What if the Case is ‘Elusive’? Methodological Challenges for Qualitative Research in International Relations and Global Governance

2.1. Elite Interviews

In total, I conducted 26 elite interviews with people involved or familiar with the activities of the selected MSPs. Interviewees included individuals working in a wide range of capacities, such as IO staff, MSP staff, NGO representatives, scientists, and communication officers.

Due to the relatively small number of individuals meeting the inclusion criteria (i.e. participation in or familiarity with the activities of one of the studied MSPs), the recruitment of interview participants was mainly guided by purposive sampling. At the same time, this approach was complemented by other non-probability sampling strategies: while snowball sampling was instrumental in the recruitment of additional participants, the logic of quota sampling was helpful to avoid an overrepresentation of certain categories of respondents.

Potential interviewees were identified from a mix of sources: information available on the webpages of relevant organizations, lists of participants from MSPs’ events, potential informants’ direct participation in specific online events. In most cases, the contact details of prospective interview participants were retrieved from the same sources.

More than a hundred individuals were invited for an interview. The large majority of potential interviewees were contacted via email. When email addresses were not available, LinkedIn was used as an alternative channel to approach potential interview respondents. In a few other instances, I reached out to relevant organizations via the contact forms available on their websites. The overall response rate did not exceed 30%. In about 10 instances, interview invitations were either declined or redirected to colleagues within the same organization. Factors such as partnership composition or policy field did not appear to influence the likelihood of a response. Nevertheless, I noticed that – compared to other categories of individuals – a somewhat higher response rate was received from academics and scientific experts. This could be due to the possibility that experts may perceive an academic interviewer not only as a ‘colleague’ but also as a credible or competent interlocutor with whom to exchange specialist knowledge, notwithstanding disciplinary or epistemological barriers (Bogner et al., 2009). In sum, access proved to be challenging for the in-depth study of MSPs through elite interviews. I was an ‘outsider’ in all case studies and could rely on my network contacts only on a single occasion. Moreover, it should not be ruled out that challenges in access may have been compounded by logistical and contingent obstacles, including spam filters and ‘bad timing’.

To maximize flexibility while ensuring comparability, interviews were conducted following a semi-structured format. This allowed, for example, to adapt interviews to the time at disposal of each interviewee or to direct certain questions to specific categories of respondents. In line with recent evidence from the literature, the online format turned out be advantageous in terms of leveling power relations (Howlett, 2022; Vaagland, 2024). The possibility of meeting in a neutral cyberspace contributed to making conversations more informal and may have reduced the risks of bias and unforthcomingness stemming from meeting in a professional setting. This was further facilitated by the fact that several informants took part in the interview from home instead of their office (Vaagland, 2024).

Concerning other aspects of rapport, my experience with online elite interviews was less favorable. Short time slots in busy schedules also meant less to no time for small talk and off-the-record conversations. This often resulted in more focused but less spontaneous exchanges. The well-known sudden severance occurring at the end of Zoom meetings contributed to amplify this sense of ‘distance’. While all of these circumstances did not completely eliminate snowballing opportunities, they substantially limited their scope. These insights are in contrast with previous research suggesting that the online format does not significantly impact rapport and may even be preferable to in-person interviews (Jenner & Myers, 2018; Weller, 2017). Clearly, these considerations need to be gauged against the particular dynamics that occur when ‘researching up’, as elite interviews may not engender rapport in the same way as interviews with non-elite participants. Moreover, the online format is not immune from the compounding factor of time constraints, as elites notoriously display a tendency to be (or to appear to be) very busy or not particularly flexible with time (Harvey, 2011; Liu, 2018). When the clock is ticking and time is running out, one usually has to make a pragmatic choice between fostering rapport (for instance, through an adept combination of probes and prompts) or diligently (yet aloofly) exhausting the list of questions in one’s interview guide. There are no easy ways to escape this trade-off between depth and breadth: while good rapport is often crucial to obtaining richer and more insightful answers, the semi-structured interview format is significantly compromised when several questions remain unasked. Hence, in situations of severe time constraint, I usually prioritized covering as many topics as possible instead of engaging in rapport-building techniques.

