The political scientists — Alan Gerber, Gregory Huber, Marc Meredith, Daniel Biggers, and David Hendry — undertook a unique experiment in cooperation with the Connecticut Secretary of State. They identified a group of about 6,400 former felons who were eligible to vote under Connecticut law, but had not registered to vote. This group of former felons had been incarcerated for less than three years and had not been convicted of “sexual assault, crimes with children as victims, or crimes resulting in the death of another person.” (Several other types of felons were also excluded from the study. The paper has more details.)
The researchers then divided these ex-felons into three groups. One received no contact. The other two groups received a voter registration card and letter from the Connecticut Secretary of State a couple weeks before the deadline to register. The letter always included the statement “According to our records, you are eligible to register and vote” (emphasis in original) as well as information about the election, a reminder that they were not registered, and appeals to civic duty. For one of these two groups, the letter also said this:
According to the current election rules, once you have successfully registered, you will not have to explain why you were previously unregistered when you show up to vote on Election Day. Election officials at the polling place will also not ask you any questions about your past. You will be treated with the same respect due to all of those who vote in Connecticut’s elections. (emphasis in original.)
This language was intended to determine if providing additional assurance had additional impact.
What were the results? Rates of voter registration and turnout were very low among all three groups. However, they were higher among those who were sent a letter and voter registration card. In the group that wasn’t contacted, 6 percent registered to vote and 3 percent actually voted in November 2012. But among those sent a letter, the registration rate was almost 8 percent and the turnout rate was about 4 percent. Gerber and colleagues also found that the impact of the letters was concentrated among those who had previously voted in 2008 — suggesting, as other research has found, that get-out-the-vote efforts typically motivate a population that is already somewhat politically engaged.
These are small increases, to be sure, but they are statistically meaningful. And they raise interesting questions about whether interventions more elaborate than a single letter might have a larger impact.
Of course, Gerber and colleagues are well-aware that there are significant political disagreements about whether ex-felons should be notified about their eligibility to vote. They see their study as a piece of evidence that “policy-makers can use to help judge the efficacy of these sorts of notification efforts.”