Science diplomacy (SD), a burgeoning field at the intersection of international relations and science policy, has received limited scholarly attention in Central, Eastern, and South-Eastern Europe. This study addresses this gap by examining international scientific collaborations among Central European Initiative (CEI) member states, a diverse group encompassing EU members, candidate countries, and third countries. The inherent paradox of SD—the convergence of the national interests of diplomacy and the borderless nature of science—is highlighted. This research utilizes a network perspective, acknowledging that the frequency and direction of collaborations are influenced by actors' positions within the network. This study innovates by applying social network analysis (SNA) to explore SD collaborations, unlike existing literature primarily focused on national SD apparatuses. The study investigates two key research questions: 1) Do larger economies hold more prominent positions in SD networks, or can smaller countries achieve disproportionate influence? 2) Do shared EU memberships lead to distinct clusters within the broader CEI network, potentially excluding non-EU members?

Existing research on SD often focuses on the institutionalization of national SD structures and instruments, such as science attachés and S&T agreements, primarily in leading global powers (e.g., US, China, France, Germany). Studies analyzing specific cases of scientific collaborations, like those concerning the Arctic or climate change, are also prevalent. However, research on SD in Central, Eastern, and South-Eastern Europe is limited, with notable exceptions such as recent work in EU-funded projects and studies on specific countries' SD activities. This article bridges this gap by employing a methodological innovation: the use of SNA to analyze the relational aspects of SD, complementing the existing focus on national-level attributes.

The research employed an online survey administered in two waves to officials in the ministries of foreign affairs (MFAs) of CEI member states. The survey explored: (1) the organization of SD activities at the national level; (2) diplomatic tools for international S&T cooperation (e.g., science attachés, cooperation agreements, joint initiatives); and (3) the development of SD capabilities. Fifteen out of seventeen CEI member states participated. Data from the survey were used to construct two networks: (1) current cooperative relations between MFAs in S&T and (2) priority partners for future collaborations. Social Network Analysis (SNA) software Pajek was utilized to analyze these networks. Two centrality measures were employed to assess the prominence of countries: indegree centrality (direct links) and proximity prestige (direct and indirect links). A community detection algorithm (Louvain Method) was used to identify subgroups within the networks to address the question of whether EU membership creates distinct clusters.

The analysis revealed a complex relationship between economic size and prominence in SD networks. While larger economies like Italy and Poland held influential positions, several smaller economies (Bulgaria, Montenegro, Serbia, Slovenia) also exhibited high prominence in terms of both indegree and proximity prestige, suggesting that smaller countries can ‘punch above their weight’ in SD. The results showed a lack of a clear correlation between GDP and network centrality. Concerning the influence of EU membership, the community detection algorithm identified two subgroups of countries. These subgroups were not predominantly comprised of EU or non-EU members, suggesting that geographical proximity plays a larger role in shaping SD collaboration patterns than shared EU membership. The South/Southeastern European countries formed one group, and Central/Eastern European countries formed another, with Bulgaria being the exception, moving from the Central/Eastern to the South/Southeastern group when future collaborations were considered. These patterns persisted in both current and projected collaboration networks.

The findings challenge the assumption that economic size is the sole determinant of prominence in SD networks. Smaller countries can leverage their strategic position and relationships to achieve significant influence. The absence of clear EU-membership based subgroups indicates that regional dynamics and geographical proximity are stronger drivers of collaboration than shared institutional membership within the CEI context. The study highlights the limitations of a solely state-centric approach to understanding SD networks. Future research could benefit from examining the role of non-state actors (e.g., universities, research institutions) and the influence of specific policy areas within SD.

This study demonstrates the value of SNA for analyzing SD networks, offering a more nuanced understanding of collaboration patterns than traditional approaches focusing on national attributes alone. The findings suggest that factors beyond economic size, such as geographical proximity and established regional cooperation frameworks, significantly influence science diplomacy collaborations. Further research incorporating a broader range of actors and temporal analysis would strengthen these findings and contribute to a deeper understanding of SD in Central, Eastern, and South-Eastern Europe.

The study's findings are preliminary due to the relatively small number of nodes (15 out of 17 CEI member states) in the network. Data relied solely on a survey of MFA officials, limiting the perspective and potentially overlooking other relevant actors or collaborations. The analysis is cross-sectional, lacking longitudinal data to track changes in collaboration patterns over time and to evaluate the impact of EU membership over a longer period. The use of a binary network representation also simplifies the complex reality of SD collaborations.