Which get out the vote tactics work best? What two decades of field experiments show

The get out the vote tactics that work best are the personal ones. Randomized field experiments summarized by political scientists Donald Green and Alan Gerber find that door-to-door canvassing produces about one additional vote for every 15 voters a canvasser actually talks to, volunteer phone calls about one vote per 35 contacts, and a standard nonpartisan mailer about one vote per 273 recipients (Bustle, 2018). Robocalls show no detectable effect (CIRCLE, Tufts). On the message side, helping voters make a concrete voting plan raised turnout 4.1 percentage points among people reached (Nickerson and Rogers, 2010), and the best-known social pressure mailer raised it 8.1 points (J-PAL).

Those numbers come from randomized experiments: split a voter list, contact one half, then compare turnout in the public voter file. This guide ranks the tactics by that evidence. For the calendar, staffing and budget side, read our companion piece on how to plan a get-out-the-vote program.

What get out the vote tactics are most successful?

GOTV has one job: turn people who already support you, or who are eligible and unlikely to vote, into ballots cast. The League of Women Voters' evidence summary puts it plainly: the goal of voter mobilization is "to transform nonvoters into voters, not change minds" (LWV). Persuasion is a separate program with separate math.

Here is how the main tactics rank in the experimental record.

Canvassing, phones and texts are channels. Vote plans, social pressure and gratitude are messages that ride on a channel. Pick the most personal channel you can staff, then load it with the messages that tested well.

Why personal contact beats every other GOTV method

The modern evidence started with a 1998 experiment in New Haven, where Gerber and Green tested canvassing, phone calls and direct mail against a control group (American Political Science Review, 2000). Face-to-face canvassing produced an 8 percentage point turnout increase among the people contacted, while each piece of mail added about half a point (Gerber and Green, Handbook overview).

Hundreds of experiments later, the pattern holds. A meta-analysis of more than 200 GOTV experiments found average effects of 1.0 percentage point for door-to-door canvassing, 0.7 for direct mail and 0.4 for phone calls (Green, McGrath and Aronow 2013, as summarized by Enos et al.). Those are averages across everyone a campaign tried to reach. Canvassers knock far more doors than they get answered, so the effect per actual conversation is several times larger.

Two more findings make door conversations worth even more than the headline number:

• The effect spreads inside the home. In households with two registered voters, about 60% of the turnout boost from a canvass conversation passed to the person who did not answer the door (Nickerson, American Political Science Review, 2008).
• The effect carries into the next election. In an experiment with 25,200 registered voters in 1998, people urged to vote by canvassers and mail were also significantly more likely to vote in the November 1999 local elections (Gerber, Green and Shachar, 2003). A voter you turn out in one election is easier to turn out in the next.

Fund the doors first.

Door-to-door canvassing: the GOTV gold standard

Green and Gerber call door-to-door canvassing by enthusiastic volunteers the "gold standard" of voter mobilization (Bustle). The tactic worked across very different places: a coordinated set of experiments before the November 6, 2001 elections ran in Bridgeport, Columbus, Detroit, Minneapolis, Raleigh and St. Paul (Green, Gerber and Nickerson, Journal of Politics, 2003).

What separates a canvass that moves turnout from one that just logs knocks:

1. Real conversations. The one-vote-per-15 figure counts people canvassers spoke with, not doors knocked. A program that leaves literature on every porch and talks to almost nobody is a leaflet drop with extra labor.
2. Clean walk lists. Canvassers should hit your GOTV universe only, sorted into tight turfs so they spend the shift at doors rather than in cars.
3. Data integrity. In that 2001 study, researchers discarded walk lists in Raleigh and Detroit where canvasser contact records looked implausible. Someone has to audit canvass data every night.
4. A short script with a plan question. Remind them of the date, ask how and when they will vote, and get a specific answer. That is where the 4.1 point vote-plan finding comes in.
5. Repeat passes in the final days. Knock again on the people you did not reach, and come back to supporters who said they had not voted yet.

Strong programs use volunteers and paid canvassers together, managed to the same standard. Our guide to paid canvassing costs and doors per hour covers the staffing math, and our canvassing team runs these programs nationally.

