Microsoft Teams or WhatsApp. While there are studies that discuss conducting interviews online, methodological research written for non-elite settings is inadequate to instruct elite researchers because they face very different challenges (Stephens 2017). This paper considers the following research question: Does online policy elite interviews provide sufficient data quality to provide valuable findings? To address this question, this paper reflects on the data collected through interviews with decision-makers in the European Commission (hereafter the Commission) on a highly politicised topic, using audio-visual programs or phone.

The existing general online methods literature does address the methodological implications of conducting (non-elite) interviews online rather than face-to face. Despite some noteworthy exceptions, the majority of the online methods literature is pessimistic and suggests that conducting interviews online tend to reduce data quality. This literature conjectures that an online format increases access to potential interviewees (Bowker and Tuffin 2004:230; O’Connor et al. 2008) while reducing the quality and quantity of data produced in the interviews (Cater 2011; Weller 2017). This paper makes a novel contribution by demonstrating that the online format of elite interviews can actually produce deep and unique insights into highly controversial topics. In concurrence with existing online methods research, I find that the online approach increases accessibility to interview participants (Bowker and Tuffin 2004; O’Connor et al. 2008). However, unlike the existing literature, I find that the online format increases data quality. Furthermore, the loss in data quantity is insignificant. In conclusion, I argue that the online setting can produce high-quality data that would be very difficult to obtain in any other way. In fact, an online approach can mediate several of the challenges associated with elite interviews.

These findings are significant because, looking ahead, online research will likely become increasingly relevant, particularly for interviews with policy elites. The pandemic has changed work routines and increased technological competence among researchers and elites alike. Moreover, the economic and environmental costs associated with travelling make online meetings attractive. Online interviews are more cost-effective and climate-friendly than alternative approaches, and it has recently become more accessible than ever. Thus, the online format creates opportunities for future policy researchers to gain insights into closed negotiations and decision-making processes.

The paper is structured as follows. Section 2 reviews the general online methods literature and present the expectations it raises on how the online format affect the data collection. Section 3 presents existing research on elite interviews and identifies the main challenges with this method. Section 5 presents the exploratory case study that the discussion of this paper centres on. Section 6 discusses the quality of the data collected using online interviews. Finally, in section 7, I conclude and offer some recommendations for future policy researchers.

Expectations derived from the online methods literature

A review of the online methods literature provides some expectations on how conducting interviews online affect the resulting data (Adams-Hutcheson and Longhurst 2017; Deakin and Wakefield 2014; Hanna 2012; Johnson et al. 2019; Oates 2015; Seitz 2016; Weinmann et al. 2012; Weller 2017). Online interviews are one-to-one simultaneous interactions and do not in this case include pre-planned surveys. While most of the literature remains pessimistic about the online format, several works highlight one major advantage: The online format can facilitate access to individuals who are otherwise difficult to reach (Bowker and Tuffin 2004:230; O’Connor et al. 2008). Even though these works focus on vulnerable groups such as people who are immobile or suffer from a severe illness, it might arguably also apply to people with very busy schedules due to their work.

The COVID-19 pandemic and the subsequent lockdowns have by necessity caused a recent surge in online methods. This trend has not only reiterated the importance of online methods research, it has also demonstrated how outdated much of this literature has become. Many studies focus on the peculiarity of the online setting for interviewer and interviewee alike, while others consider the technological issues that may arise during an online interview. Some of these issues are now more or less outdates, especially when researching elites, while other issues have emerged. Even so, the online methods literature has produced two main substantial criticisms of online interviews. The first concerns the quality of information, in particular that the online setting does not create intimacy and mutual trust (Jowett et al. 2011; O’Conner et al. 2008). Weller (2017:614) stresses the salience of interview rapport, defined as ‘euphoria’ or ‘ease’ in interaction, a harmonious connection or a ‘working consensus’, for creating mutual trust and enabling information flow in an interview setting. Because online methods try to mirror face-to-face interaction despite being something else entirely, it is difficult to build rapport. Thus, it is also difficult to get the interviewee to share information openly. Without rapport, it is hard to imagine that a policy elite would be willing to share insights into highly controversial decisions and policy processes.

