How Do We Get More Democrats to Vote?
Part 3: Mail
Americans have relied on the postal service to communicate with their neighbors since the late 1700s. Today in 2017, political campaigns from village council to president lean on this same system as a critical part of their outreach to voters.
A Long Body of Research
Though political mail has been a constant for many decades, the message of political mail is constantly evolving. In more recent years, the popularity of “social pressure” mail has risen.
Campaigns on both sides of the aisle have used social pressure mail to reach out to voters they think are likely to support their preferred candidate. Rather than containing an argument for why a person should vote for a certain candidate, this mail contains several pieces of information intended to get the receiving individual to vote – period.
Social pressure mail generally includes some or all of these elements:
- A record of the individuals’ own voting history
- Information about how the individual’s voting history compares to their neighbors
- A reminder that voting is a matter of public record
- Notification that a group or person will follow-up with the individual post-election on whether they voted or not
A number of studies on social pressure mail have found that it’s an effective means of increasing voter turnout. (For more information on “social pressure” methods, see our post on our 2016 canvassing results. This piece dives into the theory of inducing feelings of accountability in more detail.)
Our organization, Progressive Turnout Project, was founded to get more Democrats to vote, with a particular focus on getting inconsistent voting Democrats to the polls by executing randomized control trial experiments to assess the most effective turnout methods. Hence, we were interested in testing the effectiveness of social pressure mail for ourselves and hopeful that we’d find similar results. So, in the 2016 cycle, our organization set out to design and test a mail program using research-proven strategies found to be effective elsewhere.
Our Mail Program
To us, a particular strength of mail is the ability to personalize it to each target voter. We aren’t able to change a canvasser’s script at each door or make a digital ad that’s unique to each person, but we can send mail that contains an individual’s sample ballot and unique voting history.
As we designed our mail, we capitalized on this ability to specialize content. We designed our mail in-house, drawing on decades of research on political mail and in particular. As we do in all our field programs, we designed our mail program as a randomized control-trial (RCT) experiment so we could draw conclusions based on unbiased results.
In an RCT, we take all the voters that fit our profile (likely Democrats who are inconsistent voters in targeted districts) and randomly assign some to be contacted, in this case by mail, and some not to be contacted. We then compare the voting behavior of these two groups. By comparing the voting rates of people we mailed to the voting rates of people in our control group who didn’t receive mail from us, we assess the impact of our program.
Across 16 highly competitive congressional districts, we sent our target voters three pieces of mail using social pressure tactics to frame voting as part of a positive self-identity. In the 2016 cycle, our organization sent more than 2 million pieces of mail to inconsistent Democrats with the goal of getting them to vote in the November elections.
The first mailing came as a letter in an envelope. The letter contained each person’s individual vote history, informing the target that he or she was part of a study on voting behavior. The letter indicated that our organization would follow up with the individual before and after the election with an updated voting history.
The second mailing, which individuals received a few days after the first letter, came in a postcard that stressed that turnout was expected to be high in the individual’s town. The postcard also included a sample ballot and a sample vote plan.
The third mailing arrived a few days before Election Day. This final mailing was a postcard that had a simple and direct social pressure message. It said “You’re a role model, you’re also an infrequent voter,” and included a peel off “I Voted” sticker.
We knew people were having strong reactions to our mailers that included their individual voting history records, because they started calling and emailing us! We heard from hundreds of people upset that we knew they had missed voting in recent past elections and often explaining the extenuating circumstances that had made them miss going to the polls.
This was a bit of a stressful time at our organization due to the influx of calls and emails on the topic. But we were also hopeful that the heated feelings about our mailings meant they were having an impact!
It is not uncommon for social pressure mailings to elicit this kind of reaction. Many other campaigns and groups that use this tool have similar experiences but decide to continue sending such mailers because of their effectiveness.
The potential negative reaction of voters is a critical reason that candidate committees tend to not send social pressure mailings, despite the research showing that they work. A candidate is understandably reluctant to anger potential voters in this way.
What We Learned
When we pulled our experiment results post-election, we found that people who received our mail did not vote at a significantly higher rate than similar people who did not receive our mail.
We had painstakingly created our mail based on reams of scientific research. What went wrong?
There are many possible answers.
First, the 2016 election was unique. The presidential election dominated coverage, and extraordinary events and personalities drove the presidential campaign. We theorized that the 2016 election was a particularly high-salience election. Perhaps in an environment when people are so saturated with electoral information, social pressure mail has a limited effect.
As our then Data Director Claire Rogerson said, “We can’t remove ourselves from the messy reality. I think we sent great mail. It was spot on. But context matters.”
As two researchers wrote after finding mixed results on a 2010 social pressure mailing experiment, “Our knowledge remains incomplete concerning which under which conditions these [GOTV] mechanisms work.”
Secondly, it is possible that sending three pieces of mail is not enough to change a voter’s behavior. Perhaps a voter needs to see six or eight or even more pieces before they internalize the message!
Finally, finding out that a certain method of reaching voters is ineffective is NOT a total loss by any means. Mail is an expensive means of communication. If we continually find no results from our mail program, we will stop sending mail and reallocate those resources on means of communication that are producing results.
In the end, we aren’t studying the impact of our voter outreach work in a clinical laboratory. We will have to test strategies and methods multiple times over several different election cycles before finalizing decisions about what works and what doesn’t work. That is a frustrating reality.
 Alan Gerber, Donald Green, and Christopher Larimer. “Social Pressure and Voter Turnout: Evidence from a Large-scale Field Experiment.” American Political Science Review. February 2008. https://isps.yale.edu/sites/default/files/publication/2012/12/ISPS08-001.pdf.
 Greg Murray and Richard Matland. “Mobilization Effects Using Mail: Social Pressure, Descriptive Norms, and Timing.” Political Research Quarterly. 2014 Vol67(2) 304-319. https://www.jstor.org/stable/24371788?seq=1#page_scan_tab_contents