Even after success in 2018, many progressives remain convinced that winning over people who voted for Donald Trump is impossible and that trying is a waste of time. The myth that he could shoot someone on New York’s 5th Avenue and not lose any support has bipartisan appeal.
Data, however, tells a different story. According to 2018 exit polls, more than 3.5 million, or 8 percent of people who voted for Trump in 2016 voted for a Democratic candidate for the U.S. House this year.
This probably understates the real number—exit polls have had methodological challenges and many people who defected from Trump are unwilling to admit that they voted for him in the first place. But it’s clear that persuasion is alive and well in American politics.
We saw it in action in elections across the country and we’re developing new tools and techniques to study it. In 2018, our firm, Clarity Campaign Labs, worked with candidates in races big and small to identify the people to engage with based on experiments that allowed us to see which voters actually changed their minds when presented with certain information. Then we scaled that analysis and created statistical models for the national electorate.
Our findings aren’t what you would expect.
One notable finding is that many partisan voters can be persuaded to vote for a different candidate. Nationally, we could move 1-out-of-every-30 voters to change their congressional vote with a single message reminding them of Congress’s power to be a check on Trump.
As is often the case on hot-button issues, that movement went in both directions and induced backlash as well as support. There were an almost equal number of people that moved to the Democrat as there were that moved to the Republican upon hearing anti-Trump messages.
But the anti-Trump message isn’t what won the day. We also found that using a different message reminding voters about healthcare issues and the GOP’s plan to cut protections for people with pre-existing conditions worked even better. Hearing that message just once, we could move 1-out-of-every-20 voters to change their congressional vote. And, unlike the Trump message, almost all of the movement was towards the Democrat, with very little backlash.
Healthcare, rather than opposition to Trump, proved pivotal for the 2018 blue wave, which won the Democrats a net 40 seats and control of the House. Our methods didn’t just tell us what message worked best, but what voters to target. Despite the backlash, we could identify a universe (almost 20-percent of the country) that still moved Democratic with the Trump message at a staggering rate of six times greater than that of the average voter.
This enabled us to help specific campaigns talk to the right voters, using the right message, and through the right medium. For healthcare messaging, we could identify groups of voters in which 1 out of every 5 we talked to would vote for the Democratic candidate instead.
Here’s how we did it: We wanted to better understand which voters are actually persuadable, so we set up experiments with a treatment group and control group. Overall, we used large sample, multi-mode surveys—automated landline, live landline and cellphone, and online panels—that went out to more than 30,000 registered voters across all the treatment and control groups. The partisan split of the surveys we did was self-ID 40 percent Democratic, 40 percent Republican, and 20 percent Independent.
Half of the sample received a message before we asked how they would vote while the control track was simply given a straightforward ballot question. This randomization lets us see the actual impact of messages on vote choice.
In ballot initiatives we worked on, such as Amendment 4 in Florida to restore the right to vote to Floridians with a prior felony conviction, we saw up to 30 percent movement in response to messaging. In partisan races, such as the competitive districts we assisted the DCCC with, we saw around 5 percent actual movement. Not people who said they might change their minds—people who really did.
We tested multiple messages using this approach and found that there are many different persuadable audiences for different messages. There isn’t one master message that moves everyone. Some folks responded to the Trump message but didn’t respond to messages about healthcare. Democrats won across the country because we had candidates who could speak to voters authentically and tailor their appeals to the voters who needed to hear them most.
Voter turnout was a key piece of Democratic wins across the country and the organizers and candidates who powered the record-setting midterm Democratic turnout deserve all the kudos they’re getting and more. But Democrats should also take a lot of comfort from the 2018 House races.
Sometimes the message may not be the one that excites our base the most. But when we have conversations with Republican voters about issues they care about, we can still convince many of them to join us.
David Radloff, John Hagner, and Dan Castleman are partners at Clarity Campaign Labs, a data and analytics firm that works with Democratic campaigns.