Reporters often describe President Donald Trump’s racism as a strategy to excite Republican voters, but it may actually work better on certain self-identified liberals.
That’s the conclusion a pair of academic researchers reached after conducting an experiment with hundreds of white Americans earlier this year.
A subset of liberals, not conservatives, could be the group “most responsive to the implicit ― and sometimes explicit ― racial appeals of Donald Trump’s presidential campaign,” Rachel Wetts and Robb Willer wrote in their paper, which appears this month in Socius, a scholarly journal published by the American Sociological Association.
Wetts is an assistant sociology professor at Brown University and Willer is a sociology professor at Stanford. A previous paper they co-authored examined the way white attitudes toward welfare soured after survey participants were told white Americans would someday lose their majority status.
For the new experiment, the researchers first screened participants to gauge their level of racial resentment, asking them how much they agreed with various statements, including this one: “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.”
Then, about two weeks later, they divided their 800-plus subjects into three groups, and presented each group with a short political statement from an unattributed source. One group got Ronald Reagan lamenting welfare and “inner cities,” absent fathers and pathological poverty. Another group got flamingly racist statements about food stamps from the American Freedom Party. The third, the control group, got Sen. James Inhofe (R-Okla.) whining about “global warming alarmists.”
The professors wanted to test a popular theory that subtly racist cues ― such as terms like “inner city” that many associate with Black Americans ― harness subconscious racial animus to sway white voters against government spending. Statements that are openly racist, on the other hand, are thought to turn off those voters.
“If the dog-whistle hypothesis is correct, we would expect implicit but not explicit racial appeals to increase the impact of racial resentment on whites’ policy evaluations,” Wetts and Willer wrote.
Their experiment confirmed the hypothesis ― but with an unexpected twist. After the participants read the political statements and were asked how they felt about welfare, the ones with the most negative attitude compared to the control group were self-identified white liberals who’d read the implicit racial appeal and who had scored highly on the racial resentment scale.
Conservatives, on the other hand, did not indicate less support for welfare after reading either the coded or the explicit statements, though they started from a less supportive position in the first place.
Liberals who came in with higher levels of racial resentment also soured on welfare after reading the more overtly racist screed, though to a lesser degree ― lending credence to another theory, that explicit racial appeals may have become more acceptable to voters than they used to be.
Wetts said the results were surprising.
“If anything, many people would think conservatives would be more responsive to racial appeals, given the historical centrality of racial appeals to much Republican political messaging,” she said in an email.
Trump has made overtly racist statements, such as calling immigrants rapists, murderers and invaders. But he has also used coded language, describing Black Americans as living in filthy urban hellholes. In his inaugural address, he lamented “mothers and children trapped in poverty in our inner cities,” and pledged to “get our people off of welfare and back to work.”
The Trump administration has also made it a priority to cut welfare through executive action. (There is no federal program actually called “welfare”; Republicans have used the term to describe any social program that provides benefits to people with low incomes.)
Wetts and Willer offered a few theories about why welfare rhetoric would move white liberals more than conservatives. One is simply that conservatives are so familiar with welfare-bashing from Republican officeholders and right-wing media that they can’t be swayed any further.
Another is that liberals may be uniquely vulnerable to this rhetoric because they’re afraid to talk about racial inequality. Wetts and Willer noted that “strong norms of colorblindness in liberal political culture mean negative outcomes among black Americans as a group are rarely discussed.”
For some extra verification, Wetts and Willer tested the same rhetoric, but swapped out welfare for gun control, a topic that may be less saturated with racial connotations. So instead of “misguided welfare programs,” the unidentified speaker in one of the statements complained about “gun access laws.”
The results seemed to corroborate what they found with welfare: “Liberals high in racial resentment were more likely to voice support for increased restrictions on gun ownership after reading the implicit racial appeal compared to a message with no racial appeal.”
The paper also examined data from the American National Election Studies, a series of surveys done before and after every presidential election. Wetts and Willer used mathematical analysis to isolate different characteristics that predicted whether a white liberal would go from voting for Barack Obama in 2012 to voting for Trump in 2016.
They found that “white, racially resentful liberals were particularly likely to switch their vote to Trump” after having previously voted for Obama. (Several studies since 2016 have reached similar conclusions.)
The upshot is that Trump’s racist appeals likely helped him win. Racially resentful liberals represent about 3 percent of the population, or some 10 million people. And it only took 100,000 votes in three states to seal the Electoral College for Trump in 2016.
“They might be a small group,” Wetts said, “but they’re really important politically.”
This article originally appeared on HuffPost.