Importantly, considerations related to time constraints and rapport-building should not be limited to the duration of the interview but rather extended beyond it, as they can encompass a much broader scope. In this regard, I could mention a personal invitation I received to participate in an event organized by an MSP, whose leader I interviewed some months earlier. Unfortunately, I was not able to attend, since the event was scheduled significantly after the conclusion of the Transformative Partnerships 2030 project. Clearly, participating in such an event would have represented a unique opportunity to conduct fieldwork on that particular MSP, enabling a significantly richer engagement with the subjects of my study. As this episode demonstrates, methodological challenges can hardly be separated from critical reflections on the short-term and project-oriented nature characterizing much social science research today: the need to meet deadlines and adhere to predefined time-horizons can pose serious hurdles to the conduct of in-depth case studies, especially if the case is ‘elusive’ or ‘distant’.

Another potentially challenging aspect of interviewing ‘distant’ elites and experts pertains to the specificity of the topics dealt with in interviews, as global MSPs operate in a wide range of policy fields characterized by specific stakes and implications. As a result, conducting interview-based research on global MSPs requires not only a basic understanding of each policy field’s ‘problem structure’, but also some familiarity with jargon and technical details. To address this challenge and to maintain credibility in the eyes of my interlocutors, I made sure to arrive well-prepared at interviews. Nevertheless, it would be naïve to believe that novice-expert knowledge gaps could be easily closed.

Overall, my experience in the Transformative Partnerships 2030 project suggests that specific dynamics and variables apply to online elite interviews. A key lesson learned is that the benefits of the online format come with trade-offs. Notably, the cost-effectiveness and efficiency of online elite interviews can come at the expense of limited rapport and reduced informal talk. In other words, whereas online interviews with elites may increase accessibility (Vaagland, 2024), they do not necessarily increase access.

2.2. Event Observation

Event observation was carried out by observing virtual events organized by the MSPs in focus in the project. The formats and channels of these events varied substantially, ranging from large conferences livestreamed on YouTube to targeted technical meetings held on Zoom or Microsoft Teams. Since most of the events were simply designed to convey information, opportunities for online interaction among event participants were limited or absent. In some events, participants were allowed to use the chat function to direct questions to the organizers. Nevertheless, Q&A sessions were usually very short, leaving no time for meaningful engagement and discussion. All of these elements contributed to orient my participation in these events toward a complete observer role (Gold, 1958). In many ways, this made me a ‘lurker’. Mine was not necessarily a deliberate choice. In this regard, it has been argued that viewing naturalistic lurking and active involvement as alternatives represents a “false choice” for online ethnographers (De Seta, 2020, p. 88). To a certain extent, my experience corroborates this perspective insofar as my position as a lurker was not dictated by a choice stricto sensu. It resulted instead from what I would describe as the ‘rules of the field’ established by the organizers. Ultimately, my unobtrusive role enabled me to observe these events like a “fly in the wall” (Reunamäki, 2025, p. 190), with minimal to no interference with the field.

More specifically, online event observation had three main benefits: a) it allowed me to identify the main actors and stakeholders within each MSP (as many were subsequently invited for online interviews); b) it provided rare and unique opportunities to see these actors ‘in action’; c) by enabling the identification of patterns and repetitions, it provided a sense of the ‘hot topics’ in each MSP’s agenda.

Nevertheless, if compared to traditional in-person observation, this approach had a number of drawbacks. Not unlike online interviews, online observation of these ‘elite events’ conveyed a strong sense of physical separation from the study subjects and objects. The “comfortably detached” feeling (Gold, 1958, p. 222) typical of the complete observer rapidly turned into ‘discomfortable detachment’ because, as outlined above, such a status was not deliberately chosen by the researcher but rather imposed by the circumstances. Being that ‘far’ and physically detached from the field also meant not having any chance to rely on gatekeepers. Yet, the main limitation was the impossibility to immerse oneself in the events to garner deeper meaning from them. In this respect, whether remote fieldwork can produce descriptions as ‘thick’ as those obtained through on-the-ground fieldwork is a subject of scholarly debate, with some arguing that the divide between the two approaches may be overstated (Postill, 2017). However, relying just on fixed-camera streams and superficial glimpses, I was only able to obtain ‘thin’ – rather than ‘thick’ – descriptions. This contributed to making the interpretation of actions and happenings more challenging.