The second main criticism the online methods literature presents relates to the quantity of information that can be collected in an online interview. The researcher can only register what is caught by the camera and the microphone, which does not cover all the information that would be obtained if the interviewer was present in person (Cater 2011). This criticism builds on a recognition of the importance of the information conveyed through body language and non-verbal communication.

To summarise, the online interviews literature raises some expectations about access, data quantity, and data quality. However, it does not address how going online effects elite interviews which have their very unique set of challenges associated with them.

Defining ‘policy elite’ and identifying the main challenges of elite interviews

‘Policy elites’ are individuals who have disproportionately high levels of influence on policymaking processes and policy implementation processes. In this paper, decision-makers in the Commission of the European Union make up the policy elites. They are in an influential position because they can influence policy outcomes and they have knowledge about the policy processes they engage in. This means that they have expert knowledge on the policy processes the researcher is interested in, but it also means that they have stakes in how these processes are disseminated. This influences what information they choose to share and what issues they choose to highlight or downplay in the interviews. Furthermore, because of their position, they are likely skilled in communication and know how to persuade and convince third parties. Elite interviews can be considered a subcategory within expert interviews (Littig 2009). Littig (2009) argues that an interviewee can be an elite and an expert, and that it is the epistemological premise of the data collected that differentiates the two categories. Information about the order of events is for example something an expert interview can account for, whereas the underlying assumptions and understandings held by individuals can be revealed through elite interviews (Littig 2009). This paper, situated in the practical methodological discussions about online interviews uses the term elite interviews because it speaks specifically to challenges raised in this strand of literature (see Ross 2001). One of the important differences between expert interviews and elite interviews in this regard is how power and bias influence the outcome of the data obtained (Liu 2018, van Audenhove and Donders 2019). My definition of policy elite is not to be mistaken for social elites, which refers to people of high social status, such as is the case in Stephens (2017) study on telephone interviews. Ultimately, however, several findings from the study have implications for the broader expert interview literature.

Political science scholars are often concerned with explaining policy outcomes. Negotiation theory, organisational theory and public administration theory can guide the researcher in doing so, focusing the research on the dynamics of the policy processes and how they shape policy outcomes. What actors are involved, how they are involved, what their preferences are, how these preferences are shaped, and how they act to achieve their preferences can be necessary pieces of information. This is information that can be difficult to obtain and is sometimes even purposely hidden from public domain. Interviews with policy elites are crucial sources of information in this type of research (Lancaster 2017; Natow 2019). The elite interview research literature describes the unique challenges associated with the method. It can be challenging to gain access to the elites (e.g. Dexter 1970; Glassner and Hertz 1999; Hertz and Imber 1995; Kezar 2008; Odendahl and Shaw 2002); the elite can intentionally and unintentionally provide biased information (DiMaggio 2014; Lancaster 2017; Natow 2019; Ross 2001:160; Smith 2006:644); the positionality of the elite create an unbalanced relation with the researcher (Liu 2018, Ross 2001; Smith 2006; van Audenhove and Donders 2019). All of these three issues are deliberated in further detail in the following discussion sections: Access to elites online; Biased information from elites online; Positionality in an online context. When researching politicised issues, such as migration policy, it is likely even more difficult to get elites to agree to be interviewed because they are unwilling to share information. And even if they agree to participate, it is likely difficult to get them to share openly during the interview. Because of the politicised nature of the topic under discussion, it is difficult for the elite to openly share information about the policy processes and at the same time present their own role, or the role of their institutions, in a favourable light.

