I am an Assistant Research Professor in the Duke Social Science Research Institute (SSRI). My research is focused on understanding two aspects of political accountability: the determinants of citizen capacity to hold politicians accountable, and elite strategies for subverting mechanisms of accountability. This fall I will be teaching Practical Data Science, the course site for which can be found here, and a preliminary syllabus for which can be found here, and this spring I will be teaching Unifying Data Science (listed in some places as Practicing Data Science II), the preliminary site and syllabus for which you can find here.
In my efforts to understand why some communities are better able than others to hold politicians accountable, I use novel sources of big data to measure social networks and test previously untested theories of how the structure of social networks allow citizens to sanction free-riders and share information for political purposes. In Friends Don’t Let Friends Free-Ride, I use eight months of cell-phone meta-data — which includes information on the communication patterns of more than 20 million Venezuelans — to measure the structure of social networks for an entire county, then relate individual-level measures of network position to political participation. Similarly, an effort to explain why ethnically fragmented communities experience worse development outcomes, my work in Zambia shows that ethnically fragmented communities have more fragmented social networks, potentially impeding their ability to organize to hold politicians accountable. In work in Uganda, we have shown that social networks can help explain variation in electoral turnout, a critical prerequisite to electoral accountability. And finally, in my work on Somaliland, for example, I have documented how dependency on taxation created a critical mechanism of accountability.
In my efforts to better understand how politicians seek to undermine mechanisms of accountability, I study the ways in which politicians manipulate aspects of election administration to influence the composition of the electorate, undermining accountability. In particular, my work has improved our ability to understand and measure gerrymandering, improved our understanding of how aspects of election administration, like polling place placement, impact voter behavior, and measured the extent of partisan manipulation of polling place locations in North Carolina.
Finally, I am also passionate about ensuring the integrity of social science research, and making computational tools accessible to social scientists to empower them to advance our understanding of the world. To that end, I have compiled a number of tutorials which are available on this site, developed a social scientists’ guide to the world of Python and given talks on about the potential value of new tools like the Julia Language for academic research.