Ballot Rolloff and Ballot Design

I had an opportunity this week to attend the annual conference of the California Association of Clerks and Election Officials. I was on an interesting panel discussing ballot rolloff and the implications of long ballots in California for voters.

I talked about preliminary research that Yimeng Li and I have been doing on a natural experiment that has been ongoing in Los Angeles County. As the result of state legislation, LA County reordered some of the races on recent ballots, effectively moving some of the local races closer to the top of the ballot and pushing some of the more salient statewide races towards the bottom of the ballot.

At this point, we’ve looked at election results from LA County, compared to other large counties in the state where this ballot change did not take effect. We use the natural experiment to produce some causal estimates (using “difference-in-difference” methods) that show that this change in LA County did increase the percentage of ballots where voters cast ballots in city and school board elections, relative to counties where the change did not take place. The slides from the presentation are available. We are actively working with new data to extend this analysis in a number of ways, stay tuned.