The case: online interviewing Commission representatives on a highly controversial topic during lockdown

The case study at hand is based on my own experience from conducting 20 elite interviews online with decision-makers in the Commission, between April and September 2020. The topic discussed in the interviews was a highly politicised one, namely EU policy responses to the 2015 migration crisis. The EU denied entry by Syrian refugees and other migrants, it made deals with undemocratic regimes such as Turkey and Libya, and more than 20.000 people died in the Mediterranean Sea trying to reach Europe (European Parliament 2021:80). Arguably the most salient policy area in the EU, the migration crisis offers an excellent issue for investigating how to get good, reliable data from policy elites. The more politicised the topic, the more difficult it is to gain access to participants and to collect high-quality data by interviewing them. This paper considers a case in which one would expect it to be very difficult, or unlikely, to obtain access to participants and to obtain high-quality data and it is therefore a ‘least likely’ case (Gerring 2010:116). If online interviews can produce high-quality data in a least likely case, then the method will likely be appropriate in other cases as well.

Although the case study does not offer a comparative perspective including in-person interviews, the case is highly relevant for addressing the knowledge gap concerning online elite interviews. Most interviews were conducted on Zoom or Skype, while a minority used WhatsApp or Microsoft Teams. Personnel from different Commission units had different preferences, because of security reasons or habit. A couple of interviews were off-camera for the entire duration due to preference, while a couple of others were off-camera in short periods, due to a fragile internet connection. The different formats and their implications for the data collection will be discussed below. The interviews were semi-structured; following an interview guide defined after having analysed relevant policy documents. I first interviewed people who were listed on the Commission’s website as having highly relevant positions. To gain access to these individuals, I drafted emails with explicit reference to the policy process I was interested in learning about, and I tied the elites’ position in the organisation to my request. For example, I would ask how a ‘senior officer’ in ‘this particular unit of the Commission’ viewed ‘the negotiations in 2016 that led to the EU-Jordan Compact’. Being specific about both the policy process and their position serves two purposes. First, it filters out people who work in the relevant unit now but were not there at the time of the events you are interested in. Second, by specifying that you are interested in that particular person, it leaves little room for them to decline by saying they believe their perspective is not relevant. Subsequent interviewees were approached based on recommendations made by first-round interviewees. Interviewees were asked about which actors were involved, what their interests were, and what activities they engaged in, when forging the EU foreign policy responses to the 2015 migration crisis that I was interested in. The final interviews pointed me only to individuals I had already interviewed, signalling that I had already covered all the central actors for my cases.

The online interviews took place in the beginning of a global lockdown, which constitutes an important scope condition. Thus, the interviewees were participating from their homes rather than from their offices, which provides interesting evidence of the importance of environment even in an online setting. Furthermore, access to the elites was likely influenced by the fact that they were all locked in their homes and thus more accessible than usual. Even so, several of the lessons learned during this data-collecting effort point towards conducting elite interviews online even in a pandemic-free future. Climate concerns, limited research resources and lack of security and stability in several regions of the world mean that online methods are here to stay. Furthermore, COVID-19 has permanently changed the way we work; hence, even in a future, pandemic-free world, people will continue to work remotely and meet online. In particular, this applies to typical policy elites’ workplaces (Berger, 2024). During the pandemic, researchers and elites have gained technological competence which makes online interaction easy and accessible.

Lessons learned from conducting elite interviews online

The following will consider in some detail the three main challenges of elite interviews: access, biased information, and positionality. The expectations raised from the generalised online methods literature and from the elite interview literature will be compared and contrasted with my own reflections from conducting interviews with the Commission on the European Union’s policy response to the migration crisis.

Access to elites online

A major challenge associated with elite interviews, noted in several works, is to gain access to the elites (e.g. Dexter 1970; Glassner and Hertz 1999; Hertz and Imber 1995; Kezar 2008; Odendahl and Shaw 2002). This challenge involves securing that your request actually reaches the elite you want to interview and furthermore that you get the elite to accept. Depending on the specific elites investigated, they might be difficult to get a hold of. There are several potential bottlenecks, such as undisclosed email addresses and phone numbers, secretaries guarding their schedules, and security measures around their offices (Lancaster, 2017:95). If you do manage to reach them, they might not reply, and even if they are positive to meet for an interview, they are often busy individuals with full calendars. Moreover, if you want to discuss a sensitive issue, they might be less inclined to make space for you. Odendahl and Shaw (2002) suggest that the likelihood of successful access tends to be higher if the elite network identifies the researcher as a credible actor. While there are several measures to increase credibility, Odendahl and Shaw (2002) recommend networking the relevant group so that they know who you are before you approach them for an interview.

