Americans’ implicit biases are changing for the better, new Harvard study finds
- Writing Sample
- Apr 10, 2020
- 3 min read
3 NOVEMBER 2019 (intended for ABC News)
Americans have become less biased towards those of different demographics within just a decade, say researchers at Harvard University.
The study, published this past January in the Journal of Psychological Science, found decreases in both implicit biases, or unconscious attitudes towards others, and explicit biases, or beliefs and attitudes about others we deliberately think about.
Using 4.4 million online tests from 2007 to 2016, the study measured biases in six categories: sexuality, race, skin tone, age, disability, and body weight. While explicit biases decreased for all categories, implicit biases had some exceptions. Nevertheless, this study is the first to find long-term changes in implicit biases, previously thought to be unlikely to change over time.
“The fact that some [implicit] biases ebbed over a ten-year period is cause for hope,” Harvard researchers Tessa Charlesworth and Mahzarin Banaji wrote in their Harvard Business Review (HBR) article. “It shows that even seemingly automatic biases can and do change.”
They added that their research, which used data from Harvard’s “Project Implicit” website, should be used to enact policies that motivate positive behavior and attitude change. The website, created in 1998, includes the popular Implicit Association Test (IAT) for users to test their own implicit biases and surveys to measure conscious and self-reported explicit biases.
The results found that attitudes on sexual orientation showed the fastest, steadiest change in both explicit and implicit anti-gay biases, with a 49% and 33% decrease, respectively. Notably, the researchers predict implicit sexuality bias to phase out completely between 2025 to 2045, attributing the ability to conceal one’s sexual orientation (unlike the other five categories) as a potential reason for the change.
Race attitudes showing pro-White preferences decreased at slower rates, with 37% for explicit bias and 17% for implicit bias. Skin-tone attitudes, or pro-light skin preferences, decreased by 21% for explicit and 15% for implicit bias. Changes in both race and skin-tone attitudes have been particularly rapid since 2012.
The remaining three attitudes, however, have not seen the same improvement in implicit biases over the decade.
Implicit pro-young and pro-able-bodied attitudes, for instance, have remained stable. While explicit age biases decreased by 34%, implicit age attitudes saw a 5% decrease. Similarly, disability biases decreased by 24% for explicit and 2% for implicit.
The only category that has seen an increase in implicit bias is body weight, even with a 15% decrease in explicit bias. The increases in implicit weight-bias were based on both figure (full body) and face, with a 40% increase for pro-thin facial preferences from 2004 to 2010.
The researchers speculate that the increased attention to the obesity epidemic and the health benefits of lower body weight may explain the anomaly.
“That was surprising in and of itself,” Charlesworth said to WBUR. “We typically think about body weight as something that people can control, and so we are more likely to make the moral judgment of, 'Well, you should just change.'”
Despite the lack of progress in implicit attitude change for three of the six categories, the authors concluded that the results are promising.
“These findings highlight the fact that our minds can and do change toward greater equality of opportunity,” they wrote. “Business leaders … must not forget the subtle forms of bias and discrimination that exist today, especially given the stability and even the deepening of some implicit attitudes over time.”
Charlesworth and Banaji are now working on a second study to find why and how some attitudes and groups are changing more than others. But the progress in implicit biases found in their study did not arise on its own, they added, providing implications for policy and future research.
“It's really exciting to show potential for attitudes to change,” Charlesworth said. “Because now we can try and understand why.”
Comments