Voter Registration Experiments 2019-2020
Reporting and discussing the results of letter writing campaigns for voter registration
At Vote Forward, most of our campaigns try to address voter turnout directly, by sending handwritten letters that encourage people to vote. But there’s an even more fundamental barrier to participation: getting registered. In the U.S. - unlike many other democracies - the burden of registering to vote falls on individuals, not the government. As a result, many people just don’t do it: Around 30% of voting age Americans were unregistered as of 2016, according to the Census Bureau.
Over the past few years, we’ve experimented multiple times with adapting our handwritten letter program for voter registration. Most recently, we ran a randomized controlled trial in the 2021 Georgia runoff elections for U.S. Senate, where volunteers sent handwritten letters to over 200,000 unregistered people who we believed were eligible to vote. These letters included a URL for Georgia’s official state website, where the recipient could register online. In this trial, and in the others we’ve analyzed to date (including earlier trials in Georgia and Wisconsin), we hoped to see that the same kind of letters that have proven so effective for voter turnout would work their magic on voter registration too. To our disappointment, that’s not what we saw: These letter writing campaigns did not positively impact voter registration rates, based on the data we currently have.
But if this is true, why would voter registration work so differently from voter turnout? Once we started pondering this question, we started to see a clearer picture. For one thing, the data just looks a whole lot messier. When a voter registers, their name and address are recorded with the state government. While even this data is never perfect, it’s as close to perfect as we can get. But if someone is unregistered to vote, that data is less consistent and less accurate, and therefore it’s harder to ensure that each letter reaches the right person at the right place. For another, the people we’re trying to reach are, almost by definition, less engaged with politics in general. (If they were highly engaged, they’d probably already be registered.) We already know that this lack of engagement is a huge factor in why some people who are registered choose not to vote. It’s likely that, even if our data were perfect, Vote Forward’s registration letters would be reaching a lot of folks who are unregistered precisely because they’re detached from the democratic process in some way. That could be because they face barriers to participating, because they feel cynical or skeptical that it matters, or because they just don’t care. Whatever the reason, it means that they’re unlikely to get involved in general - and indeed, we see that the people we’ve included in our registration experiments registered to vote at extremely low rates over time, regardless of whether they got a Vote Forward letter.
Does this mean that it was a waste of our time - and your time - to try voter registration? Absolutely not! We run experiments like this to learn new things: If we knew what the outcome would be ahead of time, we wouldn’t need to test at all. In this case, learning that our standard approach to voter turnout doesn’t seem to yield the same results for registration is a crucially important lesson. It doesn’t mean that Vote Forward won’t try voter registration again, but it does mean that when we try again, we’ll have to try out something meaningfully different. And it means we can now spend our time figuring out how to create something that works better, rather than asking volunteers like you to spend your time on less effective campaigns.
Adapting what we do in response to new information - good or bad - has always been a core value for us. So, while we’d hoped to share more positive news in this update, we are excited about what future reimagined versions of our registration campaigns might look like.