The online interview literature highlights how online methods might facilitate access, by focusing on physical barriers to face-to-face meetings (Mann and Stewart 2000:17–18). With online research, you can widen your sample to potentially cover different parts of the world and reach people who are not mobile because of physical disabilities (Bowker and Tuffin 2004:230; O’Connor et al. 2008). For very busy individuals, going online can increase the likelihood of them being able to participate (Deakin and Wakefield 2014). However, the literature also argues that going online can limit access significantly because potential interviewees may have limited access to the internet or to relevant software. They may also lack the required computer competence (Cater 2011; O’Connor et al. 2008).

In my experience, going online significantly enhanced access to elites. My initial plan was to spend April and May of 2020 in Brussels, and to conduct around 20 interviews with Commission representatives while I was there. Several of these interviews were confirmed. When travel bans made my trip impossible at the last minute, I asked the participants I had already scheduled interviews with if they were willing to attend online instead. They all replied positively. In addition, most of the people I had contacted but not heard back from also agreed to participate online. Thus, I obtained a considerably higher success rate in setting up online interviews than in arranging office interviews. Furthermore, because I was no longer restricted by physical proximity, I also contacted Commission representatives stationed at EU delegations in other parts of the world. It would not have been feasible for me to travel to remote destinations for single interviews; however, online I could interview them as well. All my respondents were technologically competent and experienced in conducting meetings online. All I had to do to set up the online interviews was to figure out what their preferred program was and schedule the meeting in their calendar. A contemporary technical challenge in online methods, especially when contacting elites, is to navigate their digital security systems. Across units within the same organisation, there were different security-based recommendations for which program should be used to facilitate online video calls. This can be solved specifically asking for their preferred program while making the appointment, to ensure that you have the necessary programs installed before the meeting.

In conclusion, going online can increase access to elites by facilitating the participation of busy or geographically dispersed individuals. However, the pandemic was most likely a very important factor. In April and May, when I conducted most of my interviews, the Brussels elites were all at home in their house or apartment, in lockdown. The city had shut down and at least for the specific group I was interested in, the workload was manageable, meaning they had time for me. In a post-COVID world, access to these individuals may not be as easy as during lockdown. Several of the interviews were conducted after working hours, and because the participants had little else to do due to the lockdown, they were perhaps more open for being interviewed in their spare time. Moreover, because of the lockdown, they were forced to work remotely, meaning that they had a steep cultural and technological learning curve concerning online meetings. While policy elites are always viewed as highly educated and competent individuals, during the first 18 months of the pandemic they were mainly working online, meaning that online was their ‘normal’ mode. Going online was no challenge in terms of gaining access to these individuals. Concerning access to information, an unexpected drawback of the online format was that in two separate interviews, participants referred to unofficial documents available to them and of high relevance to me. On both occasions the participants made it clear that had I been in the room with them they could have easily provided me with said document, but they were not comfortable sending me the documents online, considering that they were unofficial. Because I could never access the documents, it is hard to tell how important they would have been for my analysis. However, according to other senior EU researchers, extensive fieldwork in Brussels can result in getting access to very relevant unofficial documents that can be quite important sources of information, particularly when insider interviewees can help you interpret them.

