Search
Browse By Day
Browse By Time
Browse By Person
Browse By Mini-Conference
Browse By Division
Browse By Session or Event Type
Browse Sessions by Fields of Interest
Browse Papers by Fields of Interest
Search Tips
Virtual Exhibit Hall
Change Preferences / Time Zone
Sign In
X (Twitter)
Many states require election workers to try and validate signatures on the envelopes in which some voters return ballots by mail. However, even forensic experts make errors when checking signatures. I apply Bayes’ Theorem to show that, under a range of plausible prior beliefs about the (imperfect) precision of signature checks, and about the (low) prevalence of invalidly signed ballot return envelopes, one should expect dozens or even hundreds of valid signatures to be wrongly rejected for each invalid signature that is correctly rejected. This implies that, when more votes are cast by mail, election workers are likely to wrongly reject many mail ballots. And yet, there is a feedback mechanism likely to reveal that ballots were mistakenly rejected, since some voters whose ballots are wrongly rejected complain and/or seek to “cure” the issue. I expect that election workers calibrate their behavior in response to this feedback, becoming more cautious and rejecting fewer mail ballots. I test this prediction using variation in mail ballot return rates and rejection rates across states and counties. I also test whether the sudden increase in the numbers of ballots returned by mail in the 2020 primary elections resulted in fewer rejections for alleged signature mismatch when it came to the general election in the fall. I find support for the calibration argument, although other factors, such as the ease of the cure process, are also relevant.