Does the Implicit Association Test (IAT) Really Measure Racial Prejudice? Probably Not
Recent research shines new light on the Implicit Association Test.
Posted January 28, 2011 | Reviewed by Jessica Schrader
The Implicit Association Test (IAT) was created by Anthony Greenwald and colleagues  and measures the strength of automatic associations people have in their minds. Many people have taken the test online and have found that they are faster to associate positive words with names of white people rather than Black people. Mass fear has ensued that perhaps most of America really is racist. An even greater fear is that Americans are racist but don't even know it; a situation that seems difficult to change.
Should people be this concerned about their results on the IAT, or is everyone worrying needlessly?
Recent research is shining new light on the IAT, offering an alternative explanation of what the IAT really measures. And the results have important real-world implications.
It's well known that people are prejudiced against the "out-group." Perhaps the IAT-effect is just a result of the human capacity to associate positive stimuli more easily with their in-group, and negative stimuli more easily with their out-group. In other words, perhaps the IAT is tapping into a more general quirk of human nature rather than a specific race effect.
A few recent studies are consistent with this idea. In one study, researchers administered two different versions of the IAT . In one version, the in-group was "French and Me" and the out-group was "North African." Using this version, they found an IAT-effect. In another version, the two categories were "French" and "North-African and Me." In this version, the effect completely disappeared! This suggests the crucial factor was in-group/out-group membership, not nationality. In another study by the same researchers, they established the association with either the in-group or out-group before administering the IAT, and again found that when people associated themselves with the out-group there no longer was an IAT-effect.
In another study, a different team of researchers administered the IAT to three different groups of Americans: a Caucasian group, an African-American group, and a Latino group . They found that the White-Black IAT-effect was largest for those in the Caucasian group, and smallest for those in the African-American group. Conversely, the White-Latino IAT-effect was largest for the Caucasian group and smallest for the Latino group. For those in the Caucasian group, there was no difference in the White-Black IAT-effect and the White-Latino IAT-effect. Again, these findings suggest that the relevant factor is in-group/out-group, not race.
Finally, a Dutch team of researchers looked at the issue by replacing a racially charged out-group name (Moroccan) with a racially neutral out-group name (Finnish) . (Note: I take the researchers at their word that in Amsterdam, "Finnish" is racially neutral whereas "Moroccan" is racially charged).
When Dutch names were contrasted with either Moroccan or Finnish names, they found the IAT-effect. More interestingly, when Moroccan names were contrasted with Finnish names, no IAT-effect was found! These results suggest that the racially-charged Moroccan names were processed in a similar way as the racially-neutral Finnish names.
What factors influenced the processing of the out-groups? For both of the in-group/out-group comparisons (Dutch-Finnish and Dutch-Morrocan), they found that when positive concepts and the in-group (Dutch names) required the same button press people required less time to encode the stimuli or to map their decisions onto the response keys and were less cautious compared to when positive concepts and the out-groups (Finnish or Morrocan names) required the same button press. The same effects weren't found in the Finnish-Morrocan comparison (where both were out-groups and therefore there was no in-group/out-group comparison).
The Dutch study  ruled out potential explanations of these results such as name familiarity (maybe people came to the study with more familiarity for certain names than others) and the context in which the Moroccan category was presented (perhaps presenting two out-groups in one IAT changes the context such that the out-groups are no longer viewed as out-groups).
Instead, they prefer an explanation put forward by another group of researchers  that it is more intuitive processing a positive word associated with an in-group than a positive-word associated with an out-group. Processing a positive word with an out-group requires a switch in mental set in order to retrieve the correct category membership and this takes up more time.
Taken together, these studies suggest that the IAT-effect is due to in-group/out-group membership and is not based on racial prejudice.
Racial Prejudice in the Real World
These results have important real-world implications. Racial prejudice is still a very serious problem across the world. It's therefore important to really pinpoint what accounts for explicit prejudice and make sure we are getting the cognitive process(es) just right.
Research has shown that those who show a strong IAT-effect are more likely to demonstrate overt racist behavior [6, 7]. The correlation is not that large though. As the Dutch researchers point out, caution should be used when making claims about the IAT's ability to measure characteristics of a person that cause racist behaviors.
To me, the most interesting question is why some people with a strong IAT-effect show overt racism while others with a strong IAT-effect do not. The results of the Dutch study suggests that the IAT-effect in itself is not that revealing about racial preferences. People who show a strong IAT effect shouldn't necessarily panic that they are unconscious racists.
Perhaps individual differences in the IAT really are just measuring differences in intelligence and the ability to exert cognitive control and that is the pertinent factor that is related to overt prejudice. Some recent brain research supports this idea.
