Are Attitudes Toward Others Taking a U-Turn?

Biases against people based on their race, sexual orientation, and disability were consistently dropping across the country, says Tessa Charlesworth, an assistant professor of management and organizations at the Kellogg School. This included both explicitly stated bias as well as implicit bias—broadly defined as people’s less-conscious and less-controllable automatic reactions.  

Then in 2021, the tide started to turn. In data from 2021 to 2025, what Charlesworth saw was “shocking.” Both explicit and implicit bias increased in nearly every area they looked at and among most demographic groups. “The data shows a clear stalling, even a reversal, of attitude progress,” she says.  

Not only were the general tides turning, but the data also showed a surprising difference in the rate of change. The youngest participants (those under 20 years old) showed the largest relative increases in bias, despite previously being viewed as “the vanguards of egalitarian progress,” says Charlesworth, who collaborated on the study with Meriel Doyle, then a predoctoral fellow at Kellogg, and Mahzarin Banaji at Harvard University. The effect was most pronounced among young, conservative men. 

After generally ruling out several common explanations—such as the “spillover” of prejudice between anti-transgender and anti-gay biases—the researchers argue that the best remaining explanation may be a larger sense of threat and societal disruption in the U.S. that began in late 2020 with the onset of the Covid-19 pandemic. The pandemic not only threatened people’s health but increased their sense of economic instability, amplified political divisiveness, and moved most people (and especially young people) online and into social-media spaces that had high toxicity.  

The pandemic created a “perfect storm of many simultaneous threats at once that would completely reshape people’s daily lives and therefore also reshape the direction of their prejudices,” Charlesworth says. 

Changing cultural tides 

The study grew out of previous research by Charlesworth’s team, which had analyzed online surveys on explicit and implicit biases from 2007 to 2020 hosted by the research organization Project Implicit.  

The surveys measured explicit bias by asking participants to rate how much they prefer one group over another, including groups defined by race (e.g., white versus Black Americans), sexuality (e.g., gay versus straight people), or age (e.g., young versus old people). The researchers also measured implicit bias. They did so by tracking how quickly people categorize a rapid series of images and words. For example, if someone is told to categorize pictures of older people and positive words as well as younger people and negative words, they may perform that categorization more slowly than if they are asked to perform the reverse stereotype (older people and negative words; younger people and positive words). The more a survey respondent showed this discrepancy in categorization speed across the pairings, the more that suggested a higher implicit bias against the group. 

In her earlier work tracking biases from 2007 to 2020, Charlesworth found that explicit bias had dropped in every category, spanning race, skin tone, sexuality, age, disability, and body weight. Implicit biases also decreased for race, skin tone, and sexuality. 

In the current study, the team sought to update this record by adding five new years of data.  

They examined the explicit and implicit attitudes toward the same topics from an additional 2.8 million people across 2021 to 2025. Even though the past trajectories (from 2007 to 2020) were widespread and strongly decreasing, the researchers still suspected that 2020 could mark an inflection point. The Covid-19 pandemic could have had an impact on people’s attitudes; much research has shown how health threats, economic instability or job insecurity, and political sectarianism tend to make people reinforce group boundaries and find scapegoats. Still, Charlesworth expected that biases would probably continue to decrease overall, just more slowly than they had in the past.  

So she was surprised to find that the decreases in implicit and explicit bias not only stalled but, in some cases, even reversed. “I had absolutely no expectation that we would see reversals,” she says. 

The team had predicted using computer models that both explicit and implicit bias against gay people would decrease by more than 60 percent if past trends continued. Instead, they found that both had reversed course: explicit bias more than doubled (moving away from near-neutral or no anti-gay bias), and implicit bias increased by 22 percent.

Similarly, explicit bias rose by 53 percent for race, more than doubled for skin tone, and rose by 16 percent for body weight. Implicit bias rose by 15 percent for race, by 12 percent for skin tone, and by 6 percent for body weight. New data on attitudes toward transgender people, which the researchers hadn’t gathered in their past studies, also showed that explicit bias rose by 24 percent, and implicit bias by 15 percent, over the five years of new data collection from 2021 to 2025. 

The only category where bias dropped was age: explicit bias against older people decreased by 9 percent, though implicit bias still rose by 4 percent. And while explicit bias against people with disabilities remained relatively unchanged, implicit bias rose by 7 percent. 

Swimming in different waters 

Reversals happened among both political conservatives and liberals. While liberals’ explicit bias declined for most topics, their implicit bias based on race, skin tone, disability, body weight, and age rose. 

The results suggest that broader cultural forces are influencing most groups of people, regardless of their political orientation. “We’re swimming in different waters today than we were just five years ago,” Charlesworth says. 

The reason, the team surmises, stems largely from the destabilization that the pandemic caused on multiple fronts at once. In addition to causing threats of illness or death, it triggered job losses and shook up financial markets. Political sectarianism intensified as people argued about vaccines, mask requirements, and stay-at-home orders. Finally, the pandemic pushed people into spending more time online and on social media, where toxic messages and misinformation abound. All those factors tend to increase our suspicion and scapegoating of “other” groups. 

And when the researchers broke down respondents by age, the group where attitudes reversed most dramatically was people under 20 years old. In their own previous studies, Charlesworth and colleagues had found that younger people typically showed the fastest decreases in bias. “What we found in the new data was the complete opposite: younger people are rising faster than other age groups,” Charlesworth says. 

For instance, implicit biases against gay people and racial minorities didn’t change much among people over 40 years old. But among people under 20, implicit biases rose by 105 percent for sexuality and 23 percent for race. Charlesworth speculates that “the threat dial was turned up” for this group of younger people because they were exactly the group that felt the most threatened by the destabilization of the pandemic, including economic instability and toxic social media. 

The team then examined smaller subgroups of participants by age, gender, and political orientation, with a focus on bias based on race and sexuality. Young conservative men were consistently among the top two or three demographic groups with the fastest-rising bias against people based on race and sexuality. The researchers suspect that the cause of this trend was potentially because they were more exposed to online spaces endorsing “manosphere” beliefs. 

Re-righting the ship 

The results are “a sobering wake-up call,” Charlesworth says. “Based on past data, we might have assumed that all was well. If we just let things unfold over time, we would become an egalitarian society. That is clearly not true.” 

Undoing this reversal will require reigniting the major cultural forces that initially had society on a path toward egalitarianism before 2020, including fighting for legal changes and reducing economic threats, political divisiveness, and social-media toxicity. Charlesworth acknowledges that “these are the hardest things to undo in our society,” but past trends could offer hope that society might once again “re-right the ship.”  

After all, she says, “the data show clear evidence that people’s attitudes can change.” Even though the trends have been moving in an anti-egalitarian direction recently, bias was moving in egalitarian ways in the past. “Now it is especially critical that we, as individuals, and as collectives with power to influence our leaders and our laws, feel the urgency of turning change back in the direction of our values.”