Cover Story

Jerry Seinfeld popularized the phrase "not that there's anything wrong with that," after noting that someone was gay, to assure himself and others that he was not homophobic. It's a refrain that echoes through popular culture--likely one people often identify with because, for many, being politically correct is an important social statement. But how much truth is there to what people say--and maybe believe--about their private nonstereotypical thoughts?

It appears they're usually doing more unconscious stereotyping than they realize, find University of Virginia (UVA) assistant psychology professor Brian Nosek, PhD, and his collaborators, psychologists Mahzarin Banaji, PhD, of Harvard University, and Tony Greenwald, PhD, of the University of Washington, who've set out to answer that question through a computer-based test of implicit attitudes. Over six years they've developed and used their Implicit Association Test (IAT) to collect data from more than 3 million completed online tests, learn about the nature and magnitude of stereotypes and help people face up to their unconscious attitudes and beliefs.

Hidden biases

The IAT, part of a research initiative called Project Implicit, asks participants to pair concepts, such as "good" and "fat" or "bad" and "thin," in a series of timed tests. The researchers measure response times and check small variations to determine if someone is biased toward thin over fat, gay over straight or white over black, Nosek says.

Presumably, if a person associates "fat" more closely with "bad" than "good," he or she will respond more quickly when "fat" is paired with "bad" than when "fat" is paired with "good," says Nosek. The test also collects information on demographics and explicit attitudes. For example, participants state how much they weigh and how strong their preference is for fat or thin people.

A study of 600,000 tests completed in 2000 shows that both white and black participants had an implicit preference for white faces and names. White respondents indicated an explicit preference for white people, but showed a stronger implicit preference on the test, according to the results, published in 2002 in Group Dynamics: Theory, Research and Practice (Vol. 6, No. 1). Meanwhile black people indicated an explicit preference for black people, but showed a statistically weak implicit preference for white people--results that raise important questions about how stereotypes permeate even to those who are being stereotyped, Nosek says.

Biases revealed

Indeed, revealing to people their own implicit biases is a major purpose of the IAT, Nosek says.

"The idea that we have implicit thoughts, feelings and behaviors that are contrary to how we'd like to behave is shocking," he says. "So this is a tool that is interesting and important not just for the advancement of science, but also for self-reflection and for getting people talking about what it means to have implicit associations."

The 600,000 online tests also revealed general population preferences for straight people over gay people, preferences for thin people over fat people and preferences for young people over old people. They also demonstrated that people generally associate women with the liberal arts and men with science.

Now funded by a National Institute of Mental Health grant, Project Implicit and the IAT are making a mark in several psychology fields including clinical psychology, social psychology and neuroscience, and outside fields such as diversity education, health behavior and the law, Nosek adds. Researchers in those fields are using the IAT to separate out implicit effects--often attributed to the fuzzy category "intuition"--in their own larger research projects.

As yet, there is little data about what sort of effect experiencing one's own implicit bias with the IAT has on individuals, but the anecdotal responses of teachers, trainers and others who use the test to teach people about their personal biases are more than the researchers say they hoped for.

"We get people saying that the tool was important in a workplace or a classroom for starting up a dialog about what it means to be biased, what prejudice is and what people can do about it," Nosek says. "That makes me think that we're doing exactly what science and education are supposed to do: Bring new awareness."

Further Reading

Take the Implicit Association Test or find out more about the multi-university collaboration, Project Implicit, at www.projectimplicit.net.