Sat 14 Mar, 2009
Tags: Implicit racial bias, Racial prejudice, Racial segregation
A new study came across my desk this week, which suggests that whites who have experience distinguishing the faces of individual black people may display less “implicit racial bias.”
In this research, white subjects were given two distinct types of training with the faces of black Americans and an equal number of Chinese faces. In the “individuation condition,” subjects were trained to discriminate among different individual black faces, while in the “categorization condition,” subjects simply practiced categorizing faces as black or Chinese.
At issue in this study is so-called “implicit racial bias,” which purports to measure subconscious racial bias through one of several types of tests. This is racial prejudice which does not match conscious judgments about race, but which manifests in laboratory tests and which might play a role in more subtle forms of real-world prejudice.
Implicit racial bias is commonly measured through such tests as the “Implicit Association Test” or the “Bonafide Pipeline,” which try to detect subtle differences in how subjects respond to people of different races. This research, however, used a new test, called the “Affective Lexical Priming Score” (ALPS).
The ALPS is generated by showing the subject a series of faces. Each image is displayed with a string of letters, which might be a positive, negative, or neutral English word, or a nonsense word. The subject is simply asked whether the string of letters constitutes a valid English word. The assumption is that implicit racial bias against blacks by white subjects will manifest as quick response times when a black face is paired with a negative word, and slower response times when a black face is shown with a positive word.
Subjects who were given individuation training showed an improved ability to distinguish among black faces. This corresponds to a reduction in what is known as the “Other-Race Effect,” where people of all races have more difficulty distinguishing individuals of races other than their own, presumably because they have more experience distinguishing individuals of their own race.
Along with their improved facial recognition skills, these same subjects also showed a reduced implicit racial bias, as measured by the ALPS. Moreover, the degree of improvement in facial recognition was significantly correlated with the degree of reduction in ALPS score.
Subjects who were merely given the categorization training, on the other hand, displayed no improvement in distinguishing black individuals and no reduction in ALPS-measured implicit racial bias.
It is difficult to draw broad conclusions about real-world social phenomena from relationships among laboratory training and tests. However, this research suggests that implicit racial bias includes a perceptual processing component which can easily be altered with simple experience or training.
I think that two rather different conclusions might be drawn from this result. The first is that an important component of subconscious racial bias in the real world might stem from nothing more than a relative lack of familiarity and experience with people of other races. This would explain why people score so differently on tests of implicit racial bias, depending on their demographic characteristics and life experiences. It would also reinforce the importance of racial diversity in one’s environment, especially as a child.
The other possible implication is that tests of “implicit racial bias” may be largely measuring perceptual mechanisms which are related to race and which can be detected in careful study designs, but which are quite distinct from prejudice, subconscious or otherwise. A lack of familiarity with distinguishing individuals of other races may be deeply unfortunate, but it does not by itself equate to racial bias at all.