The findings of this case study support the conclusion of the study of non-elite interviews by Deakin and Wakefield (2014). They achieved higher acceptance rates, especially for busy people, by offering online interviews. Moreover, this case study concurs with the interview literature which highlights the possibility of widening your sample to include more geographically remote individuals through online methods. However, much of the current literature focuses on access to technology and technological competence. As this case study demonstrates, such issues are no longer very relevant for research on elites, because for many elites working remotely has been the norm during the pandemic and even thereafter. On the other hand, with this digital shift comes new issues, such as security standards regarding software for the elites’ workplace. Finally, getting access to unofficial documents, which is uniquely relevant to policy elite research, might be more difficult to achieve through the online sphere where sharing leaves digital footprints.

Biased information from elites online

A second main challenge when conducting elite interviews is the risk of biased information. There are several potential sources of biased information. First, the interviewee’s memory might fail (Natow 2019). There is also a risk, especially when researching politicised issues, that the elite might misrepresent the facts through self-serving statements or elusiveness (Natow 2019). Moreover, social desirability bias might lead the interviewee to answer in a way that reflects positively on him or her (DiMaggio 2014). These are relevant issues for expert interviews more generally (Bogner et al. 2009; Döringer 2021). Most people have motivation and ability to skew information in an interview setting; however, research suggests that elites are better equipped to protect themselves and better positioned to manipulate research results than other interviewees (Ross 2001; 160; Smith 2006:644). Issues related to anonymity and confidentiality are particularly pressing when interviewing elites on politicised issues (Lancaster 2017). Because of their position, elites might even be viewed as a vulnerable group when they disclose information about sensitive policy processes (Lancaster 2017:99). Natow (2019) finds that it is very common for studies involving elite interviews to include a review of policy documents. Kezar (2008:397) argues that an analysis of other sources of data, such as a document review, often serves as a starting point from which the interview guide for elite interviews is made. This way, the researcher becomes more knowledgeable and better equipped to critically review and interpret the information received during the interview. To limit bias, the researcher must be well-prepared and stay vigilant throughout the interview. However, the online mode impacts the researcher’s ability to interpret information. Body language is highlighted as an important aspect of an in-person interview that is lost in the online setting (Cater 2011). The vast majority of the existing literature on online interviews is not terribly optimistic. Several studies suggest that the quantity of information is adversely influenced by the online mode, because you can only register what is caught by the camera and microphone, which amount to less information than you would get if present in person (Rettie 2009:422; Deakin and Wakefield 2014). However, some studies indicate that participants might be more comfortable being interviewed remotely, such as online or via telephone, rather than in person, and that they therefore might offer richer information in an online format (O’Connor et al. 2008; Stephens 2017; Weller 2017).

The interviews in this case study were well prepared, based on an interview guide prepared by reviewing relevant policy document, news articles and press releases. Some of the interviews had no video, only sound. In some cases, the reason was poor connection in the context of home office, in other cases the interviewee simply preferred no camera. An interesting finding is that not having visual contact with the elite can be freeing for the researcher. It means that the researcher does not have to think about facial expressions and reactions while the interviewee speaks. The uneven positionality between elite and researcher puts pressure on the researcher to act in a certain way, to demonstrate that they are competent and understand what the elite is talking about. Thus, the elite interview setting demands a lot from the interviewer. She must be perceived as registering and comprehending all of what the elite is saying, while at the same time critically reviewing the answers, consulting the interview guide, and searching for questions that can provide the empirical evidence she is looking for. This case study finds that the researcher has much more capacity for critically reviewing the information and coming up with good follow-up questions in the off-camera interviews. When not having to focus on anything else, the information attained during the interview can be of higher quality. Elite interviewers have to carefully sequence their questions, probe issues, and listen to what is said and not said during the interview (Hermanowicz 2002; Johnson et al. 2019). Stephens (2017:211) similarly finds that in telephone interviews with elites it was easier to read questions from the interview guide during the conversation. Even for the most experienced interviewer, being on camera while conducting an elite interview might prove a challenge.

Furthermore, the lack of access to physical cues may rightly be viewed as a drawback of online methods, causing a higher risk of misinterpreting the information provided. On the other hand, if the elites are believed to be better at manipulating information than other interviewees are, are the elites then not also more capable of misleading the researcher through physical cues? This case study argues that there is some control gained in the off-camera setting, a setting making it easier to push the interview in the desired direction.

