The takeaways

• Letters boosted turnout in our Social, but not our Political, campaign
• Democratic-leaning white voters turned out at an extremely high rate

The backstory

In the world of politics, odd-numbered years are often called “off years,” since there are no big national elections. 2023, however, was an “off year” that felt more like an “on year” in many places, including Wisconsin, where a state supreme court race garnered national attention and over 50 million dollars of campaign spending. The next elected justice would rule on key issues such as abortion access and fair maps, and though the race was nominally nonpartisan, the two candidates displayed dramatically different stances. As one example, Janet Protasiewicz had stated her personal opinion on abortion: “That should be a woman’s right: to make a reproductive health decision.” On the other side, her opponent Dan Kelly compared abortion to murder. We expected that most voters would be unfamiliar with the two supreme court candidates in Wisconsin, so this race offered another good opportunity for us to provide information about candidates to voters through handwritten letters.

Nevada’s secretary of state race in 2022 had been the first time that Vote Forward ventured into naming candidates in our handwritten letters. Following that model, both our Social and Political letters in Wisconsin named the candidates and provided a link to a nonpartisan voter guide, and the Political letters went a step further by including quotes that contrasted the candidates’ opinions. Neither letter told the voter who to vote for; at Vote Forward, we believe that one of the reasons our letters succeed is that we demonstrate respect for voters’ ability to make their own choices.

Our Social letters were sent to historically underrepresented voters - specifically, voters of color - while our Political letters were sent to Democratic-leaning voters. Since these filters are applied sequentially, this meant that the Political letter recipients were almost entirely white. All voters were modeled as relatively unlikely to vote in off-year elections.

When we looked at the results, that difference in voter targeting turned out to be highly meaningful. Our Social letters demonstrated solid evidence of impact at +0.6 percentage points, while our Political letters showed an effect too small to interpret at +0.1 percentage points. (For more info on how we design our experiments and interpret results, please see our Research FAQs.) And when we looked at voter turnout numbers, a striking difference emerged: baseline turnout among Social campaign voters (33%) was less than half that of Political campaign voters (69%). It’s a rule of thumb that higher baseline turnout means smaller measured effects, and we certainly observed this pattern here. Essentially, we saw that the huge amount of money and media focused on Wisconsin’s supreme court race really had worked to mobilize voters - at least, it had mobilized Democratic-leaning white voters.

Unlike our previous study in Nevada, this study was not structured to measure vote choice, so we can’t directly say whether our letters affected the race’s outcome (which Judge Protasiewicz won by a large margin). However, we were happy to again see evidence that letters with information about candidates can boost voter turnout. In 2024, our Political letters build on this line of research by naming and providing information about candidates in key U.S. House and Senate races. Check them out, and our other letter campaigns, on our Campaigns page.