Understanding Prejudice
Understanding Prejudice
Return Home

Reading Room

Exercises and Demonstrations
Multimedia Center
Teacher's Corner
Directory of Experts
Links on Prejudice
About Us
Privacy Policy
Contact Us

Reading Room
The Psychology of Prejudice: An Overview

Probing Implicit Biases

How is it possible to measure implicit attitudes and beliefs when people may not even know that they have them? One of the most common ways is with an experimental technique known as "priming" (Wheeler & Petty, 2001; Wittenbrink, Judd, & Park, 1997). Typically, participants in these studies are exposed to a word or image that brings to mind thematically related ideas or associations concerning a target of prejudice (e.g., an ethnic minority group). Then, once an implicit prejudice or stereotype has been activated, researchers can assess its strength, content, and effect on other attitudes, beliefs, and behavior.

In an early experiment using this technique, Patricia Devine (1989) had White college students watch a screen capable of displaying words so rapidly as to be undetected. In one experimental condition, participants were shown a subliminal series in which 80% of the words were stereotypically associated with African Americans (e.g., jazz, rhythm, athletic, basketball, slavery). In another condition, only 20% of the words were associated with African Americans. Next, people were asked to read a brief scenario and judge the actions of a person it described. Devine found that people in the 80% condition -- who, unbeknownst to them, had been heavily primed with stereotypic words -- later judged the person as relatively more hostile (in keeping with the general activation of a stereotype concerning African Americans). Furthermore, this activation occurred regardless of how high or low participants had scored on explicit measures of racism, suggesting that even when people do not believe in racial stereotypes, merely knowing about the stereotypes may be enough to trigger discrimination.

Bullfighter teasing a bull One of the most popular techniques for probing implicit biases is the Implicit Association Test, or IAT (Greenwald, Banaji, Rudman, Farnham, Nosek, & Mellott, 2002; Greenwald, McGhee, & Schwartz, 1998). The IAT is a computer-based test that measures how rapidly people are able to categorize various words and images, and it capitalizes on the fact that most of us identify words and images more rapidly when they come from closely related categories than when they come from unrelated categories. For instance, if you associate librarians with intelligence and bullfighters with violence, you can probably tell in a split-second that synonyms for intelligence like smart and brainy relate to the dual category "librarians or intelligence," and synonyms for violence like aggression and brutality relate to the dual category "bullfighters or violence." But what if we switch the elements around, and you are asked whether smart and brainy relate to the dual category "librarians or violence" or to the dual category "bullfighters or intelligence"? In this case it will probably take you longer to match smart and brainy with the category containing "intelligence," because these dual categories contain elements that are not stereotypically related to each other. Thus, by comparing the speed with which people categorize words or images, the IAT indirectly assesses how closely people associate certain elements with each other. To examine racial stereotypes, for example, the test might replace librarians and bullfighters with Whites and Blacks. With this version of the IAT, faster responses to "Whites or intelligence" and "Blacks or violence" (compared with "Whites or violence" and "Blacks or intelligence") could indicate the presence of an implicit stereotype.

The Implicit Association Test has been used to measure a variety of hidden associations, such as implicit racial and gender stereotypes, attitudes toward elderly people, and preferences for particular political candidates (Greenwald, McGhee, & Schwartz, 1998; Nosek, Banaji, & Greenwald, 2002). Implicit associations have even been detected in minimal group research, when people have no prior group experience yet display positive associations with ingroup member names and negative associations with outgroup member names (Ashburn-Nardo, Voils, & Monteith, 2001). As with other measures of implicit stereotyping, IAT scores have also been linked to behavioral measures of discrimination. For instance, one study found that White students with pro-White IAT scores later treated a White conversation partner better than a Black conversation partner, as judged by independent raters who watched videotapes of the conversations (McConnell & Leibold, 2001).

Previous page
Page 13 of 27
Next page