The recorder was not as noticeable as it would have been in a physical meeting. Indeed, several studies have highlighted this as a benefit of online methods (Howlett 2021; Rettie 2009; Weller 2017). In face-to-face interviews, there is often an awkward moment when the recorder is placed on the table. In contrast, in this case study the starting of the recorder went unnoticed because the issue of recording was settled via email before the interview. Some of the elites had expressed by email that they would be able to speak more freely without recording. In those cases, rigorous note-taking replaced recording. However, there were instances during the interviews where the elite would hint that they wanted to disclose something ‘off the record’. It was not always clear if they meant this in the strict sense of ‘please don’t record this part’, or simply that they did not want to be quoted on it. Anyway, such a remark entailed a temporary stop of recording to facilitate frank conversation. Such exchanges illustrate that although the awkwardness of the physical recorder is not as pressing in online interviews, elite interviewees are very vigilant. They take into consideration that they are being recorded and they weigh what information they are willing to give under which circumstances. All of the interviewees in this case study made it clear that they were open to being contacted again with any potential follow-up questions. This is likely something they would have offered even if the interview had taken place in person. However, the benefit of only having met online is that the threshold of contacting them online at a later stage is much lower. All informants remain a simple phone call away in case new questions emerge. Having already spoken with them over the phone or online, it may be perceived as less daunting to call them up again in that same format.

Positionality in an online context

Positionality in elite interviews is a third main challenge. Expert interviews differ from elite interviews concerning how power and bias influence the quality of the data obtained (Liu 2018, van Audenhove and Donders 2019). Smith (2006:644) claims that several academics conducting elite interviews have admitted to not treating the elites according to ethical best practices. Ross (2001: 157) describes her experience from conducting elite interviews in the Australian Parliament in the following way: ‘It is an arrogant and imposing environment which puts all its visitors firmly in their place, which is some way below its own lofty confidence’. She argues that the interview environment reinforced the unbalanced power relation between the elites she interviewed and herself (2001:161). Positionality constitutes an obstacle to getting reliable information, because it affects the researcher’s ability to create a balanced relationship in which it is feasible to collect sensitive information and to judge the reliability of the information provided. Furthermore, the interview setting can reinforce the power asymmetry between the researcher and the interviewee (Smith 2006). To reach elites, you will often be expected to come to their offices, which are oftentimes located in government buildings with strict security measures for visitors. The context in which the interview takes place is very much ‘in their court’, further underlining the interviewer’s dependence on the elite in the research situation (Ross 2001). To mediate the relational issues in elite interviews, Kezar (2003:400) dives into feminist and narrative traditions. She identifies, inter alia, ‘mutual trust’, ‘mutuality’, ‘egalitarianism’, ‘empathy and ethic of care’ as relevant concepts. By facilitating these interconnected concepts, the researcher will minimise the power differential and create room for a more open conversation where the elite can provide information and at the same time give the researcher room for critical reflection concerning the elite’s accounts. However, the online format may affect how we as researchers achieve these norms.

A main critique of online interviews is that they constitute an inferior alternative to in-person interaction (e.g., Adams-Hutcheson and Longhurst 2017; Deakin and Wakefield 2014; Hanna 2012; Oates 2015; Seitz 2016; Weinmann et al. 2012; Weller 2017). Information quality may allegedly be affected by the online setting, which cannot create the required level of intimacy and mutual trust (e.g., Adams-Hutcheson and Longhurst 2017; Hanna 2012, Seitz 2016). Because the face-to-face interview is the standard way we communicate and how we have done interviews in the past, online researchers turn to established offline practices that do not necessarily translate well into the online setting (O’Connor et al. 2008). Because of the fast technological development and the turn to working remotely and online, many of these studies are now outdated. Several more recent studies support the claim that online interviews can in fact provide comparatively deeper connections and more open information sharing than in-person meetings can (e.g. Weller 2017; Jenner and Myers 2019; Howlett 2021). Some interviewees have even compared the online interview to speaking with a friend (Weller 2017:618).