One brain study used fMRI to examine participants while they were taking the IAT . Brain areas relating to cognitive control and conflict resolution (the dorsolateral prefrontal cortex and anterior cingulate) were most active during conditions in which items from incongruent categories (e.g., insect + pleasant) shared a response key than when items from congruent categories (e.g., flower + pleasant) shared a key. According to the researchers, their findings suggest that greater cognitive control was required in conditions in which it was necessary to overcome the strong tendency to map emotionally congruent items to the same response key. Note that this account is very similar to the one mentioned earlier .
Further research has shown the role of inhibiting strong gut reactions in determining the IAT-effect. Researchers had White participants view faces of unfamiliar Black and White males . Participants who showed greater activation in a region of the brain associated with fear and negative emotions (the amygdala) while viewing Black faces relative to white faces tended to score higher on two measures of unconscious race evaluation: the IAT and the eyeblink response. In a second experiment, they did not find the same pattern of brain activation when the faces were familiar and the participants regarded the Black and White individuals positively.
In a related study, researchers had participants view Black and white faces either below the threshold of awareness (subliminally) or above the threshold of awareness (supraliminally) during fMRI . When presented subliminally, the amygdala was more active for Black faces relative to White faces. This effect was reduced when the faces were presented supraliminally. Interestingly, control regions in the prefrontal cortex showed greater activation for Black faces compared to white faces when presented supraliminally. Also, the IAT-effect was related to a greater difference in amygdala activation for Black faces relative to White faces, and activity in the prefrontal cortex predicted a reduction in amygdala activation from the subliminal to the supraliminal condition. According to the researchers, this provides evidence for neural distinctions between automatic and controlled processing of social groups, suggesting that controlled processes may play a role in automatic evaluations.
Viewed in the light of the Dutch study mentioned above , these brain studies suggest that people with lower levels of cognitive control may be less likely to inhibit emotions about those in the out-group. The effect may not necessarily be related to race.
Racial discrimination is a real problem throughout the world. An important step toward eliminating racism is understanding how culture shapes our minds, and how our minds in turn shape the world. The research I just reviewed suggests that researchers may have overestimated the degree of people's implicit racial prejudice.
This doesn't mean we are in the clear. Throughout the course of evolution, humans evolved the ability to quickly categorize those who are in the "in-group" and those who are in the "out-group". This skill can be adaptive when processing a lot of information, but can also be harmful to society when it influences racist thoughts and behaviors. Therefore, we should be very careful how different groups are portrayed in the media, schools, and society. The faster we can automatically associate people with our in-group, the less likely we will be to implicitly and overtly demonstrate racial prejudice toward them.
Of course, there is still a lot more to learn. Researchers should continue to investigate what the IAT is really measuring and why some people become racists and others do not. Such knowledge will hopefully bring us closer to eradicating racism.
© 2011 Scott Barry Kaufman, all rights reserved.
Follow me on Twitter!
 Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74, 1464–1480.
 Popa-Roch, M., & Delmas, F. (in press). Prejudice Implicit Association Test effects. Zeitschrift fu ̈r Psychology/Journal of Psychology.
 Blair, I. V., Judd, C. M., Havranek, E. P., & Steiner, J. F. (2010). Using community data to test the discriminant validity of ethnic/racial group IATs. Zeitschrift fu ̈r Psychologie/Journal of Psychology, 218, 36-43.
 van Ravenzwaaij, D., van der Maas, H.L.J., & Wagenmakers, E-J. (in press). Does the name-race implicit association test measure racial prejudice? Experimental Psychology.
 Klauer, K. C., Voss, A., Schmitz, F., & Teige-Mocigemba, S. (2007). Process components of the implicit association test: A diffusion–model analysis. Journal of Personality and Social Psychology, 93, 353–368.
 Greenwald, A. G., Poehlman, T. A., Uhlmann, E. L., & Banaji, M. R. (2009). Understanding and using the Implicit Association Test: III. Meta–analysis of predictive validity. Journal of Personality and Social Psychology, 1, 17–41.
 Nosek, B. A., Greenwald, A. G., & Banaji, M. R. (2007). The Implicit Association Test at age 7: A methodological and conceptual review. In J. A. Bargh (Ed.), Social psychology and the unconscious. the automaticity of higher mental processes (pp. 265–292). London: Psychology Press.
 Chee, M. W. L., Sriram, N., Soon, C. S., & Lee, K. M. (2000). Dorsolateral prefrontal cortex and the implicit association of concepts and attributes. Neuroreport: For Rapid Communication of Neuroscience Research, 11, 135-140.
 Phelps, E. A., O'Connor, K. J., Cunningham, W. A., Funayama, E. S., Gatenby, J. C., Gore, J. C., et al. (2000). Performance on indirect measures of race evaluation predicts amygdala activation. Journal of Cognitive Neuroscience, 12, 729-738.
 Cunningham, W. A., Johnson, M. K., Raye, C. L., Gatenby, J. C., Gore, J. C., & Banaji, M. R. (2004). Separable neural components in the processing of black and white faces. Psychological Science, 15, 806-813.