(iPoster) Uncontested Elections down the Ballot: Problems and Solutions

• In Event: Poster Session: Public Policy & Urban and Local Politics
  In Poster Session: Poster Session: Urban and Local Politics

Abstract

Political scientists studying elections in small jurisdictions often rely on administrative data from official ballot returns, such as the California Elections Data Archive. These data have an important limitation: they only include the results from races that reached the ballot. For the types of down-ballot races considered by local politics scholars, this can lead to a substantial censoring problem: back-of-the-envelope calculations suggest that for school boards in California, around 40% of district-election-years (including off-cycle districts) feature no race on the ballot. Studies considering only the impact of some reform on the identities of race winners in such data, or trying to infer the composition of local boards, risk drawing incorrect inferences. This project seeks to achieve three goals. Firstly, I re-assess several published papers with simulated data for uncontested races to assess their sensitivity to this data issue. I also provide researchers with the sharp bounds on many classic regression estimands without additional data. Finally, I am working on collecting, for school board and city council races in California, a new panel dataset on incumbent officeholding and candidate registrations. The new data will allow scholars to answer previously intractable questions in local political economy.

Author

• Samuel Asher, Stanford University