Resources

In this paper, authors examine the effects of automatic voter registration (AVR) on both registration and turnout. They find that ind it does raise registration rates substantially, that the effect of AVR gradually builds the longer it is in place, and that the different types of AVR have significantly different effects on both registration and turnout.

The guide provides a basic framework for testing risk limiting audits in diverse contexts by outlining foundational prerequisites and operational, legal and regulatory considerations

Research shows that printed ballots pose challenges for blind voters and low-vision voters, who cannot read them directly.
This is a report of a project examining the legibility of summary ballots printed by ballot marking devices. The goal of the investigation was to to identify aspects of design, layout, or typography that can make a summary-style ballot easier to read and to increase the likelihood that voters detect a mistake or change on their ballot.

This report highlights key trends in voter turnout among voters with disabilities in the 2020 election. Authors note a significant increase in turnout among these voters compared to the 2016 general election, reducing the turnout gap with voters without disabilities to 5.7 percentage points. However, approx. 11% of voters with disabilities reported some difficulties voting.

In this paper, authors examine whether use of vote by mail exacerbates existing inequalities in mail-in ballot rejection rates. They find that younger voters were more likely to have their vote by mail ballot rejected than older voters due to lack of timeliness or deficiencies with the return envelopes. These findings highlight potential adverse impacts of expanded vote by mail on certain groups.

This paper examines the barriers that Native Americans face
when trying to register and participate in elections. Through conducting several field hearings, authors uncover several important findings related to polling place access and first generation voting barriers that prevent them from casting a ballot.

Authors describe the difficulties that vote-by-mail presents for Native American voters. Specifically, “members of the 574 federally recognized tribes” face barriers to political participation to a greater degree than any other racial or ethnic group. The authors also define measures that can be taken to level the field, all while respecting social distancing.

Cantoni examines the effect of voting costs - specifically distance to a polling location - on ballots cast. He finds that small increases in distance to a polling location reduce ballots cast. He also finds that during non-presidential elections, these effects are three times larger in high-minority areas than in low-minority areas.

In this paper, Merivaki and Smith seek to examine who is more likely to cast a provisional ballot and why some provisional ballots are rejected. They suggest that beyond individual-level factors, there are administrative reasons why some prospective voters are more likely to be required to cast provisional ballots, and why some provisional ballots are rejected.

In this paper, authors argue that risk limiting audits can verify tabulation of paper records but cannot by themselves ensure ballot marking device printouts reflect voter intent if voters do not verify them.

This report summarizes projections of how many people with disabilities would be eligible to vote in the November 2020 elections, using data from the Census Bureau’s 2014-2018 American Community Survey combined with Census Bureau population projections for 2020-2021. The report breaks down key demographic characteristics of eligible voters with disabilities.

This report details the design and usability testing of an accessible ranked-choice voting ballot interface built to work for voters who are blind or very low vision, have limited or no use of their hands, or have cognitive or attention disabilities. Testing with participants with various disabilities found that no single design can be fully optimized for all modes, but a workable default is achievable.