We write peer-to-peer scripts that change the question a voter is answering. Not "will you vote?" but "will you do this one thing for me?" Delivered one person at a time through Culper, our P2P texting system, they turn supporters into voters who show up.
"Please vote" is a request from many people to many people, so each voter feels their part is small and optional. Our scripts turn it into a specific request from one real person to one voter, like "Can you stand our sign up at your polling place Tuesday at 7:00?" A personal, specific ask carries social weight, gets a commitment, and builds a voting plan, and the volunteer ends up at the polls. URSA writes these scripts for Culper's peer-to-peer texting, where volunteers send them one conversation at a time.
Every get out the vote message makes a request. What decides whether it works is who seems to be asking, and how many other people it seems to be asking.
So we stop asking people to vote. We ask them to help, and the vote comes with the favor.
Psychologists have a name for what happens when a request goes to a crowd: diffusion of responsibility. When many people could act, each one feels less responsible for acting. It is the reason a person in trouble on a busy street can be passed by hundreds of people, and the reason a turnout message aimed at an entire precinct quietly asks nobody in particular.
Voting has this problem built in. Every voter knows one ballot rarely decides an election. A generic reminder from a campaign, even a well designed one, reinforces it: it is obviously going to thousands of people, it asks for something abstract, and it never says when or where. The voter agrees that voting is important, and then does not vote.
Turnout research points the same direction. Across hundreds of field experiments, contact that feels personal and conversational moves more people to the polls than contact that feels mass produced. A real person talking with a voter beats a postcard, and a postcard beats a robocall. The more a request feels like it came from someone, for someone, the more it works.
This is the script that shows the idea best. A volunteer texts a supporter and asks them to stand up a campaign sign at their own polling location on Election Day, at a specific time.
Look at what that yes contains. Maria agreed to be at her polling place, at a set time, on Election Day, for someone who is counting on her. She is not being asked to vote. She does not have to be. Once she is standing at the door of her polling place, she votes. The request moved from "many to one" to "one to one", from an abstract duty to a small favor, and the favor carries the vote with it.
Hi Maria, it's Dana with the Yes on Prop A team, I live over on Briar Creek. Quick favor?
Sure, what's up?
We need someone to put our sign up at Lakewood Elementary, your polling place, on Tuesday. Could you take 7:00 to 7:30 when the polls open? Takes two minutes.
I can do that before work
Thank you! I'll text you Monday night with where to pick up the sign. Reply STOP to opt out.
None of this is a trick. Each piece of the script lines up with a well studied finding about how people decide to act. Here is the research and how we put it to work.
What the research foundWhen people believe many others could respond, each person is less likely to step in. The effect drops sharply when one person is clearly the one being asked.
How we use itEvery script names the voter, comes from a named volunteer, and makes the task theirs alone. "Can you take 7:00?" leaves no one else to do it.
Darley and Latané, Journal of Personality and Social Psychology, 1968.
What the research foundIn a classic study, homeowners who first agreed to display a small sign about safe driving were far more likely, weeks later, to agree to a large, unattractive sign in their front yard than homeowners asked only the big request. Saying yes to something small changes how people see themselves.
How we use itWe start with a request almost anyone will grant: a quick favor, a sign, a time slot. Each yes makes the next one, including the ballot, more natural.
Freedman and Fraser, Journal of Personality and Social Psychology, 1966.
What the research foundPeople feel strong pressure to act in line with what they have already said, especially when the commitment is active, voluntary and made to another person.
How we use itThe voter types the yes themselves, to a real person. That reply is a public, active commitment, and we follow up to confirm it the night before.
Robert Cialdini, Influence: The Psychology of Persuasion, 1984.
What the research foundVoters asked to think through when they will vote, where, and how they will get there turned out at higher rates than voters given a standard encouragement. The effect was strongest for people who live alone, who have no one else in the house to remind them.
How we use itThe sign ask is a voting plan in disguise. It sets the place (their polling location), the time (7:00) and the reason to be there. The voter never has to make the plan, because the favor made it for them.
Nickerson and Rogers, Psychological Science, 2010.
What the research foundPeople asked about "being a voter" turned out at higher rates than people asked about "voting". A noun invites people to claim an identity, and people work to live up to identities they value.
How we use itOur scripts make the voter part of the team: "our sign", "we need someone", a volunteer, a precinct captain, a neighbor helping neighbors. People who help run the effort see themselves as people who show up.
Bryan, Walton, Rogers and Dweck, Proceedings of the National Academy of Sciences, 2011.
What the research foundA large field experiment found that turnout rose by about eight percentage points when voters were told their neighbors would see whether they had voted. Visibility to other people is one of the most powerful turnout levers ever measured.
How we use itWe do not shame anyone, and we never publish who voted. We get the same effect in a positive way. A person standing a sign up at the polls is visibly part of the election, in front of their own neighbors, and a volunteer is expecting them.
Gerber, Green and Larimer, American Political Science Review, 2008.
What the research foundDecades of randomized turnout experiments find that the closer a contact feels to a real conversation with a real person, the more it raises turnout. Impersonal, mass produced contact does far less.
How we use itThis is why the scripts run through peer-to-peer texting, not blast texts. A volunteer sends each message by hand, reads the reply and answers it. It is a conversation, which is the thing the research says works.
