The Takeaways

• Vote-by-mail letters encouraged voters to turn out for Pennsylvania’s municipal primaries
• Small changes to letter templates didn’t make much difference

The Backstory

2020 was a historic year for voting, and especially for voting by mail. That’s why we ran two experimental campaigns in Maine and Wisconsin focused on boosting mail and absentee voting. In those campaigns, Vote Forward letters not only encouraged voters to choose mail voting more often, but also raised voter turnout overall.

But we want to make sure this is an approach that works for all voters, not just some. Our first experiments showed positive results in a highly limited subset of voters. To really feel confident, we wanted to test our vote-by-mail campaigns in a more diverse group of voters — after all, the United States is a diverse country!

We took advantage of Pennsylvania’s municipal primary elections in May 2021 to run our test with 200,000 registered voters of color. As of 2019, no-excuse mail voting is available to all Pennsylvania voters, but those who are not on the permanent list must request a mail ballot for each election. Thus, our letters focused on helping voters learn about options to request their ballot, as well as key deadlines.

We also took this opportunity to test a question we’ve had on our minds for a while: Do small wording changes in the letter templates make a difference? In Pennsylvania, we compared three versions of the template head-to-head. One version resembled our usual letters, and two included an additional sentence to frame mail voting as a normal, common activity:

• “Many people vote by mail” (“Static” norm)
• “More and more people are voting by mail” (“Dynamic” norm)

Other research has found that this kind of language can be especially motivating, so we wanted to see if these templates were more effective compared to our standard design. 100,000 voters were randomly split across these letter types, and the other 100,000 were held out as a control group.

Once again, we saw that vote-by-mail letters work! Compared to the control voters, the voters who received a letter voted at a rate that was 0.4 points higher, and in addition, they voted by mail at a rate that was 0.2 points higher.

These impacts didn’t seem to change much across the three template types. While there are small differences, they are so small that they could just be random noise.

The good news here is that all of the letters were effective at motivating voters to show up and vote by mail. These are smaller effects than we observed in our previous vote-by-mail campaigns, but they are still meaningful, especially in local elections where just a few votes can have outsized impact. And the fact that we didn’t see differences across letter templates is useful information—it means we can focus instead on trying other kinds of changes that might cause bigger effects.

For us at Vote Forward, this means that mail voting is here to stay. Our 2020 experiments provided the first evidence that vote-by-mail letters work, and 2021 reconfirms it. In our Virginia campaigns (happening right now!), in next year’s midterms, and beyond, we’ll keep helping voters take advantage of this safe, convenient way to vote.