GOTV phone banking: conversation quality decides it

Phones work when they sound like people. Early experiments with brief professional calls found no detectable mobilization effect, but David Nickerson's eight volunteer phone experiments in six cities in 2000 and 2001 found that personal, volunteer calls do raise turnout, at an estimated $26 per vote using his assumption of a $15 hourly cost (Nickerson, American Politics Research, 2006).

The commercial phone data makes the same point from the other direction. CIRCLE's compilation reports that a short commercial call took about 400 contacts to produce one new vote (CIRCLE). Volunteer calls, by Green and Gerber's count, need about 35.

Practical rules for a GOTV phone bank:

• Use callers who believe in the race and let them go off script.
• Keep the plan question in every call: day, time, and how they are getting there or returning the ballot.
• Skip robocalls for turnout. CIRCLE's compilation reports no detectable effect from them.

GOTV texting: does it work?

Text reminders have real experimental support. In a field experiment of 8,053 people during the 2006 election, a text sent the day before Election Day produced a statistically significant 3.0 percentage point increase in turnout (Dale and Strauss, American Journal of Political Science, 2009). A follow-up in Norwegian municipal elections found a 2.93 point effect among newly eligible foreign nationals and 4.58 points among young native Norwegians (Political Behavior).

The 2006 study texted new registrants who had given their mobile numbers when they signed up, people who expected to hear from the sender. Collect cell numbers at every registration table and door.

What makes GOTV texts perform:

• Send close to the deadline: the day before, and the morning of Election Day.
• Include the specific polling place or ballot return deadline, not a slogan.
• Use peer-to-peer texting so replies get a human answer ("Where do I vote?" deserves a real reply).
• Ask for the plan by text, too: "What time are you heading to vote tomorrow?"

For per-message pricing and the compliance rules on political texting, see our guide to political direct mail and texting.

The messages that raise turnout: vote plans, social pressure and gratitude

Channel decides reach. Message decides how many points you get out of each contact.

Vote plans

In a 2008 experiment with 287,228 voters, calls that helped people form a voting plan raised turnout 4.1 percentage points among those contacted, while a standard encouragement call had no significant impact (Nickerson and Rogers, Psychological Science, 2010). Among voters in single-eligible-voter households, the effect was 9.1 points; voters in multiple-voter households were unaffected by any script (same study). Put the plan question in every door script, call script and text.

Social pressure

The best-known GOTV experiment mailed 180,002 Michigan households before the 2006 primary. A civic duty reminder raised turnout 1.8 points. A mailer showing households their neighbors' past voting records raised it 8.1 points, about as much as in-person canvassing (J-PAL). Few candidates will sign a mailer listing neighbors' voting histories. The usable lesson is milder: voting is public record, and reminding people that their participation is noticed moves them.

Gratitude

Costas Panagopoulos tested postcards that thanked people for voting in a prior election, across three very different elections (Gerber and Green, Handbook overview). Turnout among thanked prior voters ran 2 to 3 points higher, and on Staten Island the gratitude card raised turnout 2.4 points over voters who got no mail (Fordham University). This is the social pressure lesson in a form any campaign can sign.

Mail, events and digital in a GOTV program

Mail. A standard nonpartisan GOTV mailer moves about one vote per 273 recipients, so mail earns its place through message, not volume: gratitude cards, vote plan cards with the voter's polling place, and ballot return reminders. Do not count your persuasion pieces as GOTV.

Election Day events. Researchers picked 38 precincts in 14 areas across the country and randomly assigned each one to get a polling place festival or not (J-PAL). Across all 38 precincts, the festivals and their advance publicity raised turnout about 2 points (Campaigns & Elections), and the researchers estimate 6.5 points in a precinct where turnout runs around 50% (J-PAL). Some communities in the study had turnout below 10%, which tells you where a festival has the most room to work. Our guide to organizing a community event covers the logistics.

Digital ads. Run paid digital as a reminder layer on the same universe as your doors and texts, and send the traffic to a page with polling places and ballot deadlines. See our political digital advertising guide.

Who to target with get out the vote tactics

GOTV is about turning eligible people into voters, so spend your contacts where turnout is in doubt. Your universe is the middle.

A working GOTV universe usually includes:

• Identified supporters with spotty vote history.
• Newly registered voters, especially ones your own voter registration drive signed up.
• Mail ballot requesters who have not returned a ballot yet, refreshed daily from the absentee file.
• Single-voter households, where the vote plan effect was largest.

