Mapping Social Networks at UACES 2018
Academic conferences are key meeting points in the research community, a vital outlet for new research, and an indispensable opportunity for collaborators to interact in person. But, asks Brendan Moore, how do we conceptualise these complex events, and what can they tell us about the wider European Studies discipline?
UACES 2018 in Bath. View more photos on Flickr.
During the UACES 48th Annual Conference in Bath (UACES 2018), I used the publicly-available data on the conference website to begin exploring these questions using social network analysis, which aims to ‘understand a community by mapping the relationships that connect them as a network’.
Social networks are divided into what is being connected (the ‘nodes’) and the connections themselves (the ‘edges’). Both are integral to any network. Without edges, nodes are just an unrelated group; without nodes, edges do not exist.
Here I analyse UACES 2018 from two angles using social network analysis: the first focuses on panels (connected by the people who participated in them), and the second focuses on people (connected by their shared panels).
Panels as Social Networks
The public research programme for UACES 2018 lists 127 panels and 522 participants (defined as the people who contributed to conference panels, including both those that attended the panels in person and their co-authors).
My first network analysis focused on the panels themselves, which covered the gamut of topics in European studies, including Brexit, foreign policy, gender studies, the environment, and space policy.
The visualization below shows every panel. Two panels are connected when they share at least one participant (the more participants they share, the thicker the line connecting them).
Figure 1: UACES 2018 panels connected when they share participants
How often did these diverse panels share participants?
The short answer is: not often. This can be shown with a statistic known as graph density, which ‘measures how close a network is to complete’.
The maximum possible value for graph density is 1.0 (meaning 100% of all possible connections are present). The graph density of the UACES 2018 panel network is 0.026 (or 2.6% of the theoretical maximum), rising to 0.03 if the nine panels that had no connections are removed.
However, several groups of panels with shared topics were much more connected than average (Figure 2).
Figure 2: Graph density of selected groups of panels, in comparison to the overall average for the conference
Six are highlighted in the map above: foreign policy (with a graph density of 0.18), energy and environment (0.29), trade (0.43), Euroscepticism (0.67), gender (0.73), and differentiated integration (0.83).
These panel groups make up densely connected ‘islands’ in the more sparsely connected ‘sea’ of UACES 2018.
People: Who Bridges the Gaps?
The second angle I took to analysing the conference was to look at the participants and how they were connected through shared panels. In other words, I ‘flipped’ the analysis, making people into the nodes and panels into the connections.
The result is shown in Figure 3. Larger, greener nodes have a higher ‘betweenness centrality’, a measure of how important a person is to the connectivity of the network.
Figure 3: UACES 2018 participants connected when they participate in the same panel. Larger/greener nodes have higher scores on ‘betweenness centrality’
One thing to notice is that the number of participants connecting panels together is small: only 30% – 150 people – were part of two or more panels.
This makes these ‘connectors’ even more important to the network. If the top 5% of participants by betweenness centrality are removed (only 26 people), the main, connected part of the map in Figure 3 breaks into ten separate networks. If the top 10% are removed (52 people) it breaks into 19 networks.
UACES in Time: From Snapshots to Moving Picture?
There are many additional ways in which UACES 2018 can be analysed using social network analysis.
One approach is to analyse the network based on participant categories. For example, I recently wrote a brief blog post about the gender dimensions of network centrality, prompted by questions from Dr Toni Haastrup and Dr Katharine Wright, co-convenors of the UACES Research Network ‘Gendering EU Studies’. I then used the same data to analyse the ratio between men and women in each UACES 2018 panel.
Another possibility is to trace changes over time. My analysis was only of the 48th Annual Conference, how would these findings look if the analysis was extended back five years, ten years, or more? How would attention to issues change? An obvious current example is Brexit, but Prof Paul James Cardwell pointed out on Twitter that interest in many topics waxes and wanes from year to year.
These and other topics would be interesting to explore in the coming years, especially as UACES marks its 50th anniversary and prepares for the next half-century of European Studies.
About the Author
Brendan Moore is a Senior Research Associate focusing on EU climate and energy policy at the Tyndall Centre for Climate Change Research, University of East Anglia. He is also the coordinator for the Brexit & Environment research network. Further analysis can be found on his blog and by following him on Twitter.
