Any discussion of faces has to include some discussion of race. There is a cliche that has been around for years: “they all look alike to me.” Often, this cliche is offered with Caucasians talking about Asians, but all other groups can be considered as well. More important, and unhappily, the cliche has a core of truth.

Many studies have compared people’s ability to recognize faces of their own race with their skill in recognizing other faces. Some of the evidence comes from laboratory studies. Troublingly, some of the evidence comes from police records. The results are clear. The error rate for Whites identifying Blacks (or vice versa) is about 50 % higher than the error identifying a same-race face. Other results show a similar pattern for Whites and Asians, and also for young people recognizing old people and vice versa.

Interestingly, this effect is not a direct psychological result of racism, but of course it can and does have racist consequences. Specifically, the evidence says that the degree of difficulty you’ll experience in recognizing other- race faces is independent of your degree of prejudice. What does matter (at least to some extent) is degree of contact, but we need to be careful here; the effect does not depend on whether you see a mix of black and white faces as you go down the street. Instead, the effect depends on whether there is a mix of black and white faces in your social life – so that you routinely need to distinguish and identify specific individuals of other races. Even here, though, the difference for other-race recognition remains, it’s just smaller. The effects of racial segregation, then, do persist and lead to something that has dire consequences in criminal proceedings, since mis-recognizing a face by witnesses to a crime can lead to wrongful convictions.

Oddly enough, there is at least some indication that the specialized mechanisms for face recognition (see all of this week’s blogs) are less used for other-race faces. As a result, in recognizing them, people often seem to shift into a less effective feature-by-feature analysis. Peculiarly this sometimes makes them more accurate in describing other- race faces but less so in recognizing them.

I am attaching a classic result for the details: http://digitalcommons.utep.edu/cgi/viewcontent.cgi?article=1004&context=christian_meissner

If you want to read a compilation of more recent exploration of the topic look here (truth in advertising, it’s a volume I edited heavily since it was written on or near our kitchen table.)

https://www.amazon.com/Science-Perception-Memory-Pragmatic-Justice/dp/019982696X/ref=sr_1_3?ie=UTF8&qid=1487348239&sr=8-3&keywords=daniel+reisberg