Skip to the content


As the nation’s election community continues to grapple with the impact of the COVID-19 pandemic, calls are increasing for the nation to shift primarily to vote-by-mail (VBM) ballots as a way to protect voters and poll workers in November. That push is raising some concerns that populations at greater risk of COVID-19—especially voters 65 and older and African Americans–are also those who would be less likely to cast ballots by mail. That is, an increase in VBM could result in a decrease in electoral participation by those at highest risk of COVID-19.

To test this concern, I analyzed county VBM patterns from 2018, with an eye on usage by African American and older voters. That analysis revealed that counties with higher proportions of these voters had higher VBM rates in 2018. This suggests that moving the November 2020 elections primarily to VBM could help reduce the exposure of high-risk groups to COVID-19 without reducing their anticipated participation in the election.

The data for this analysis comes from the U.S. Election Assistance Commission’s (EAC) 2018 Election Administration and Voting Survey (EAVS), which, among other things, asked counties to report their VBM rates for the 2018 General Election. Using those totals, along with estimated population figures from the Census Bureau’s 2018 5-year American Community Survey (ACS), I constructed a model to test the relationship between a county’s VBM rate and the proportion of its population that is “at risk,” defined as either African American, aged 65 or older, or both. You can see the results below:

Predicted VBM Rate

The results are statistically significant at 95% or better and are further detailed in the full analysis here.

These results suggest that a hypothetical county with no at-risk voters (i.e., neither African American voters nor voters 65 and older) would see a 19% VBM rate, with those totals increasing (especially for counties with a higher number of older voters) as the proportions of these at-risk populations increase.

The analysis suggests this relationship is even stronger in counties that do not require voters to provide an excuse when requesting VBM ballots:

Predicted VBM Rate

The bottom line? Contrary to concerns about a disproportionate negative effect of VBM on participation by African American and voters 65 and older, expansion of VBM to protect these groups from increased risk of contracting COVID-19 this fall won’t disadvantage them at the ballot box.

Here at Fors Marsh Group, we will continue to review the data and look for insights on what it tells us about the nation’s election system in November 2020 and beyond.

About the author

Carl Turner, Ph.D.

Carl Turner, Ph.D.

Carl Turner joined Fors Marsh Group (FMG) in October 2012. Dr. Turner works on FMG’s Elections Research team where he has been working on the Overseas Citizen Population Analysis project for the Federal Voting Assistance Program (FVAP). In addition to various market segmentation projects for the Department of Defense (DOD) and the Food and Drug Administration (FDA)Dr. Turner has also worked on a national survey of tipping behavior for the Internal Revenue Service (IRS). 

Dr. Turner graduated with a BA in economics and history from the University of North Carolina at Chapel Hill and received his master’s degree in public policy from the College of William and Mary. He received a PhD in public policy from George Mason University in 2012Dr. Turner’s research interests include local economic development, urbanization, and their relationship with political institutions. He has authored research papers that have been published in journals such as The Annals of Regional Science and Growth and Change. 

Let's Work Together

We’re here to help, so ask us anything.

Insight Delivered

Subscribe to our newsletter to stay up-to-date on our latest news and outlook.

Please enter a valid Email address