Assessing the quality of the relationship with an interviewee is a complicated matter. In fact, ‘the quality of online relationships’ is a developing subject of study for psychologists (e.g., Cummings et al. 2002; Wolak et al. 2003). Even so, I believe some lessons can be drawn from this case study’s somewhat unexpected findings. The interviewees were surprisingly open and frank, even though they were discussing a highly politicised issue. The online format could potentially facilitate more open dialogue. There are several reasons why this could be the case. Studies suggest that the location of the interviewee matters also in an online setting (Howlett 2021; Jenner and Meyers 2019). The interviews in this case study were not conducted inside the Commission building with colleagues in the next room. This important contextual matter relates not only to the elites and their ability to share unbiased information with me, but also to my positionality as an outsider. Powerful and intimidating surroundings, such as official buildings, can reinforce the power imbalance characteristic of elite interviews (Ross 2001). In contrast, the home-office context in these online interviews took place, was very personal. Most of the interviewees were at home, and their bed, their bookshelf or a piece of art was visible in the background. One interviewee introduced their daughter on camera, the dog of another could be heard in the background. Such personal exchanges, more so than most in-person office interviews, builds down the uneven positionality that very much categorises elite interviews. By inviting the researcher into their personal space, a starting point of mutual trust is created. The pandemic created an ideal context to shed light on the role that physical surroundings play in online elite interviews.

Discussion: online elite interviews can provide unique insights in policy research

While studying a least likely case in terms of obtaining access to reliable data, this paper finds that online interviews provided unique and valuable insights. Indeed, the online approach mediated several of the main challenges with elite interviews. The online format allowed access to hard-to-reach and very busy individuals. This finding is partly related to the fact that the elites were in lockdown, which likely affected their schedules. Moreover, contrary to expectations from the online literature, it is possible to get high-quality data in an online setting. Building rapport with elites can be achieved also online, and it is possible to have conversations that run smoothly and have the elite speak freely, even on highly controversial issues. Finally, in keeping with the online methods literature, this case study finds that the online setting provides less visual information than a traditional face-to-face interview would offer. However, the case study raises a potential benefit with online methods: the less visual information, the easier it was to stay vigilant and focus on what was being said and to refer to the interview guide which is crucial in elite interviews. These factors served to increase the quality of the information. Furthermore, policy elites are likely as capable of being persuasive visually as verbally and I am not convinced that in-person interaction would reveal more relevant information than the online interviews did. The drawback is that the threshold for the elite to share unofficial documents is higher online.

It was clear that the participants in this case study found the topic of the interview difficult to discuss. They sometimes signalled that they found their own discourse on migration problematic by cutting themselves off or by explaining how their work affects their thinking on these issues. For example, when asked about the EU’s motivation for establishing the Emergency Trust Fund for Africa, one interviewee explained that:

Something needed to be done and especially in the countries of transit and where the influx of migrants stem from, to sort of – and you know it comes out extremely negatively whenever you say it, but the thing is, when you work with it, you do not really see it that way (Commission representative 1)

Although the participants recognised that from an outsider’s perspective, it could be problematic that the EU wanted to hinder individuals fleeing from war, they willingly shared information on several such policy processes openly, while reflecting on how their role could be questioned from a moral standpoint from outsiders. Several elites bluntly admitted that during Syrian the civil war, the EU’s foreign policy efforts in countries bordering Syria were directly linked to the refugee crisis in Europe at the time. For example, Commission representative 6 said ‘we did not want refugees to leave their countries of temporary residence bordering Syria’.

