How Do We Get More Democrats to Vote?
Part 1: Canvassing

One of Many Conversations

Andrew is a lifelong resident of Waterloo, Iowa, so the streets he walked in the five months leading up to the 2016 election were familiar.

On an early summer day, he knocked on the door of Laura, an elderly woman. State records indicated she voted inconsistently in recent years, casting a ballot in 2010 and 2012 but not 2014. Answering the door, Laura looked surprised to see someone here to talk with her so far in advance of the election. Introducing himself, Andrew began asking about her voting habits. Did she believe voting was important? What kind of issues had motivated her to vote in the past?

After ten or so minutes of conversation about the issues facing their town and the importance of standing up for their community, Andrew pulled out a Commit to Vote card and asked her to sign her name committing to vote that year. The woman hesitated, telling Andrew she wasn’t planning on voting this year. A little confused, he gently pushed back, asking why. She told him she couldn’t drive anymore and had no way of getting to the polls.

Andrew and his fellow Iowa staff ready to knock doors

Andrew offered the option to vote by mail, but the woman said she was unable to fill out the mail ballot request form due to broken reading glasses and poor eyesight; Andrew offered to help her fill out the request form. They completed the form together as well as a Commit to Vote card, which would be mailed back to her when vote by mail ballots went out in the fall.

As Andrew left her house, Laura put her hand on his shoulder and thanked him. He told her he’d be back to speak with her in a few weeks and would give her a call before Election Day. Heading down the driveway, he picked up his pace. He had a long list of other voters in her neighborhood with a goal of at least 100 similarly in-depth conversations that week.

Andrew’s ten other team members were fanned out across Iowa’s 1st Congressional District engaged in similar conversations. That night they would all meet up to talk about the encounters they had and write personalized thank you notes to the voters they spoke with that day.

The Turnout Problem

As members of the first ever Progressive Turnout Project field team, Andrew and his fellow Field Representatives were part of our effort to change the way political organizations communicate with voters and boost voter turnout. Founded in the aftermath of the 2014 election, our organization is focused on finding new solutions to one of the most basic and debilitating flaws of the Democratic Party: low voter turnout. We design, test and execute voter turnout programs to engage voters who are often excluded from the political process.

Democrats have struggled to get voters like Laura to the polls for years. In 2014 turnout was the lowest it had been in 70 years, and Democrats suffered crushing losses around the country as a result. Turnout was just 36.7% nationally but far lower among key Democratic constituencies: people of low-income, minorities, and young people. More than 44 million eligible voters of color did not vote, and 66 million eligible voters earning less than $50,000 did not vote.

Andrew’s home district, Iowa’s 1st Congressional District (IA-01), is a perfect example of this phenomenon. Widely considered one of the most competitive districts in the county, IA-01 had been represented by popular Democrat Bruce Braley since 2006. However, when Braley decided to run for Senate in 2014, the seat opened up, and one of the most heated races in the country ensued. Low turnout across the state, especially among low income and minority voters, helped lead Tea Party Republican Rob Blum to a narrow victory.

Inconsistent voters are politically more liberal than the general electorate. Thus, when turnout is high, the electorate is more liberal; when turnout is low, the electorate is more conservative. Unless we can learn to get voters like Laura to the polls, we’re not going to hold a majority in Congress anytime soon.

Low turnout isn’t a new problem for our party, but it’s one we struggle to effectively address. It’s not a lack of awareness. Anyone who follows politics has heard the slew of pundits and party officials bemoaning low turnout after a tough election cycle. Instead we would argue it’s a lack of knowledge–we’re just not sure what to do to change voting behavior.

Many inconsistent voters are more mobile than consistent voters, and therefore less involved in community and social networks that encourage voting. Their mobility makes it harder for campaigns to contact them. Many of them are disenfranchised by overly restrictive voting laws or unable to go to the polls because they work multiple jobs.

Coming out of 2014 we saw a need for a long-term investment in mobilizing inconsistently voting Democratic voters and in research around turnout. Thus, Progressive Turnout Project was born. The swell of grassroots support we saw immediately confirmed our belief that this was something the progressive community was hungry for. In the 2016 campaign cycle, that support allowed us to run programs in 19 congressional districts, experimenting with everything from canvassing to digital ads to mail.

