1.5 Interpreting the findings
Why might this pattern of results be of interest to prejudice researchers?
Looking at Fazio et al.’s findings, it could be argued that they show that white Americans have deep-seated, automatic, negative associations with the category ‘Black Americans’. These associations facilitate (thus speed up) the classification of negative words, but interfere with (thus slow down) the classification of positive words. In other words, reaction times indicate implicit prejudices that emerge in ways that individuals find difficult to control; in fact, individuals are often unaware of the degree to which they harbour such prejudices.
But is this really prejudice?
Implicit measures of prejudice are controversial. Some researchers have argued that the subtle negative associations measured by research such as Fazio et al.’s (1995) study are not the same thing as the overt negative feelings and beliefs studied in traditional prejudice research. Critics, such as Hal Arkes and Philip Tetlock have argued, for example, that implicit measures set the ‘bar too low’ and that the negative associations tapped by reaction time measures fall well short of bona fide prejudice. For example, such associations may reflect shared cultural knowledge of such associations rather than an individual’s personal feelings or beliefs. Alternatively, it may show bias in favour of certain groups without necessarily showing prejudice against others. In other words, ‘… relative difference in RT [reaction time] between two target sets does not necessarily imply hostility or prejudice toward either group.’ (Arkes and Tetlock, 2004, p. 267).
Countering this argument, other researchers have pointed out that implicit prejudice measures are often good predictors of other negative behaviours towards minority groups, strengthening confidence in their validity. In Fazio et al.’s (1995) study, for example, participants who had high scores on the implicit prejudice task, similar to the one you have just completed, later behaved in a less friendly fashion during an interaction with a black experimenter.
It is also noted that such measures overcome many of the limitations of explicit prejudice measures, notably the problem of social desirability. Because these measures tap responses that are automatic and difficult to control, such responses are far more difficult to fake.
It is worth noting, in conclusion, that although implicit prejudices are more difficult for individuals to control, we should not assume that they cannot be changed.