My dissertation expands the Extended Party Network theory of political parties, which conceives of parties as networks of related groups. Existing scholarship has only tested theories that explain the effects of this network on individual candidates contingent on their place within the network. My dissertation takes a broader view by testing how the structure of the network of organizations outside of the legislature affects how parties act within the legislature.

I argue several points in my dissertation. First, that more influential groups within the party network have more control of party ideology and that network characteristics more generally affect the ability of a party to cooperate. Second, I argue that as groups within the extended party network become more connected that there will be more ideological agreement between members of the legislative party. Finally, I explore how how state racial and cultural diversity, size, and electoral laws effect how well groups are able to connect with each other. I use social network analysis methods and state level donation data from 2000 to 2014 to argue this.

Political Opinion and Behavior

I am working on several additional projects that focus on the role of opinion and behavior in American politics. Working with Lee Ann Banaszak (Penn State), Daniel Gillion (UPenn) and John McCarthy (Penn State), I surveyed protesters at the 2016 Republican and Democratic national conventions. This survey was designed to test theories of political mobilization. Lee Ann Banaszak and I have two working papers from this project:

In another paper, with Nicholas Dietrich (Penn State), I have tested theories about the role of media on public opinion. We used daily data on media coverage, support, and interest for candidates in the 2012 and 2016 Republican primary to test these theories.

Methodology and Measurement

Finally, I am working on a project with Christopher Fariss (U Mich) and Michael Kenwick (UPenn) to improve existing Item Response Theory (IRT) and factor analysis methods incorporation of temporal information. Together, we have developed a robust dynamic latent measurement model that allows a latent trait to move slowly in most cases with periodic sudden changes. This is akin to punctuated equilibrium which has been shown in many social science phenomena, which tend to be robust over long stretches of time followed by small stretches of rapid movement.