Designing Our Program

When we sat down to craft programs that would engage and activate inconsistent Democratic voters, we were guided by two main principles. First: a strong belief that voters need to be engaged in a serious, respectful and authentic manner. Second: a belief that research, data and analytics can shed light on human behavior and enhance our work. There are mountains of behavioral science research out there about how to change people’s minds and influence their behavior. We wanted to take all of that knowledge and utilize it in comprehensive programs whose effect we could measure. Fortunately, as we dove into the research, it became increasingly clear that those two principles go neatly hand in hand.

Starting in the early 2000s, political scientists have been evaluating campaign tactics by running experiments that measure and compare the cost per additional vote generated. Some of that research has made it into the practice of politics; much of it has not. The most famous example of that work is Get Out the Vote, a collection of experiments and insights compiled by Don Green and Alan Gerber.  Researchers across the country have replicated and built on their work, developing an extensive literature on the relative efficacy of different methods of communicating with voters.

Those decades of research have demonstrated what many of us know intuitively to be true: an in-depth, face-to-face conversation with someone trusted, rooted in values, is the best way to persuade someone to take an action. The closer you can get to replicating that experience, the more successful you’re going to be.

We were also inspired by research done by behavioral scientists like Richard Thaler, Cass Sunstein and Robert Cialdini focused on “nudges.” Humans aren’t perfectly rational creatures, and sometimes we struggle to actually do things that we know are good for us. “Nudges” are subtle interventions that take advantage of the irrational way we sometimes process information to help us make good choices. Research on using subtle interventions to get people to stop littering or buy slightly healthier options at the grocery store covers techniques that are applicable to voting as well. Some of the key insights from that research included:

  • An Iowa voter with our materials emphasizing identity as a voter

    Self-prediction and commitment: Asking someone to predict whether or not they are likely to perform an action increasestheir likelihood of eventually doing it. We incorporated this by asking people to predict whether they would vote or not.

  • Inducing feelings of accountability: Feelings of accountability are induced by leading people to believe they will be forced to justify their behavior. When you make it clear you will hold them accountable or that their behavior will be public, they are much more likely to behave in socially desirable ways. This is why we make people feel like someone is checking to see if they vote by showing someone their own vote history and/or letting them know we will follow up with them on their participation after the election.
  • Expression of identity: Reframe the desired action to align with an identity that someone might highly value. For example, we would ask someone about reasons why they vote such as, “I care about setting a good example” or “I care about my family’s future.”

Our goal was to put those individual insights into action in a comprehensive, long-term program.

Throughout spring 2016, we laid out guidelines for hiring motivated and local staff members, carefully wrote scripts that encouraged long, in-depth conversations and laid out plans for multiple repeated contacts to encourage commitment, planning and follow through.

We also spent time designing how to execute a randomized control trial experiment (RTC), similar to the trials drug companies use to see if a drug is effective or not, in all of our field work. To do this, we take all the voters that fit our profile and randomly assign some to be contacted and some not to be contacted. We then compare the behavior of the two groups.

For our program in Iowa’s 1st, we found a total of 90,000 voters that fit our profile. We designated about 20,000 voters to the control group and did not communicate with them at all. That left about 70,000 voters for the “treatment” group—voters that staff would work to reach at their homes.

A few months after the election when the state releases data on who did and did not vote, we compare the rates of turnout between people in our treatment group and people in our control group to show the effectiveness of our work.

Our Research Director Claire Rogerson summed up the importance of testing to the organization: “With RTCs, you know the effect of your work. Campaigns sometimes don’t know if what they did was effective — if that’s the case, you’re always guessing and never moving forward. With RTCs, you can start to compare different methods of contacting voters and start making better choices about the ways you work.”

Getting Off the Ground in Iowa

As we looked for a district to pilot our program, Iowa’s First Congressional District emerged as an ideal testing ground. The race was expected to be close, and Democrats nominated a candidate with impressive experience and strong local ties. But even more crucially, the district had a large percentage of Democrats who were conspicuously absent during midterm elections and were at risk of not voting in 2016. Using state records and demographic data, we identified more than 90,000 voters who fit that profile.

In March of 2016, we took the plunge and hired our first District Director, Brett, to execute our program. Based in Cedar Rapids, Brett was tasked with hiring local staff who could effectively communicate with their friends and neighbors.

Potential hires came from a variety of places and a wide range of backgrounds. Brett met Rachel at a resale furniture shop in search of affordable office furniture. Struck by how persuasive, outgoing, and personable she was as a salesperson, Brett told her about his work and asked if she’d consider working in politics. Rachel ended up joining our team and spent her time talking to voters in her hometown of Marion.

