Your Neighbors Will Know: Social Pressure, Gratitude, and the Line Between Them

This is the first post in a four-part series we’re calling “What Actually Persuades.” Each post takes one mechanism with real experimental evidence behind it and shows how to put it on a postcard or in a letter to a voter. This week is social pressure, and the gentler version of it that a nonprofit can use. Next week is plan-making, and why asking a voter when and how they’ll vote beats asking whether. The third post is identity: the famous “be a voter” finding, what happened when a bigger team tried to replicate it, and what survives. The last post is persuasion itself: what the evidence says about changing the mind of someone who doesn’t already agree with you, and when in the calendar that’s even possible.

If you’re planning messages now for late-summer writing, that’s the order to think in: turnout first, persuasion last.

The mailer that worked too well

In August 2006, Alan Gerber, Donald Green, and Christopher Larimer mailed one of four postcards to about 180,000 Michigan households before a sleepy primary. The results, published in the American Political Science Review in 2008, are the most cited turnout experiment in the field.

The “Civic Duty” card reminded people that voting is a duty. It raised turnout 1.8 points. The “Hawthorne” card added that researchers would be checking whether they voted: 2.5 points. The “Self” card printed the household’s own voting record for recent elections and promised an updated copy after the election: 4.9 points. The “Neighbors” card printed the voting record of everyone on the block and promised to mail the whole street an updated chart afterward. That one moved turnout 8.1 points, an effect the size of door-to-door canvassing for the price of a stamp.

It also made people angry. The card listed a phone number, and many recipients called it to complain and ask to be taken off the list. Green and Gerber themselves have advised organizations not to use the heavy version because of the reputational risk. When real campaigns copied it anyway in later cycles, local news described the “voter report cards” as creepy and insulting. That’s the deal on offer: the biggest turnout effect in the literature, paid for with your organization’s name.

A nonprofit shouldn’t take that deal, and it doesn’t have to.

Where the effect actually comes from

Christopher Mann ran the obvious follow-up. In the 2007 Kentucky governor’s race, he mailed nearly 96,000 unmarried women either a copy of the “Self” mailer or one of three softened versions that kept the voter’s own record but dropped the heavy-handed language. Every version raised turnout by about the same amount, roughly two points on a very low baseline. The softer versions drew less anger when he tested them in a survey. His conclusion: the work is done by the credible sense that someone can see whether you voted. The shaming adds nothing except complaints.

Gerber, Green, and Larimer found something similar in a 2010 follow-up with 369,000 Michigan voters. A mailer showing the voter that they had “voted” in a past election raised turnout 4 points. One showing “did not vote” raised it 6.3. Shame is stronger, but praise works too, and praise doesn’t generate phone calls.

The gratitude version

Costas Panagopoulos took the praise idea further and tested a message with no surveillance at all: thank you for voting. In three experiments, a Staten Island special election in early 2009, the New Jersey governor’s race that fall, and a Georgia primary in 2010, a postcard thanking people for voting last time raised turnout 2 to 3 points over a plain reminder. He told Fordham that’s about five times the effect of a generic get-out-the-vote mailer, and that a generic thank-you worked as well as a thank-you with social-pressure language bolted on. The effect held across low- and high-propensity voters.

This is the one to build your fall program around. A handwritten thank-you from a neighbor is already the most credible version of this message a voter will ever receive, for reasons we covered in why handwritten communication works.

What a volunteer can write

Split your list by whether the voter cast a ballot in 2024. That’s on the file you already have.

For people who voted last time:

Thank you for voting in 2024. I’m writing because the election this November 3 matters just as much for our schools and our hospitals, and I hope you’ll vote again.

For people who didn’t, the honest, plain version of the “Self” mailer:

Whether each of us votes is a matter of public record. I’ve looked at my own record, and I’m glad it shows I vote. I hope yours will show November 3, 2026.

A line that works for either group, and that we’ll come back to in the third post of this series:

People in our neighborhood vote in every election, not just the presidential ones. I hope you’ll be one of them this fall.

Keep the rest of the card about the writer. The social-pressure sentence is one line, not the theme.

The line we don’t cross

The rule: you may tell a voter that voting is public. You may thank them. You may tell them, truthfully, that their neighbors vote. You may not show them anyone else’s record, name anyone, promise to report back to the street, or imply that you’ll be watching.

Our programs depend on volunteers who are proud to sign their first name to a card. Nobody signs a card that threatens someone. The voter who gets a Neighbors mailer remembers the organization that sent it, and the turnout gain over the gentle version is not worth a single angry call to your executive director. Mann’s numbers say you keep most of the effect without it. Take the gentle version and move on.

Next week: why asking a voter when and how they’ll vote beats asking if.