Why insisting you’re not racist may backfire

University of California - Berkeley Haas School of Business

When you insist you’re not racist, you may unwittingly be sending the opposite message.

That’s the conclusion of a new study* by three Berkeley Haas researchers who conducted experiments with white participants claiming to hold egalitarian views. After asking them to write statements explaining why they weren’t prejudiced against Black people, they found that other white people could nevertheless gauge the writers’ underlying prejudice.

“Americans almost universally espouse egalitarianism and wish to see themselves as non-biased, yet racial prejudice persists,” says Berkeley Haas Asst. Prof. Drew Jacoby Senghor, one of the authors. “Our results suggest that the explicit goal of appearing egalitarian might blind people to the possibility that they could be communicating, and perpetuating, prejudicial attitudes.”

Co-authored by Derek Brown, PhD 24, and Michael Rosenblum, PhD 20–a post-doctoral scholar at NYU Stern School of Business–the study builds on past research finding people’s racial prejudice “leaks out” through nonverbal behavior, such as facial expressions or physical distance. In a series of experiments published in the Journal of Experimental Social Psychology, the researchers looked at perceptions based solely on written content.

They selected a group of white participants, screening out the small percentage who expressed overt prejudice, and scored subjects’ racial attitudes with two widely used assessments. The subjects were then asked: “Do you believe that all people are equal and should have equality of opportunity? Why or why not?,” and “Are you prejudiced toward Black people? Why or why not?” A second group of white participants, asked to read the written responses, accurately estimated how the writers had scored on the prejudice scale.

Linguistic cues

In a second experiment to parse out whether people were signaling racial attitudes intentionally or inadvertently, they asked one group to answer as honestly as possible and another group to answer “in the least prejudiced way possible.” There was no difference to the readers, who accurately scored both groups’ answers.

“That gave us some confidence that people are naturally trying to come across as egalitarian, but something about the language they choose is betraying them,” Rosenblum said.

What were those linguistic cues? The most powerful indicator, they found, was language that dehumanized or objectified African Americans–for example, “I have a great relationship with the Blacks.” Other characteristics such as defensiveness, references to personal responsibility, or a belief that equal opportunity exists were strongly associated with higher levels of prejudice, and cues such as focus on equity or an acknowledgement that inequality exists were associated with lower levels of prejudice. Interestingly, references to being colorblind or mentions of personal contact with Black people weren’t indicative of the white participants’ attitudes.

“This demonstrates that peoples’ use of the cues are meaningful not only for how prejudice is expressed, but also how egalitarianism is perceived,” said Brown.

Contagion effect

A third experiment had a sobering result. The researchers found that white participants reported greater prejudice towards Black people after reading statements from the self-avowed white egalitarians who scored high on underlying prejudice. In other words, the readers mirrored the attitudes of the writers, even when they identified themselves as ideologically dissimilar (conservative vs liberal).

“We don’t know reading other people’s views gave them permission to express more prejudice, or whether they thought that this is the norm and their actual prejudice level changed, but there seemed to be a contagion effect,” Rosenblum said. “One of the lessons here is that words carry weight. It does seem that this is one way that prejudice is unwittingly spread.”

/Public Release. This material comes from the originating organization/author(s)and may be of a point-in-time nature, edited for clarity, style and length. The views and opinions expressed are those of the author(s).View in full here.