Green and Gerber, Get Out the Vote: How to Increase Voter Turnout.
The sign is one script. Each of these makes the same move, from a general duty to a specific favor from a specific person, and each one puts the voter at or near the polls.
"Can you stand our sign up at your polling place at 7:00 Tuesday?" Time, place, and a reason to be there when the doors open.
"Mrs. Alvarez on your street needs a ride to vote Thursday. Could you take her at 10:00?" Two voters, one trip, and nobody wants to let a neighbor down.
"Could you bring two people from your street when you vote? Text me their first names." The voter becomes the organizer, and naming people makes it real.
"We need someone to hand out our card outside the 100 foot line at the library for one hour Saturday. Morning or afternoon?" Offering a choice of two times turns "maybe" into a pick.
"Can you text me how long the line is when you vote? I'm telling the rest of the street when to go." A small job that only makes sense once they are at the polls.
"Early voting is open at the MUD building through Friday. What day works for you, so I can check you off our list?" Asking which day, not whether, assumes the yes.
Every script we write for Culper P2P follows the same structure. Each line has a job.
The volunteer says who they are and why they are the one writing: same street, same school, same church, same subdivision. A name turns a campaign into a person.
"Quick favor?" invites a small yes before the request arrives. People who answer at all are already in a conversation, not reading an ad.
Never "help out" or "get involved". One concrete thing they can picture doing: this sign, this ride, this hour.
Their polling location and a time slot. Specific beats flexible, because flexible is how good intentions slip to next week.
"Takes two minutes." The ask has to be small enough that saying no would feel strange.
"7:00 or 7:30?" "Morning or afternoon?" Choosing between two options is easier than deciding whether to do it at all.
The volunteer thanks them and promises a reminder, then actually sends it the night before. The commitment is renewed when it matters.
Opt out language on the first message, texting hours honored and every STOP respected. Culper enforces these automatically.
A one to one ask only works if it is actually one to one. Mass texting platforms that blast the same message to a list look and feel like what they are. Culper, the campaign software we built, runs true peer-to-peer texting: a volunteer works a queue of voters, loads the campaign script, makes it sound like themselves, and sends each message by hand. When the voter replies, it lands in the same conversation and the volunteer answers.
Culper also handles what makes P2P safe to run at scale. The first message to any voter carries opt out language. Sends outside the campaign's texting hours are held. Every STOP goes on a suppression list no campaign can override. Each household gets one conversation, not five.
Most important, every conversation ends with a disposition: supporter, committed to a sign shift, needs a ride, undecided, opposed. That disposition goes straight to the voter record. The night before Election Day we know exactly who said yes to what, and the follow up texts go to exactly those people. After the polls close we match the early vote roster and the voter file against the commitments, so the campaign can see who kept their word.
Supporters and lean supporters with a usable mobile number, deduplicated to one conversation per household, scored for how likely they are to vote without help. The people who need the nudge get the script.
We write the opener, the asks and the replies to the common answers ("I already voted", "I work that day", "Who is this?") in the campaign's voice and the community's language.
Volunteers work their own neighborhoods wherever possible, because "I live on your street" is the strongest opener there is. Culper splits contacts evenly and rebalances as new texters join.
The first asks go out a week or more ahead. Confirmations go out the night before and the morning of. Early voting is part of the plan, with its own asks.
Each day of early voting, voters who have cast a ballot come off the list, so volunteers spend their time on people who still need to vote.
Commitments made, commitments kept, and turnout among people who said yes compared with similar supporters who were not asked. Real numbers, not impressions.
Everything in our scripts is true. We ask real people to do real things that really help the campaign, and every sign shift, ride and hour at the door is real work that needs doing. We do not invent urgency, we do not shame anyone for not voting, and we never reveal anyone's voting history to anyone. People can say no, and when they say stop, we stop.
A one to one ask is a request from one identifiable person to one voter for one specific thing, like standing up a sign at a polling place at a set time. It works better than a general "go vote" message because the voter is clearly the one being counted on, the task is concrete, and saying yes is a commitment to a real person.
Most turnout messages come from an organization, go to everyone, and ask for something abstract. That spreads responsibility across thousands of people so each voter feels their part is optional, a pattern psychologists call diffusion of responsibility. Specific, personal requests with a time and place avoid it.
Yes. Peer-to-peer texting, where a person sends each message individually, is widely used by campaigns. It still has rules: identify who is texting, include opt out language, honor every STOP request and respect texting hours. Culper enforces these automatically. Campaigns should still review their own program with counsel.
Mass or broadcast texting sends one message to a whole list at once. Peer-to-peer texting has a volunteer send each message individually and carry on a two way conversation with the reply. P2P feels personal because it is personal, which is why our persuasion scripts run on it.
It puts them at their polling place, at a set time, on Election Day, for someone who is counting on them. That combines several of the strongest turnout effects in the research: a specific plan, a public commitment, a personal request and being seen. URSA tracks commitments and turnout in Culper so each campaign can see the results in its own race.
We write them with the campaign. URSA drafts the opener, the asks and replies to common answers, then tunes them to the candidate's or committee's voice and the community. Volunteers are coached to make each message sound like themselves.
We will build the list in Culper, write the scripts for your race and train your volunteers to run them.
Schedule a Consultation