Join NULab and David Sparks, of the Boston Celtics, as he presents a talk on a social network approach to explaining ideological polarization.
The lack of quantitative data on House candidates makes many theories of electoral politics difficult to test. Perhaps the most fundamental shortcoming of current data is the lack of ideological estimates for both primary and general election challengers. I propose a unique solution to this problem that exploits data from the rapidly expanding realm of social media by estimating ideal points for each House candidate based on the patterns of connections between their Twitter profile and their Followers.
I identify a latent ideological dimension from the matrix of following relations, which corresponds closely to roll-call based estimates for Congressional office holders. I illustrate the utility of this measure by application to House party primary elections, and find that ideological extremity is positively associated with likelihood of winning the nomination. I then argue that this phenomenon is a likely source for the polarization we observe in Congressional politics today.
David Sparks is the Director of Basketball Analytics for the Boston Celtics. He received his PhD in Political Science in 2012 from Duke University, where his research focused on American Politics and Methodology; specifically, political parties, ideology, and polarization.