2020 Was a Banner Year for U.S. Election Administration

MIT’s Election Performance Index shows improvement across the board.

March 10, 2022 Claire DeSoi MIT Election & Data Science Lab

Despite widespread claims that the 2020 election was fraudulent or poorly managed, election administration did not just persevere under unexpected and challenging conditions—it improved.

The 2020 election was an anomaly in many ways, but the Election Performance Index (EPI) shows us that election administration continued to trend in the direction it was already heading: up. While there was more early voting and voting by mail than previous years (a pre-existing trend that 2020 accelerated), overall, more states improved their practices, data, and reporting.

When looking at the index for 2020, we must of course remain mindful of the 2020 election atmosphere and context. In an election year like 2020, though, where administrators and election officials had to adapt quickly to unprecedented challenges, data-driven measures became even more important in finding and telling the story of how elections in the US are managed.

About the EPI

The EPI was the first objective measure of election performance in each state for U.S. midterm and presidential elections. It began tracking election administration with the 2008 election. Innovative and unique, it provides a comprehensive assessment of how election administration functions in all 50 states and Washington, DC. Importantly, the EPI allows us to identify the specific parts of election administration that have improved, and to see the parts of the system that have lagged behind. Its data-focused approach means it can serve as a foundation for discussions about the challenges that arise for election administrators, and shed light on the opportunities each state has to address them.

What does the 2020 Index tell us?

Overall, the 2020 EPI tells us that U.S. election management was better than ever. The majority of states improved their election administration, despite adverse conditions; a whopping 45 states saw their index scores rise from the previous presidential election.

“The blocking and tackling of running elections has been very good,” says MIT Election Lab director Charles Stewart III, “and it got better in 2020.” The curveballs delivered by 2020 may have forced election administrators away from the expected, but broadly speaking, they delivered results: states performed essentially as they have in previous elections. While that might not seem triumphant, it means that for all of 2020’s challenges, the fundamental outcomes of election administration were consistent between 2016 and 2020. States held their own, wherever they were, with most states improving.

Two highlights to whet your appetite:

• Voter turnout increased for every single state in 2020, despite the added complexities COVID introduced.
• Data quality continues to improve. While some patchiness remains, we’ve seen significant progress in overall data quality (just compare the indicator for the 2008 and 2020 elections). This is undoubtedly helped by recent improvements in the EAC’s Election Administration and Voting Survey, as well, which is a key source of data for the EPI.

"The blocking and tackling of running elections has been very good, and it got better in 2020."
-Charles Stewart III

Effects from COVID-19

When the Covid-19 pandemic shut most public life and spaces down in early 2020, election scientists and administrators scrambled to rework long-standing plans in order to ensure that the primary and general elections were able to proceed safely and fairly. (For more information on those efforts, see the Healthy Elections Project, our collaborative effort with Standford.)

As many states shifted to expand their vote-by-mail policies, and pandemic shut-downs necessitated changes in polling place locations, worries rose that voters would be confused and the election would be chaotic. Happily though, we don’t see significant evidence of this in the EPI. Whether that’s due to the increased efforts of election officials, or voters simply paying closer attention to details, it is a welcome insight.

A few other interesting notes on the effects we see from Covid-19 in the EPI:

• Absentee rejections. Looking at the EPI’s calculations for absentee ballot rejections, most states improved their scores overall from 2016 to 2020, which indicates that the volume of mail ballots increased faster than rejections—good news, especially for a year where many states had to take a significant and accelerated pivot to mail voting.
• Voting by mail. When looking at the data for mail ballot rejections, we noted that signature-related reasons for rejection were higher in vote-by-mail states when compared to other states, as other reasons for rejection decreased.
• Wait times. Looking at data from the Survey of the Performance of American Elections, we noted that wait times to vote increased in 2020. We took a closer look at where wait times rose versus where they did not in this blog article about the effects of states’ decision to change (or not) their policies on absentee voting and voting by mail.
• Residual vote rate. This rate decreased from 2016 to 2020, although we might have expected it would go up with the increase in mail ballots. In 2016, however, there were many more abstentions than usual; 2012 may in fact be the better comparison. Look for a blog post on this issue in the coming days.

New indicators in the 2020 EPI

Longtime users of the EPI will note that we added new indicators this year! We will continue to write about these indicators and our decision to include them, but a few notes of explanation are appropriate here.

We have been preparing to adjust some of the indicators in the EPI for several years, with the goal of more accurately reflecting the evolving landscape of election administration. We’ve had time to review and improve some of the methodology behind the index indicators, and in many cases the reliability and improvements in the data available has opened up the possibilities available to us. Here is a quick run-down of our changes; stay tuned for more information on all of these to come:

• Replacing “Disability- and illness-related voting problems” with a “Voting Access” indicator, complete with an adjusted calculation, better reflects the intent to measure the access to voting among people with disabilities.
• Adding an indicator on Risk Limiting Audits, in addition to the existing “Post-election Audits” indicator, recognizes the advances in auditing, and boosts those that have made that move without reducing index scores for states that are pursuing more traditional auditing practices.
• Including an indicator on ERIC Membership encourages states to improve their practices on list maintenance while maintaining transparency and privacy.

What effect do these changes have on the EPI?

As policies and best practices change, it makes sense that the EPI must change as well. These updates make the EPI a better and more accurate measurement of the state of election administration in the U.S. today. It hones the indicators to reflect the improvements in data we’ve seen in the past decade. We are confident that the 2020 EPI will stand right along with its previous iterations as the premier assessment tool to evaluate election administration and management across the United States.

This article was originally published on the MIT Election Lab website.