An early entry for "Worst Science Reporting of the Year"

From AP: [faked racist event]

Those asked to predict their reaction to either comment said they'd be highly upset and wouldn't choose the white actor as their partner.

Yet students who actually experienced the event didn't seem bothered by it — and nearly two-thirds chose the white actor as a partner, the researchers report Friday in the journal Science.

"didn't seem bothered" based on what scale? what measurement? Was this simply anecdoctal evidence from a camera? Was there some sort of way to measure it?

And: two-thirds of who? All of the participants (because one-third of them didn't receive any special stimulus...) or two-thirds of the ones who received the stimulus?

What was the task? Although this is certainly not going to remove the tag of "racism" from the choice one way or the other, it is going to modify the specifics of how that racism plays out. For instance, a stereotypical "white" task is going to lend itself to racist choices simply based on the task itself, not the belief system of the actors regarding social control mechanisms, which is the variable this study seems to be studying. Of course, considering the shoddy reporting, the study itself might have been trying to study the effects of friendship on race perception, and the reporter just never got to that part. (or, to continue this totally random example, didn't understand the term "dyadic relationship")