A perennial problem with asking people about their views is that they are likely to moderate their answers to make them socially acceptable. The ‘implicit association test’ (IAT) is supposed to get round that problem by tapping into people’s deep-held attitudes below the level of their conscious control. But now Guy Boysen and colleagues report that the IAT is also affected by what is socially acceptable.
Boysen’s team used the IAT to measure participants’ attitudes toward homosexuality. Participants responded to pictures of gay and straight couples and pleasant and unpleasant words as fast as possible using two keys. In one condition, the same key was pressed in response to the sight of gay couples and unpleasant words, with another key used to respond to straight couples or pleasant words. A homophobic person would be expected to respond more quickly in this condition because the key pairings are consistent with their attitudes, but more slowly in a second condition in which the same key is allocated to gay couples and pleasant words, with another key for straight couples/ unpleasant words.
Over one hundred and fifty straight students performed this test, with half told their responses would be kept private and the others told the researchers would see the results. If the test is truly implicit, the participants shouldn’t have been able to moderate their performance. However, the researchers found the participants who believed their results would be public showed a significantly reduced bias against homosexuals compared with the participants who thought their results would be private.
So the IAT can be affected by social desirability, but does this happen deliberately or at a subconscious level? In a second experiment, all the participants were told their results would be seen by the researchers. Crucially, however, half were also told physiological measures taken afterwards would reveal whether they had tried to fake their results to make them more acceptable. If moderation of the IAT is under deliberate control, participants subjected to this latter manipulation should have been put off faking their attitudes, but actually there was no difference between the groups. This suggests social desirability can affect the IAT, but only at a subconscious level beyond participants' control.
Boysen, G.A., Vogel, D.L. & Madon, S. (2006). A public versus private administration of the implicit association test. European Journal of Social Psychology, 36, 845-856.
Link to demonstration of the IAT.