Scientists study bias with association test

Yawu Miller | 2/5/2014, 11:01 a.m.

It may not be surprising to learn that most people in the United States — 80 percent of whites — harbor a pro-white bias. Perhaps more surprising is that a large minority of blacks — 40 percent — hold a pro-white bias.

These biases, uncovered through more than 15 years of research, are not the sort test subjects would commonly admit to. Not even to themselves. They are unexamined attitudes being uncovered by cognitive scientists working in the growing field of study on implicit biases, using tests that measure attitudes that most people cannot, or do not want to, admit to harboring.

“We all have these cognitive shortcuts we use to make sense of the world,” explains David Harris, director of the Charles Hamilton Houston Institute at Harvard University. “In moments of decision, it is more likely than not that these biases will influence our behavior.”

The project that gave birth to the study of implicit biases humans harbor is the Implicit Association Test, founded in 1997 by cognitive scientists working out of Harvard University, the University of Virginia and Washington State University.

The earliest version of the test measured variations in the speed with which subjects can match images of black people or white people with positive or negative words. It is designed to measure subconscious attitudes. Since 1997, different versions of the test measuring attitudes toward skin tone, gender, homosexuality and other categories have been added. In all, the test has been administered 15 million times.

The test has granted behavioral scientists unprecedented access into the recesses of the human mind.

“It’s really been a game changer in a lot of ways,” says Jessie Daniels, a professor at City University of New York who writes extensively about racism. “It’s kind of become a Rorschach Test on how you view race.”

“The main research that has really exploded is how different kinds of implicit attitudes affect decision-making,” says Carlee Beth Hawkins, a researcher with Project Implicit, the multi-university collaboration that administers the Implicit Association Test.

One such research project in 2005 found that the 50 test subjects, all trained police officers, were more likely to mistake wallets or cellphones in the hands of blacks for guns than they were if whites were holding wallets or cellphones.

In a 2008 study using the Implicit Association Test, researchers at the University of Pennsylvania and Stanford University found that many whites do not see blacks as fully human and under certain circumstances associate blacks with apes.

And multiple studies involving doctors in recent years have linked doctors’ implicit bias to overt biases in medical care, ranging from a lower likelihood that black patients will receive pain medication to a lower likelihood that black patients complaining of chest pain will be treated for a heart attack.

Black physicians in the studies were far less likely to hold an implicit bias against black patients or provide inferior treatment to black patients. But on the whole blacks are not immune to implicit biases against their own race.