Racial Bias in Election Related Signature Verification: A Laboratory Experiment

• In Event: Causes and Consequences of Variability in Localized Election Administration

Abstract

The COVID-19 pandemic prompted a significant shift towards non-traditional voting methods, with 43% of voters opting for mail-in ballots in the 2020 general election. However, this transition has introduced a new challenge: ballot rejection. In the 2020 election, over 560,000 mail-in ballots were rejected nationwide with nonwhite voters disproportionately getting their ballot rejected due to a signature discrepancy compared to white voters.

This study investigates the potential racial bias among ballot reviewers contributing to these disparities. I present a series of original laboratory experiments simulating working conditions and signature specimens encountered by poll-workers, using undergraduate students as experimental subjects. My preliminary results indicate that participants with negative racial attitudes are significantly more likely to reject nonwhite signatures compared to white signatures. This finding underscores the risks associated with discretion in election administration, reminiscent of historical practices such as the use of understanding clauses to screen out minority voters in the Jim Crow South.

Moreover, I assess the alleviating effects of four election reforms (de-biasing training, increasing cognitive resources, increasing review committee size, and the presence of ethnic confederates) designed to reduce racial bias in signature review. By focusing on the specific mechanisms and outcomes of these reforms, this research provides valuable insights for policymakers and election officials seeking to address the issue of ballot rejection and promote fair electoral practices. In other words, this study aims to scientifically validate best and worst practices for mail-in ballot signature review.

Author

• Michael Herndon, University of California, Los Angeles