Replication Data for: Editor’s Choice: Measuring Candidate Quality using Local Newspaper Endorsements

I construct a new measure of candidate quality differentials using local newspaper endorsements. I argue that political endorsements made by newspapers can be used as expert opinions that reflect quality differences between the candidates in an election. Using a dataset of 21,095 local newspaper endorsements, I simultaneously estimate the quality differences between candidates in 6,432 elections, along with a dynamic measure of the partisan bias of 368 local newspapers. Using the new measure, I show that a one standard deviation increase in relative candidate quality increases a candidate’s two-party vote share by 3.4 percentage points, and that candidate quality accounts for about one-fourth of the incumbency advantage. These findings advance debates on the source of incumbency effects and demonstrate the broader electoral impact of candidate quality. I conclude by discussing the potential of these endorsement-based measures to enhance our understanding of candidate quality in electoral politics and governance.