Supporters who vote every time belong on your volunteer list instead. See our guide to recruiting and managing volunteers.

How to calculate cost per vote and test your own GOTV program

Cost per vote settles budget arguments. The formula:

1. Cost per contact = total program cost divided by completed conversations (or deliveries, for mail and text).
2. Contacts per vote = taken from the experimental record above (15 for doors, 35 for volunteer phones, 273 for standard mail).
3. Cost per vote = cost per contact multiplied by contacts per vote.

An illustration, not a quote: if a canvass costs you $5 per completed conversation, at one vote per 15 conversations you are paying about $75 per additional vote, before counting the 60% household spillover. Run the same math on every line in your GOTV budget and the priorities sort themselves. Our political campaign budget guide has the line items.

To measure your own results, hold out a random slice of your target list and contact everyone else. After the election, pull vote history from the voter file and compare turnout in the two groups. It is the same method every study in this guide used, applied to your district.

Frequently asked questions

What get out the vote tactics are most successful?

Door-to-door canvassing has the strongest record, at about one additional vote for every 15 voters contacted. Volunteer phone calls follow at roughly one vote per 35 contacts. Adding a vote plan question to those conversations raised turnout 4.1 points among people reached in a 287,228-person experiment. Robocalls and generic mail rank at the bottom.

What are get out the vote strategies that a party or candidate might use?

A party or candidate typically runs a door canvass of supporters and low-propensity voters, volunteer and paid phone banks, peer-to-peer text reminders, mail ballot chase, gratitude or vote plan mail, and Election Day events or rides to the polls. The experimental evidence says to weight the plan toward live conversations and treat mail and digital as reinforcement.

Does door knocking actually increase voter turnout?

Yes. Door-to-door canvassing is one of the most tested GOTV tactics, starting with Gerber and Green's 1998 New Haven experiment, and it raised turnout again in six cities in 2001. The effect also spreads to other voters in the household and carries into later elections.

Do GOTV text messages work?

Yes. A 2006 experiment with 8,053 people found a text reminder sent the day before the election raised turnout 3.0 percentage points, and a Norwegian follow-up found effects of 2.93 to 4.58 points in the groups it tested. Texts work best as a timely reminder with specific polling place or deadline information, sent to people who gave you their number.

Do robocalls increase voter turnout?

No detectable effect has been found in GOTV experiments with robocalls, according to CIRCLE's research compilation. Spend that money on live callers or on doors.

How many contacts does it take to get one vote?

Based on Green and Gerber's summary of the experiments: about 15 door conversations, about 35 volunteer phone contacts, or about 273 recipients of a standard nonpartisan mailer. Short commercial phone scripts needed about 400 contacts per vote in studies compiled by CIRCLE.

What is the difference between GOTV and persuasion?

Persuasion tries to change which candidate or position a voter supports. GOTV gets people who already support you, or who are eligible and unlikely to vote, to cast a ballot. Each needs its own list and script.

When should GOTV start?

Start building the GOTV universe, volunteer base and canvass staff months out, and start contact when voting starts: as soon as mail ballots go out or early voting opens in your state, not the final weekend alone. The final days still matter most for reminders and repeat door passes. Our GOTV planning guide covers the calendar.

How Leo Operations approaches this

Leo Operations is a national field, media and creative firm, and GOTV is a core field service. Our field programs have knocked more than 650,000 doors, and we have run campaign, ballot and organizing work in 25 states for more than 40 campaigns, officials and causes.

Founder Rodriquez Wright II has worked campaigns since 2012. He ran regional GOTV for Hillary Clinton in the 2016 New Hampshire primary, coordinated canvass programs in Las Vegas and Chicago that year, directed a statewide PAC voter-contact program in Tennessee in 2018, and served as Tennessee State Organizing Director for Mike Bloomberg's 2020 presidential campaign.

Our field programs cover paid canvassing, voter registration and GOTV. Leo also produces political media, campaign ads, video, websites and digital strategy, so the same firm can handle the doors and the reminder content.

If you are planning a GOTV program, email us your race, your turnout target and your budget, and we will tell you what it takes to hit the number.