Research Seminar: Wednesday 31st March 12-1pm GMT

MS Teams Click here to join the meeting

Dr Tom Mills (Aston, Sociology and Policy)

Dr Narzanin Massoumi (University of Exeter)

Examining state society interactions from lobbying disclosure data: a computational social science approach

Extensive research in political and social science has examined the activities of interest groups and their relationship with policy makers, from studies of think tanks, to the activities of professional lobbyists and business groups, to civil society campaigners and social movement actors. This work provides empirical insights into the strategies and influence of organised interests and advocacy groups, but the focus on particular sectors, industries and issues gives only a partial picture of state-society interactions. The introduction of lobbying transparency legislation in the last decade has made a broader analysis possible. In the UK, for example, the Government has since 2010 published details of all ministerial meetings with external groups. Yet there is to date little research making use of this disclosure data. This is likely due to its scale and the fact that data are spread across different sites with discrepancies in format and structure, all of which pose significant challenges for systematic analysis. Techniques from computational social science, however, have made it possible to collect, clean and restructure large volumes of data automatically, and data scientists working with Transparency International have compiled details of over 68,000 UK ministerial meetings into a structured format. This text data though is still not clean or consistent at the level of single lobbying individuals and organisations. In this paper, we resolve this with a series of automated text processing techniques. Having cleaned the data, we then analyse the relative access afforded different organisations and policy communities. Rather than manually coding organisations, which has been the approach taken in previous work, we make use of a community detection algorithm to identify interest groups from within the data itself. This allows us to systematically map state-society interactions in the UK at the level of individuals and organisations, as well as the wider policy networks of which they are part.