Tuesday, February 11, 2014

Stereotypes can drive racial classification

It is widely recognized that people are stereotyped by the race they belong to.   It is less widely recognized that the race of a person is driven to some extent by stereotypes.

NPR, on today's Morning Edition, broadcast this story.   The audio and transcript are there.   My summary below the fold.

Aliya Saperstein @Stanford University, Andrew Penner and Jessica Kizer examined data from the National Longitudinal Survey of Youth (NLSY), conducted by the Bureau of Labor Statistics.

In a longitudinal survey, they keep track of a random sample of people for a long period of time; people in the sample are periodically interviewed, and have recorded all kinds of facts, and also race.

Saperstein & co. found that some 20% of the NLSY sample experienced at least one change in racial classification - the interviewers' perception of a sample person's race changed between interviews.

You might think that this is because of people of mixed race,  who might be randomly assigned to one category or another by interviewers.

But what Saperstein & co. found is that the change of race was driven by changes in people's life circumstances.

SAPERSTEIN: If someone went from being employed to being unemployed, or being out of prison to being in prison, or being off welfare to being on welfare, the interviewer was more likely to see the person as black - after they experienced that sort of downward mobility - than before.
INSKEEP: Wait a minute. You're saying you've got a guy, he's a lawyer, he's presumed to be white. He does something wrong, he goes to prison, he comes out of prison, he's a black man.

VEDANTAM: Well, I think this is what Saperstein is trying to say. She's not saying that happens with every lawyer and every person. But she's saying there's a tendency to sort of see race not just through physical characteristics, but through social characteristics.
 We are further told:
In another study, Saperstein looked at differences in how funeral directors listed the race of people who had died. And she found that when people had died as a result of homicide, funeral directors were more likely to list the person as being black, even when family members listed the person as belonging to another race.

There was another study where she found that if the dead person had died of cirrhosis, which is a disorder commonly caused by alcohol abuse, the funeral director was more likely to list the person as being Native American, even when family members listed the person as belonging to another race.
One fascinating thing that Saperstein has found is that it isn't just other people's perceptions of you that change. The survey that she followed also asked people to report their own race. And she found that when people went to prison, they became more likely to think of themselves as black.

Think about e.g., the fact that Murray and Herrnstein's "The Bell Curve" that tried to show that blacks are intrinsically have less IQ, and that IQ is strongly correlated with life success, relied on just such a longitudinal survey sample.   If we assume that the race of 20% of that sample was mutable, just as Saperstein found, what it might do to their already dubious conclusions.

I think it also helps make clearer, to anyone who is not a third+ generation American, just how much the election of Obama to the Presidency upset the racial applecart, and why so many people hate Obama unconditionally. 

When the FBI breaks down homicides into white-on-white, black-on-white crime, relying on racial classifications the police have filed in their reports, is there a significant effect of this prejudice on the statistics? 

Finally, as Vedantam said:
And it's a troubling idea because we say we track people's race in order to address prejudice and disparities, in all the ways that you mentioned at the start of our conversation. But it turns out that the way we track race itself is subject to the very same prejudices.