Attitudes About Race Affect Actions, Even When They Don't

Tiny effects of attitudes on individuals' actions pile up quickly.
Publish date:
Social count:
Tiny effects of attitudes on individuals' actions pile up quickly.
Protestors stand in solidarity with Ferguson. (Photo: 40969298@N05/Flickr)

Protestors stand in solidarity with Ferguson. (Photo: 40969298@N05/Flickr)

Even when attitudes have only a small impact on actions, individual prejudices can easily turn into racist institutions, according to a recent study in the Journal of Personality and Social Psychology.

That conclusion was one of several mostly technical results in a paper about the real-world value of something called the Implicit Associations Test. The IAT uses a special categorization task and translates the time it takes people to perform that task into a measure of a person's underlying, perhaps even unconscious, attitudes about anything from names to racial and ethnic groups. And, as dozens of studies show, most of us, regardless of who we are and where we come from, show signs of underlying racial and other sorts of prejudice, at least as measured by the IAT. Sad, but apparently true.

It doesn't take much at the level of individual attitudes to make society a little—or a lot—less equal.

So far, none of this is precisely what authors Anthony Greenwald, Mahzarin Banaji, and Brian Nosek—all key figures in the development of the IAT—were after. Mainly, they focused on a fairly technical debate about the IAT's ability of their test to predict behavior in the real world. Indeed, some say, there's only a weak link between how a person does on an IAT and how they act on a day-to-day basis—for example, between scores on an IAT for race and police shooting unarmed black teens and preteens.

Greenwald, Banaji, and Nosek dispute that contention, but even then that debate misses a broader point, they argue: Even if the connection between an individual's implicit attitudes and explicit acts is very weak, that connection can have big consequences for society.

Here's an example, based in part on the idea that interactions with employers or police happen over and over again. Let's say that the effect of attitudes on actions is tiny. To be concrete, suppose the differences across police officers in an IAT measure—or any attitude measure, really—of race attitudes predicts just 0.17 percent of the difference between the fraction of blacks and whites a police officer stops. By any reasonable standard, that's a very weak relationship. Yet if whites are stopped one percent of the time they encounter a police officer, blacks will get stopped twice as often. And over 25 such encounters, blacks will get stopped at a rate 17.4 percent higher than whites.

Setting aside for the moment the inside-baseball debate about whether IAT is the most useful measure of prejudiced attitudes, the big picture is that it doesn't take much at the level of individual attitudes to make society a little—or a lot—less equal. On the statistics side, it's not even a particularly new argument. But in this case, it makes a big difference. From a nearly non-existent connection between attitudes and actions, institutional racism is born.