Lifelong Iowan Matt was preparing for medical school in Dubuque. While studying for the MCAT, he joined the staff and talked to his neighbors in Dubuque about the election and his passion for affordable healthcare.

Angie had lived in Cedar Rapids for decades, and her political activism stretches back almost as long. The 2016 election was so important to her that she took a break from 25 years in the field of dentistry to work as part of our team full time.

“Being from the community, our Iowa staffers were really proud of the work they were doing day-to-day. These were people who had lived in northeast Iowa for a long time, and they really wanted to work to elect Democrats to represent them,” said Alex Morgan, PTP’s Executive Director.

Before heading out into the field in May, the team underwent an intensive training in our Cedar Rapids office with Brett. They carefully reviewed both the state of the race in Iowa and the candidates’ positions, but also went over all the research showing that deep conversations and multiple contacts had proven to be effective. They practiced their conversations over and over again with each other in front of the group, giving each other feedback on how to overcome nervousness, dive right into conversation, and make a strong ask for a commitment to vote at the end of the conversation.

At the end of the training, each representative was assigned an area of the district with approximately 5,000 voters. They would speak with each of their voters three times before the election, twice at their door and once on the phone in the days immediately leading up to the election. Most canvassers were assigned to their own neighborhood so they could communicate directly with people they knew. Many of them already knew plenty of people on their list.

Our Program

If you were a voter like Laura in Andrew’s turf, you would receive as many as seven communications from him from late April up to the election. See the below timeline below and click each event for detailed information about each piece of the program. Click the image to see our materials in more detail.

Making connections at the door wasn’t always easy. For every positive interaction like Andrew’s with Laura, there were a slew of knocks that resulted in no answer. Every now and then, our staff would meet someone who would stubbornly refuse to engage with them or slam the door in their face. Further, many of our voters are hard to find. Our universe was comprised in large part of people of low income, who tend to be very mobile. It was common to knock on a door and find that the family living there had moved months ago.

Losing the Election

On the night of November 8th when Monica Vernon lost, our team was devastated. Together they knocked more than 135,000 doors, had thousands of conversations, and secured thousands of commitments to vote. It was a tough loss to swallow, especially for our Field Representatives. Like Democrats across the country, it left all of us wondering if we could have done something differently or if we had taken the wrong approach.

Fortunately, due to the scientific nature of our work, we were able to make an unbiased assessment of our work. Often on a campaign, it’s easy to think you did everything right if you won or everything wrong if you lost. But our team is able to take away impartial lessons and draw on those insights in our future work.

“We started this organization because we felt that knocking on doors and having face-to-face conversations was the most effective way to talk voters. But we also wanted to make sure that was true and adjust course in the future if what we were doing wasn’t working,” said Alex, our Executive Director.

In Iowa, our field representatives knocked on more than 135,000 doors. The voters we spoke with at the doors turned out at a rate 12.1% higher than those in our control group, 80.24% voter turnout to 68.14% respectively. We saw extremely similar results across our districts, with double digit jumps in turnout across the board. That’s an incredible jump, resulting in thousands of votes generated across the district.

We got valuable insights about what kind of people we can reach at the doors and on the phone, as well as what kind of people we can persuade to vote. This will help us refine our targeting in the future.

Looking to the future, our work feels more critical than ever. If we’re going to defeat Donald Trump in 2020 and take back the house in 2018, it’s going to have to be with a more energized and mobilized base than ever before. We’re going to need real information about what does and doesn’t work.

Our program proved to be extremely effective at mobilizing people. We hope to encourage other organizations to adopt our methods where applicable and invest in on-the-ground efforts earlier than ever before.

But there is always room to improve. Moving into next cycle we’re focused on refining our audience, looking for new and innovative ways to make sure we have the right addresses for our voters, and screening out people we won’t be able to persuade to vote. We’re also looking for new, more creative ways to reach those voters. We’ve brought on talented new staff and developed partnerships with leading researchers at Northwestern University and the University of Chicago to pursue new tactics and analyze our data from last cycle.

We’re excited to be on the ground in Georgia, Montana, and South Carolina and preparing to launch a program in the Virginia legislative elections.

Bookmark the Progressive Turnout Project blog, like us on Facebook, or follow us on Twitter so you can be among the first to see our upcoming posts about additional results from 2016!