Implicit and explicit prejudice.

During the 1990s the idea of the existence of new forms of prejudice was taken a step further. The distinction was now made between explicit prejudice, operating at a conscious level and therefore including both traditional and modern racisms, and implicit prejudice, which was assumed to be automatically activated by target persons and to operate largely unconsciously.

A variety of measures of implicit prejudice have been used. Some are indirect or covert behavioral indices of prejudice, such as linguistic biases, eye contact, or nonverbal behaviors, which could be subject to intentional control but typically would not be. Others are truly implicit measures of automatic cognitive or physiological responses that cannot readily be intentionally controlled. The most common have been "priming techniques," which measure how quickly positive or negative stereotypes can be activated after individuals are briefly exposed to a group or social category label, and variations of that procedure, such as the Implicit Association Test (IAT), which assesses how strong the temporal association is between a group or social category label and positively or negatively evaluative terms.

It was initially assumed that these implicit measures might provide more accurate information about peoples' real attitudes, but research findings have suggested a more complex picture. Generally whites have shown markedly greater negativity toward blacks on implicit measures of racism than on explicit measures, as would be expected. The difficulty with the concept of implicit prejudice, however, has been that different measures of implicit stereotyping and prejudice have tended to be only weakly correlated with each other, suggesting that they might not be measuring a single dimension at all, and they have also tended to be uncorrelated with explicit measures of prejudice or behavioral indices of discrimination.

One possible reason for the discrepancies between implicit and explicit measures is that they might reflect different response systems. Dovidio et al. have shown that automatically activated implicit evaluative stereotypes predict people's spontaneous interracial reactions (such as eye contact and rate of blinking) but not their more deliberative, controlled interracial judgments or behaviors (overtly expressed attitudes, interview evaluations, or judgments of guilt in simulated court cases). Explicit prejudice measures such as the modern racism scale on the other hand predicted deliberative responses but not spontaneous reactions.

Two theories of prejudice have been used to explain why there might be considerable discrepancies between individuals in their explicit and implicit prejudice against a particular group. Both suggest this discrepancy might arise from individuals acquiring values that motivate them to attempt to control prejudiced attitudes or stereotypes that were acquired earlier in socialization and that are therefore automatically activated by target persons. However, the nature of the values and motives proposed by the two theories differ in certain crucial respects. The theory of aversive racism suggests that many whites have underlying, covert racist attitudes but adopt and express egalitarian attitudes at an overt level in order to maintain a self-image of themselves as nonprejudiced and egalitarian. Because of their covert racism, they will act in discriminatory ways in situations in which they can rationalize or excuse it and will do so without guilt. Patricia Devine's theory suggests that whites acquire negative stereotypes of blacks during socialization. Later some whites come to internalize explicitly antiracist values, which are genuine and deeply held. As a result, their underlying and automatically activated prejudiced reactions engender guilt and are therefore inhibited whenever they are under conscious control.

These theories, and most of the research on implicit prejudice, have assumed that implicit prejudice and stereotypes are stable evaluative orientations that are automatically activated in response to stigmatized group members. Recent research, however, has shown that implicit prejudices seem to be malleable and fluid and to vary markedly in response to situational cues and motivational influences. Andrew Karpinski and James Hilton have therefore suggested that implicit attitudes reflect the evaluative associations with out-groups that people have been exposed to in their environment and not the degree to which they endorse these associations. One implication of this is that because these "implicit" associations have not been consciously articulated and organized in "explicit" form, the actual evaluative out-group associations elicited in any situation may be markedly influenced by the nature of the activating cues, procedures, and situation. This could explain the variability in implicit prejudice across situations and measuring techniques and the weak correlations between different implicit measures and between them and explicit measures of prejudice.