The interviewees also disclosed major epistemological differences as well as diverging interests between different units within the Commission. They were quick to call out colleagues in other parts of the Commission and sometimes paraphrased what was said by them in internal meetings. Furthermore, the interviewees disclosed information about unofficial diplomatic activities that Member States engaged in to influence the Commission. They also disclosed statements made in unofficial diplomatic settings with countries outside Europe. One elite paraphrased a non-EU state:

“look, we could also somehow, you know, try to send these refugees to Europe and then maybe you will give us this 3 Billion as well, you know, as you gave to Turkey”. (Commission representative 20)

The interviewees provided insights into the decision-making processes behind policies and this information is necessary to understand why the policies turned out the way they did. An example is the 2016 EU-Jordan Compact, which established that some of the EU restrictions on Jordanian exports to Europe, called ‘rules of origin’, would be relaxed. This relaxation would only apply to Jordanian companies whose employees include a minimum share of Syrian refugees. The political decision adopted by the EU and Jordan in 2016 states that this agreement is meant ‘to enhance Jordan’s exports to the Union and create additional employment opportunities, especially for Syrian refugees as well as Jordanians’ (EU-Jordan Association Committee 2016, p.1). While this agreement was marketed as securing refugees’ access to work, in reality this was merely a positive side-effect of a completely unrelated goal, namely protecting EU commercial interests. The linkage to refugees was a strategy to avoid requests for a relaxation of the rules of origin from countries such as Tunisia and Egypt, which have much larger economies and pose more of a threat to EU commercial interests than Jordan does, while not hosting nearly as many refugees.

It is difficult to imagine any other research method being able to produce this kind of insights into closed-off decision-making processes. Some policy research studies lean primarily on document analysis of policy documents, but as the EU-Jordan Compact example demonstrates, the rationales underlying policies can be very different from what official documents state. To conclude on motives and interests based solely on an analysis of the output can be misleading. In contrast, through online interviews researchers can get access to influential decision-makers and to their unique insights into what really took place when they formulated policies.

Summary and recommendations

The lesson that can be drawn from this paper is that online policy elite interviews constitute a justifiable method in policy research which can produce high-quality data even in a least likely case for accessing relevant information. Online elite interviews should not be considered a mere subpar alternative to in-person interviews. Online interviews increase access to elites because it offers flexibility regarding time and space. Access might in this case study also have been facilitated by the lockdown the pandemic caused, but accounts from before the pandemic support this finding (Bowker and Tuffin 2004:230; O’Connor et al. 2008). Biased information is a challenge in elite interviews in particular. Online interviews reduce visual cues that can be telling, however, they accommodate an approach more focused on what is being said, on following up on the interview guide, and on the research agenda. Finally, regarding positionality, the study concludes that an online setting removes some of the conditions that underline the perceived distance between the researcher and the elite, especially in the context of home office. Importantly, the experiences this study builds on were made during lockdown early in the pandemic. This has provided interesting insights into how the physical surroundings impact the relationship between researcher and elite also in the online setting. COVID-19 has changed the way we work and, in the future, online meetings and working from home will likely continue (World Economic Forum 2021). In the top 10 US cities, office occupancy is only around half of the pre-pandemic levels (Kastle 2023 cited in Fulford 2023). Policy elites as a group are particularly likely to continue working in an online space (Berger 2024). Even though the findings in this paper relates to an exceptional context of a global pandemic, the specific findings related to online interaction and working from home remain very much relevant. Access to in-person visits to elites’ homes to conduct interviews is very unusual in policy research, however, conducting interviews online while the elites work from home is a more likely scenario.

There are several reasons why policy researchers should contemplate using online elite interviews in their research. The online format can increase access to relevant elites. Moreover, like in-person interviews, it can provide unique insights into highly controversial and politicised decision-making processes, without which wrong inferences might be drawn concerning how policies come about. Furthermore, researchers can reduce their carbon footprint by choosing online methods. The technological learning curve has been steep and the competence that researchers and elites have acquired should be utilised in online research.

This paper has focused on policy elites, however, the issues of access and of biased information are very much relevant for interviews with the more general category of experts. Because elites are an extreme case, in terms of gaining access and of presenting biased information, the findings of this paper are generalisable to expert interviews broadly, and it provides a recommendation to consider conducting such interviews online.