This data collection consists of 16 interview transcripts from semi-structured interviews with the staff of two international organizations - the United Nations (specifically the Department of Peacekeeping Operations) and NATO. Interviews were aimed at gaining an understanding of how the staff of international organizations create and maintain legitimacy internally for themselves, rather than for external audiences, like the public, member states, or beneficiaries. Interviewees consist of staff (or former staff) familiar with the substantive mission and policies of the organizations and represent a range of ranks and functions. These files have been redacted to remove identifying or sensitive/classified information. For this project, a total of 87 interviews were conducted across three international organizations – the United Nations, NATO, and the World Bank – but 71 refused permission for deposit of interview transcripts, even if redacted, including the entire cohort of interviewees from the World Bank. The interviews archived here took place during five separate rounds of fieldwork in Geneva, New York, and Brussels between January 2017 and February 2018. However, due to the risk of identification, files do not specify the date or location of the interview. These interview data were combined with analysis of primary and secondary documents and written materials and observation of spaces and objects in the project's analysis. Recent crises in Iraq, Syria, and Ukraine, as well as the current refugee crisis in Europe have highlighted the difficulties national governments have in working collaboratively with international organizations (IOs). Governments and other policymakers often view IOs as inefficient and overly bureaucratic, and therefore engage with them only selectively and superficially (Malloch-Brown 2015, Gowan 2008). However, these interactions are partly unproductive because of a misunderstanding of the conflicting obligations IOs face, how they take decisions, and how they gain and maintain legitimacy in the international system (Barnett and Finnemore 2004, Zaum 2013, Allison & Zelikow 1999, Hanrieder 2010, Rittberger, Zangl, & Kruck 2012). Indeed, across the social sciences, legitimacy is increasingly conceived of as sought-after and instrumental, as not only a static attribute but also a resource that actors seek and draw upon for strategic reasons. Within International Relations specifically, interest in how actors generate, maintain, and use legitimacy has grown in tandem with the rise of authority that exists beyond or in opposition to the state, among non-state actors, transnational movements, and IOs. However, to date most analyses of legitimacy and legitimation in International Relations focus on perceptions of the legitimacy of one actor by another (Coleman 2007, Hurd 1999). In so doing, they fail to account for self-legitimation, where an actor undertakes practices of legitimation not for external audiences, but internally, as a way of developing and reinforcing its identity (Barker 2001). Moreover, many understandings of legitimacy neglect the fact that identities are rarely uniform and instead are multiple and conflicting, and thus that legitimacy cannot be treated as a binary characteristic, but must instead be conceived of as variable, inconsistent, and contested. This project addresses these two omissions by examining self-legitimation by IOs, and looking in particular at how self-legitimation affects IO behaviour for organizations with multiple institutional identities that may conflict. Indeed, many IOs are at once operational actors that participate actively in combat, peacekeeping, development, and global political and economic processes and normative actors that develop and promote international norms. These different identities sometimes dictate contradictory goals and practices, forcing the institutions to violate the principles and activities considered appropriate to one of their identities, thus complicating legitimation. Accordingly, IO self-legitimation often entails the use of discourse and symbolic acts to convey or 'package' actions a certain way in a bid to affirm a coherent institutional identity. However, this implies that IOs often 'say one thing and do another,' and these contradictory discursive and operational practices have negative implications for perceptions of legitimacy, both internal and external. The project will undertake a comparative analysis of the self-legitimation practices of three IOs-the UN, the World Bank, and NATO. It will make a two-fold contribution, one theoretical and one practical. First, it will add to existing theories of IO behaviour and of legitimacy in International Relations, which focus primarily on external assessments of IO legitimacy and neglect the multiple and at times contradictory identities of IOs. Second, it will lead to a greater understanding among policymakers of how and why such bodies take certain decisions, how they interact with other actors in the international system, and how they can be effective actors in international affairs. While it is unlikely that the contradictions they face can be eliminated, a greater understanding of why these contradictory behaviours exist and how IOs seek to overcome them will enable more constructive and efficient cooperation, avoiding the inconsistency that has plagued recent interaction with IOs.

These data were collected through semi-structured, in-depth, qualitative interviews with the staff of two international organizations (out of three examined in the project): the United Nations and NATO. Interviewees were asked broadly similar questions, but were encouraged to speak widely and freely, and to include examples of their own choosing. These interviews were undertaken during five rounds of field work between January 2017 and February 2018. Interviews took place in New York, Geneva, Brussels, and by phone from the United Kingdom (additional interviews for the project were conducted in three additional rounds of field work in Washington, DC and Germany, but were not authorized for deposit at the interviewees' request). Where interviewees consented, interviews were recorded and transcribed and subsequently analyzed as part of the study. It should be noted that only 16 interviewees consented both for recording and deposit of transcripts, out of a total of 87 interviews conducted.