Since the 2016 election, there’s been a sprawling debate within the left over whether Democrats should abandon “identity politics” and focus instead on a forward-leaning message centered on economic populism. (Social identity is a central influence on everyone’s political choices, so when we talk about “identity politics” in this context, we really mean identity politics that don’t appeal to straight white people.) The idea is that focusing on issues of ethnicity, immigration status, gender, and sexual orientation risk alienating more of the non-college-educated whites who have moved toward the Republican Party since Obama’s election. There’s also a significant body of data suggesting that racial animus is a key driver of white conservatives’ opposition to redistributive policies and a robust safety net. Implicit in the argument is that people of color and the LGBTQ community would continue to support Democrats if they spent less time focusing on the unique concerns of those groups, both because progressive economic policies should appeal to their interests, in most cases, and because the GOP continues to signal that its “tent” isn’t big enough to welcome them.

But new data suggest that the entire debate poses a false choice, at least as it applies to racial-justice issues. Preliminary findings from Anat Shenker-Osorio (author of Don’t Buy It: The Trouble with Talking Nonsense about the Economy), Ian Haney López (author of Dog Whistle Politics: How Coded Racial Appeals Have Reinvented Racism and Wrecked the Middle Class), and researchers at the progressive think tank Demos show that a unified message that acknowledges both racial and economic disparities wins over white voters more effectively than an economic message alone. Just as significantly, the researchers found that while race-neutral economic arguments won’t cause voters of color to embrace the GOP, they do make them less enthusiastic about participating in elections. They say that more research on this is coming down the pike.

Shenker-Osorio told me that the “race-class narrative” she and her colleagues are developing “draws a causal link between issues of race and racism and the extreme and growing class inequities that are of course more acute for communities of color.” The project’s goal is to determine whether a message “that combines issues of race and class [can] actually beat both the opposition’s message—which is always the name of the game—and also the kind of color-blind, economic populist message that we’ve come to accept as a progressive standard.”

The results, according to Shenker-Osorio, were compelling. “Every single message that we tested that gave explicit mention of race” performed better with what the researchers defined as the progressive base and was “also more persuasive to the 59 percent of people who fell not in our base, and not in the opposition’s, but in that coveted middle.”

There have been several phases of studies so far. They included a series of surveys fielded in conjunction with Lake Research Partners (a political polling and strategy firm that works extensively with Democrats and organized labor), and a door-to-door canvas of 800 voters in Minnesota—half of them white and half people of color—that was conducted in partnership with a coalition of grassroots groups called Our Minnesota Future.

“Persuadable” white voters, as the studies call them, tend to be torn between two competing narratives: On one hand, these voters said that it’s important to talk about race, and that they care about racial equity. But the same voters also perceived conversations about race as being difficult in ways that tend to pull them “toward racial resentment and conservative fears.” In races against politicians who use racially coded dog whistles to fire up their base, one or the other of these views typically gets “primed,” or activated, by campaign rhetoric, and a race-neutral economic message cedes that contest to the kind of appeals to racial grievance that have long been a mainstay of conservative messaging.

A key drawback with economic messages alone, according to Shenker-Osorio, is that they attempt to counter fears situated deep in our reptilian brains with promises of greater economic security, and that tends to be a losing proposition. “When someone knocks on your door and says that crime is spiraling out of control and MS-13 is coming to kill you,” she said, “and then some Democrat knocks on your door and offers free college, the reaction is, ‘Sure, free college is great!’ But the problem is that you were just told you were going to be killed.”

Shenker-Osorio said that it’s vital to “help white people understand why it is that these severe racial disparities exist, rather than just naming the disparities, which leaves people to fill in the causal relationship for themselves.” When voters are left to draw their own conclusions about why people of color have disparate economic outcomes, they’re more likely to embrace conservative narratives that hold that poor work habits or stereotypical cultural flaws are responsible. Naming a culprit—in this case, politicians who use racial animus to divide working America—creates a sense of solidarity with people of color, rather than stoking resentment toward them. “Perhaps the most interesting finding from our research,” said Shenker-Osorio, “is that explicit references to race actually bolster economic populism.”

The point of the canvas was to test how these competing messages influenced voters’ preferences at the polls. All 800 households were shown a real flyer from a Republican congressional candidate, with the candidate’s name and partisan affiliation removed. It bashed Democrats for, among other things, “demanding more sanctuary cities for criminal and illegal aliens.” After viewing the flyer, a majority of white respondents agreed with the sentiment (as did a plurality of people of color).

Half of the respondents were then shown a flyer from an imaginary candidate that contained a progressive, race-neutral economic message, while the other half were offered one with a message that incorporated both race and class. The latter read, “Whether white, black, or brown, 5th generation or newcomer, we all want to build a better future for our children. My opponent says some families have value, while others don’t count. He wants to pit us against each other in order to gain power for himself and kickbacks for his donors.”

After viewing the GOP flyer and one of the two progressive messages, respondents were asked which candidate they were likely to support in an election. Among whites who agreed with the first flyer and were then shown the race-neutral economic flyer, 56 percent stuck with the Republican, while 44 percent shifted to the progressive candidate. But among those shown the race-class message, the numbers basically flipped—with 43 percent choosing the conservative candidate, and 57 percent switching to the progressive.

People of color were also asked which candidate they’d support, but they were offered a third option: just staying home on Election Day. More than six in 10 who were receptive to the initial dog-whistle flyer said they’d support the progressive candidate after seeing the one that incorporated both race and class—significantly higher than among those shown the race-neutral economic message (just under half switched to the progressive).

But perhaps more importantly, in terms of electoral success, the purely economic message led twice as many people of color to say that they’d simply sit out the contest than did the race-class narrative. (Among those who agreed with the Republican flyer, 8 percent of the group that then saw the race-class message said they were unlikely to vote, compared with 15 percent of those who saw the pure economic message; among those who disagreed with the first flyer, those numbers were 8 percent and 18 percent, respectively.)

The researchers think that this is likely because people of color navigate an environment in which they see structural racism all around them—they aren’t able to choose to ignore “identity politics” because they’re affected by those issues on a daily basis—and as a result they perceive race-neutral economic messages as not really being about them, or not speaking directly to them.

These findings are especially important in the Trump era. As the 2018 election cycle heats up in earnest, Republicans, who continue to see divisive stereotyping and race baiting as a winning strategy, are sticking to their dog whistles in hopes of turning out their base in November. The research from Shenker-Osorio, Haney López, and Demos suggests that Democrats would be wise to alter theirs.

Correction: This article has been edited to reflect he latest tallies of “persuadable” middle voters. They make up 59 percent of those surveyed; not 63 percent, as had previously been reported. We apologize to our readers for the error.