For all the campaigning that voters endure every election season — ads on the TV, cheerful volunteers on the front stoop, dozens of slick flyers piling up in the mailbox — few of us will actually have a change of heart about a candidate or a cause. For the most part, we leave election season with basically the same opinions we had going in. And that’s why, two years ago, a couple of researchers named David Broockman and Joshua Kalla made a huge splash when they published research confirming that a political campaign could change people’s minds on polarizing issues like the civil rights of transgender Americans.
You’d think that news might be of use to political campaigns. Yet the technique the researchers studied isn’t getting deployed much in the current election. Experts say that’s indicative of a wider disconnect between the academics who study politics and the people who do the political campaigning. Even at a time when campaigns are increasingly interested in internal data analytics and using the tools of science, it’s not easy to put published research to use in the real world.
The technique that Broockman and Kalla studied was called “deep canvassing” — a more complex and psychologically driven cousin of the door-knocking campaigns that send volunteers and candidates tromping through neighborhoods every election season. Instead of short, scripted conversations where canvassers essentially tell voters what they should believe, deep canvassing involves asking questions and listening empathetically to answers. The average interaction in Broockman and Kalla’s study was 10 minutes long.
It did seem to work, though. Using a tool called a “feeling thermometer,” which quantifies emotions by assigning them a point value, deep canvassers in the Broockman-Kalla study were able to increase positive feelings about transgender people by an average of 10 points on a 100-point scale. More amazingly, that increase in positive feelings persisted over three months. The success brought media attention and stirred excitement in the political realm. There was even a TEDx Talk.
For the most part, though, the canvassers who show up at your door this election season will still be canvassing in the shallows. Deep canvassing’s success in his study hasn’t translated into it getting widespread use on the campaign trail, Kalla told me.
That’s partly because this is still a new technique, and it’s not totally clear yet when it will work and when it won’t. A follow-up study of the deep canvassing tactic, this one aimed at changing opinions on abortion instead of transgender rights, ended with subjects’ minds basically unchanged. Kalla’s team is currently analyzing data from a second attempt at applying deep canvassing to abortion rights advocacy.
But politicians’ reluctance to implement deep canvassing is also part of a bigger trend. According to both academics and political campaign consultants, it’s not easy to take the results of scientific research and use it to shape a political campaign. “There are some points of diffusion for social science research into the political space, but by no means is it broad or widespread,” said Daniel Kreiss, a media professor at the University of North Carolina, Chapel Hill, who studies political communication and campaign tactics.
There are a couple of key problems. First, some of the techniques that seem promising in research would be complex and expensive to execute on the street. That includes deep canvassing, which requires specialized training and skilled volunteers to execute properly. Matt Morrison is executive director of Working America, a professional canvassing organization that is using some of the aspects of Broockman and Kalla’s research to advocate for Medicare expansion in Virginia. He told me that his group was able to get positive increases in support that persisted — and even grew — four months after contact. But he also said that it wouldn’t be as easy for a candidate’s campaign to replicate that success.
That’s because candidates’ campaigns are, by their nature, temporary. “They’re starting up, ramping down. Starting up, ramping down,” Morrison said. Because Working America is an issue-advocacy group that will be focused on the same topics over multiple election cycles, it can train staff once and use their skills for years; a candidate’s campaign, on the other hand, has to start fresh each time, assembling and training groups of volunteers every cycle. And that makes deep canvassing more cost-prohibitive for candidates than it is for Working America. The whole thing becomes even more difficult for candidate campaigns when you consider that face-to-face interaction seems to work best when it’s done by people who are members of the communities you’re trying to reach, said Sarah Niebler, a professor of political science at Dickinson College in Pennsylvania.
In other words, the techniques that seem to work best to mobilize voters are the ones that are least accessible to a candidate’s campaign. Melissa Michelson, a professor of political science at California’s Menlo College, agreed. She felt frustrated, she told me, that she couldn’t get campaigns to use research she’d done showing how best to engage low-income communities of color in get-out-the-vote campaigns. But that research also showed that those efforts worked best when the organizations involved had a continual presence in the community rather than appearing in election years and disappearing in off years. And that, she said, “is not necessarily something a candidate or campaign has the resources to implement.”
The other big problem with translating political science research to campaigns is that the research keeps telling us to do things that the business of campaigning doesn’t reward.
For instance, back in the 1990s and 2000s, research began to emerge showing that television advertising probably wasn’t worth the investment, on a votes-per-dollar basis, said Hahrie Han, a professor of political science at the University of California, Santa Barbara. But in Han’s experience, the campaign managers who pitched evidence-based services that avoided TV ads were usually campaign managers who didn’t get the job.
And most campaign managers wouldn’t want to do something that deviated so much from the norm to begin with, said Addisu Demissie, co-founder of the campaign consulting firm 50+1 Strategies and manager of Lt. Gov. Gavin Newsom’s current bid for governor of California. That’s especially true for a regional or state-level campaign that probably only has the resources to do either the new thing or the old — not both. “Would you roll the dice on someone’s political future that way?” he said.
That’s not to say that political campaigns don’t use some lessons from political science. For instance, if you’ve gotten a text message this year from a campaign, you’ve seen some of Michelson’s research in action. She’s one of the people whose work showed that text messages are a more effective way to communicate with voters than mail or email. And when Sen. Ted Cruz’s 2016 presidential campaign sent a series of ill-fated letters to Iowa voters accusing them of “voting violations” for failure to turn out in past primaries, his campaign was using a (poorly executed) version of an idea put forward by political scientists.
But overall, political scientists said, the election season would look a lot different if more of their advice were followed. “You’d see less indirect outreach through your TV, through your radio, through your mailbox, through your screens,” Michelson said. Which may sound great, until you consider the alternative. In an evidence-based election season, she told me, you might have fewer campaign ads clamoring for your attention in every form of media you encounter, but you’d be hearing a lot more campaign ads from your friends and family. Anyone with a Facebook account knows how that can feel.