Make a Plan: Why Asking When and How Beats Asking If

This is the second post in our “What Actually Persuades” series. Last week we covered social pressure and the gratitude version of it that a nonprofit can use. This week is the cheapest turnout mechanism in the literature: helping a voter make a plan.

The experiment

In the days before the 2008 presidential election, David Nickerson and Todd Rogers had a phone bank call registered Democrats in a competitive state. The study, published in Psychological Science in 2010, randomized about 287,000 people across three scripts.

The first script was a standard reminder: the election is Tuesday, please vote. The second added one question: do you intend to vote? The third added three more: what time will you vote, where will you be coming from, and what will you be doing beforehand?

The reminder did nothing measurable. Asking whether people intended to vote did nothing measurable either, which contradicted older, smaller studies claiming that getting someone to predict they’ll vote makes them vote. The three plan questions raised turnout by 4.1 points among the people who picked up the phone.

Among people who were the only eligible voter in their household, the plan script raised turnout by 9.1 points. Among people who lived with another voter, none of the scripts did anything at all.

Why planning works

Psychologists call this an implementation intention. An intention is “I’m going to vote.” An implementation intention is “After I drop the kids off on Tuesday, I’ll stop at the library on Elm Street and vote.” The second one names a time, a place, and a trigger. When the trigger shows up, the behavior follows without a fresh decision.

That matters because most people who skip an election didn’t decide to skip it. They meant to vote, and then Tuesday was a normal busy Tuesday, and it didn’t happen. A plan converts a vague wish into a cue that fires on its own.

The household finding makes sense in that light. People who live with another voter tend to make the plan anyway, in passing, over dinner: “Are we going before work?” Nickerson and Rogers found that people in single-voter households were more likely to answer “I don’t know” to every plan question. Nobody had asked them. The phone call was the first time anyone had.

Three questions to write on a card

One caveat first. This experiment used phone calls, where a caller asks and the voter answers out loud. We’re not aware of a published experiment that tests the same three questions on handwritten mail at this scale. Our reasoning is that the mechanism lives in the voter’s head, not the caller’s ear. The plan is what matters, and a card can prompt a plan. That’s a reasonable bet, not a proven one. If your program can hold out a control group this fall, this is a good thing to test, and we wrote about how to do that without kidding yourself.

The three questions, as a volunteer might write them:

When will you vote on Tuesday, November 3? Before work, at lunch, or on the way home?

Where is your polling place? Mine is the elementary school on Main. Yours is on your county’s elections website.

What will you be doing right before you go? I’ll be coming from work, so I’m going straight there.

The first offers choices, because “when will you vote” alone gets “sometime.” The second models the answer, because a voter who has to look something up later won’t. The third is the trigger, and it’s the one programs leave out. “Coming from work” or “after school pickup” is the cue that makes the plan fire.

If your card has room for one more line, make it a fill-in: “I’ll vote on __ at __, after ___.” A voter who writes three words on your postcard has made a plan. A voter who reads it has probably made one too.

Two targeting notes. First, your voter file has household size on it. If you have to choose, write plan-making cards to one-voter households, where the evidence says the effect is largest. Second, this message belongs on the last card a voter gets, close to the election, not in July. A plan made in July is a plan forgotten.

Vote-by-mail states need a different plan

Everything above assumes a voter walks somewhere on a Tuesday. In Oregon, Washington, Colorado, Utah, Nevada, California, and the other states where every registered voter is mailed a ballot, the decision moment isn’t Tuesday. It’s the afternoon the ballot shows up in the mailbox. That’s the moment to plan around, and it needs different questions:

Your ballot will be mailed to you around October 16. When it arrives, where will you put it so it doesn’t get buried? Mine goes on the fridge.

How will you return it? There’s a drop box outside the library on Main, or you can mail it. Mail it by Friday, October 30, to be safe.

After you send it, check that it counted. Most states let you track your ballot online.

Get the dates right for your state. Ballot mailing dates, postmark rules, and receipt deadlines all differ, and a wrong date on a handwritten card is worse than no date. And time the card to land a few days before ballots go out, not after. A voter who already mailed their ballot and then receives a card telling them to make a plan feels managed, a feeling we covered in the primary postcard post. The fix there works here: “If you’ve already sent your ballot back, thank you.”

The plan-making message is three sentences, and the evidence says it’s worth more than the civic-duty paragraph most programs spend that space on. Cut the paragraph. Ask the questions.

Next week: the “be a voter” finding, the replication that didn’t find it, and the identity language that still holds up.