Dan Thompson is a PhD candidate in American politics and political methodology at Stanford University. His research studies electoral accountability in the US with a particular focus on the role of elections in local policymaking. He collects new data on local elections which he combines with large, untapped administrative datasets on government behavior. He then uses modern empirical techniques for causal inference to study how elections influence the policies local governments choose. The first paper in his dissertation, conditionally accepted by the American Political Science Review, casts doubt on the claim that national partisanship erodes local electoral accountability, studying the case of sheriffs and immigration enforcement. The remaining papers in his dissertation explore this finding further, using different cases to study the degree to which democratic elections improve police accountability.