As I could not capture the events’ atmosphere, I could neither, for example, appreciate the possible reactions, emotions and tensions occurring not only during but also before and after the events. In this regard, global governance researchers conducting collaborative event ethnography (CEE) at mega-events, such as large United Nations conferences, have particularly stressed the importance of ‘being there’ (O’Neill & Haas, 2019). Without ‘being there’, one misses not only the performative and symbolic dimensions characterizing these transnational events but also the “situated globality” emerging from them (Aykut et al., 2024). However, as my experience shows, ‘being there’ is not always possible in global governance research. This impossibility has very little to do with accessibility (this was not an issue, although participating in some events was subject to accreditation and some degree of selectivity), but rather with the fact that a physical ‘field’ did not exist at all in many of these meetings. Recognizing this aspect encourages a deeper reflection on the epistemological dimensions of digital fieldwork. With regard to the latter, ethnographers in the anthropological tradition have highlighted how the notion of a bounded field site is increasingly challenged in favour of constructed and fluid conceptualizations (Burrell, 2009; Seligmann & Estes, 2019). In this socially constructed demarcation of the field site, the researcher obviously plays a crucial role. However, my experience in the Transformative Partnerships 2030 project suggests that the field site is not necessarily constructed by the researcher, as its boundaries may actually be imposed by the case study itself. This leads to an important consideration regarding the difference between the ‘case’ and the ‘field’, as their meaning and mutual relationship may be viewed differently both across and within social science traditions. In political science and IR, the ‘field’ is just one gateway – and only one of the possible ones – to study the ‘case’. Moreover, for mainstream political scientists, the case is unlikely to be constructed throughout the fieldwork process. Instead, it is more likely to be understood as something ‘out there’, waiting to be examined and then compared with other cases within the same population. All of this places constraints on the freedom to construct not only one’s field site but also one’s case.

Last but not least, there are important ethical considerations to be made, in particular with regard to my nearly ‘lurker’ status in the field. Although ‘lurking’ is a widespread practice among online ethnographers and other social science researchers, the ethical boundaries of this practice are still object of debate (Roller & Lavrakas, 2015, p. 193; Aykut et al., 2024, p. 33). Even though my research did not involve deception, all types of covert observation may raise ethical questions. Whereas not invisible, I was nearly indistinguishable from other event participants. As a result, my researcher status could easily go unnoticed. While it would have been unrealistic to obtain informed consent from all event attendees, a careful, considerate and attentive use of the gathered data was imperative to avoid misrepresenting the role and views of the active participants. Although nothing particularly sensitive or confidential was discussed during these meetings, the atmosphere could vary depending on the type of event and did not always convey the impression of an ‘open-door’ experience. In this respect, it may be debated to which extent awareness of the participation of a social scientist may have been perceived as troubling or problematic by the other participants. To adequately reflect on this aspect, it is important to note that the composition of participants in these meetings was not uniform. Besides the more active actors – usually organizers and invited speakers – several other passive ‘stakeholders’ populated the virtual meeting rooms. In such a scenario, I reckon that ethical concerns carry more weight for the former type of participants than for the latter. In fact, it seems of little consequence to assess the impact of my online presence on individuals who may simply have been other ‘lurkers’ in the room and who may themselves have been participating in these events for the sole purpose of gathering information. From my own viewpoint, it was similarly difficult to determine the motivation behind other participants’ involvement, as my interpretations could only be based on attendees’ observable behavior during the meetings.

In sum, observation of online events can be a valuable resource for studying subjects and objects that are ‘distant’ or unreachable otherwise. In the framework of the Transformative Partnerships 2030 project, it turned out to be a focused and cost-effective strategy for investigating MSPs and identifying their key stakeholders. However, as a low-hanging fruit, it was not devoid of limitations. Besides some lingering ethical dilemmas, limitations include obstacles in overcoming the complete observer status, an overall detached experience, and a shallow appreciation of social and contextual dynamics. While researchers in the positivist tradition may claim that some distance and separation from the study objects are not so undesirable, interpretivists accustomed to in-depth case study research are likely to see the glass half empty in these circumstances. For global governance research, an important takeaway is that a physical ‘field’ may be unavailable or inexistent for the study of particular actors and institutions, including MSPs. This, I argue, may amplify this perception of ‘distance’. It adds another layer of complexity to the previously recognized constraints that may compel ethnographers across various social science disciplines to opt for remote fieldwork, such as logistical challenges, illness, disability, or the sheer unsafety of being physically ‘on site’ (Postill, 2017). Even though many global governance scholars have embraced the perspective of conducting multi-sited ethnography (Marcus, 1995), it is clear that the challenges deriving from this ambition are not merely influenced by financial and cost-benefit considerations but also involve significant methodological and epistemological dimensions.