Prospect Theory and the Implicit Association Test

The following is my sixth-week reaction paper from a Fall 2006 class, Biological Approaches to Political Behavior, taught by Professors John Hibbing and Kevin Smith.

The study of emotions and automatic, unconscious reactions links most of this week’s readings. Mutz and Reeves find people’s well-learned manners regarding tone of voice and the preservation of personal space in conversations are violated by television’s “in your face� style, which leads people to distrust the people they usually only see on TV – politicians. Olson and Marsheutz demonstrate that it may well be impossible for even the most egalitarian person to judge people equally, for example, when one person is attractive and the other isn’t. Bower covers the controversy over the Implicit Association Test, which is just one example of a suite of diagnostic tools psychologists are using to tap unconscious attitudes people can’t report on surveys. Sanfey et al. show people respond emotionally to unfair offers in economic games. Morris et al. and Lodge and Taber both write about the “hot cognition� phenomenon, where items in long-term memory arrive in working memory with positive or negative affect already attached. Finally, Camerer et al. and Lieberman and Schreiber present classic gifts to graduate students, offering laundry lists of testable hypotheses.

The papers on hot cognition struck me as a bit of old news. The Republican Party, perhaps without labeling it as such, seem already to have noticed the potential for strong associations between political concepts and emotions. As Geoffrey Nunberg noted in “Talking Right� and George Lakoff wrote in “Don't Think of an Elephant,� conservatives have already successfully attached negative affect to many words liberals use to describe themselves and others use to describe liberals. It is just a matter of time before “progressive� goes the same way. The hot cognition hypothesis could deliver something we don’t already know if some enterprising political operative used its insights not to negatively valence political words like “liberal,� but found a way to use Morris et al.’s findings to associate nonsense words in people’s minds with the opposing party’s political concepts. Because, as Morris et al. found, this inhibits the recognition of semantically unrelated concepts, the result could be targets of these efforts left with negative affect toward liberal words on the emotional side and inability to process them on the cognitive side.

Turning to Bower’s piece on IAT, I wonder whether there is any relationship between the seeming existence of implicit associations and prospect theory. It seems that people can, with effort, work past their prejudices against the out-group, whether those prejudices spring from an evolutionarily adaptive fear of the other, are learned, or arise from some combination of genes and the environment. A white Southerner can learn to like individual blacks, for example, or a Hindu Indian can befriend a Muslim Indian. But the IAT results (methodological questions aside for the moment) seem to indicate that people unconsciously hang on to the notion that “the out-group is the enemy.� Is this because our minds, without our knowing it, are operating in accord with prospect theory? That is, are they valuing the avoidance of a loss (failing to recognize an enemy) over gaining a gain (a new friend or business associate)? One way to test this would be to find a bunch of white and black people who report large numbers of friends of the opposite race, then randomly insert those friends’ pictures into a better-run IAT and see what happens. My hypothesis would that the brain is so risk-averse in a prospect theory sense that positive association latencies would be high even when friends’ pictures are presented. (A “better-run IAT� would have the same control scheme for the two halves of the test; I agree with the letter from Mia Molvray that learning two control patterns probably contributes to measurement error in the test.)