Overview

Implicit association tests (IAT) aim to assess individual implicitly-held beliefs, as opposed to socially appropriate responses. However, IATs have been the focus of several recent research studies whose findings suggest that IAT must be treated with caution.

Background

 

The IAT involves participants pressing keys in response to the presentation of different word categories on a computer screen. The test works on the assumption that if we're using the same hand to respond to two different categories that we associate in our mind – for example, types of dog and positive words – we'll be quicker to respond than when we're using the same hand to respond to two categories that we don't associate in our mind, such as snakes and positive words.

Typically, four categories are presented, with one hand used to respond to the sight of two of the categories (e.g. European names/ negative words), with the other hand used to respond to the other two categories (Arab names/ positive words); followed by a second session of testing with the hand/category associations reversed. A participant who was prejudiced against Europeans would be expected to respond more quickly when they were using the same hand to respond to European names and negative words.

IAT Research

In support of the IAT, studies have shown that participants who register socially acceptable answers when tested explicitly, show signs of prejudice when tested on the IAT. For example, Greenwald et al (1998) showed this in the context of White participants' attitudes towards African-Americans.

However, two studies published last year suggest that attitudes recorded via the IAT may not be reliable. Firstly, Han et al (2006) from Ohio State University , demonstrated that participant attitude, as recorded by the IAT, could be easily contaminated by brief exposure to unconvincing information that contradicted their own attitude.

In this study, 123 participants were presented with detailed information about two Japanese Pokemon toys, one of which was portrayed as far superior to the other. In an explicit test, all participants rated the superior toy as being preferable. However, prior to administering a version of the IAT, that used the names of the two toys and positive versus negative words, half the participants watched a brief video showing two children saying they preferred the inferior toy. Although these participants said they found the children's reasons foolish and unconvincing, watching the video affected their subsequent performance on the IAT. The speed of their responses suggested that they found it far easier to associate the name of the inferior toy with positive words, compared with the other participants who had watched a video consistent with their own original attitudes.

Han et al (2006) concluded that attitudes recorded using the IAT can be contaminated by external information, even if that information is not perceived as credible. Han stated that "The traditional IAT is influenced both by one's personal associations and by extrapersonal associations – ones that are attitude irrelevant but that are valenced and available in memory".

 

Research by Boysen et al (2006) found that participant performance on the IAT is affected by whether the participant believes their results will be made public or not. This counteracts IATs central premise that it taps into people's sub-conscious, implicitly-held attitudes. Over 150 heterosexual students performed an IAT test of their attitudes towards sexuality. The Iowa State University 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.

Boysen and colleagues conducted a second, similar study in which they 'assessed' half of the participants with physiological equipment. Participants were not informed however, that the equipment was not working. The researchers told participants that by recording their heart rate and sweat levels, the equipment would be used to monitor responses to determine if responding dishonestly. All participants were told that their results on the IAT would be made public. Boysen et al suggested that if participants had conscious control over their performance on the IAT, then the group being assessed by the physiological equipment ought to have responded more truthfully, whilst the non physiological monitored group would have reduced bias against homosexuality (as in the first experiment).

Interestingly, there was no difference between the groups. This suggested that although social desirability can affect IAT results, this effect is not necessarily under participants' conscious control. Boysen and colleagues concluded: "In this respect, the IAT at least partially fulfils the promise that it is relatively impervious to social desirability concerns, despite the fact that the IAT responses are malleable".

IAT may be a useful tool, however as the recent research highlights, unless one can be sure of exactly what IAT is measuring and how much conscious control people have over their responses the interpretation of IAT results should be treated with caution.

Web Directions

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. http://faculty.washington.edu/agg/pdf/Gwald_McGh_Schw_JPSP_1998.OCR.pdf

Han, H.A., Olson, M.A. & Fazio, R.H. (2006). The influence of experimentally created extrapersonal associations on the Implicit Association Test. Journal of Experimental Social Psychology, 42, 259-272. http://web.utk.edu/~molson2/HanOlsonFazioJESPinpress.pdf

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. http://www3.interscience.wiley.com/cgibin/abstract/112478310/ABSTRACT

Worth a look

Science News Online article about the IAT: http://sciencenews.org/articles/20060422/bob9.asp

Project Implicit, where you can try out the IAT for yourself: https://implicit.harvard.edu/implicit/