Participants in the study judged pairs of pictures of children, half of which portrayed actual siblings. In one condition, participants were only asked to rate the facial similarity of each pair of children. In a second condition, participants were asked to classify each pair as sibling or non-siblings. Researchers found that the mean similarity ratings of the first group contained as much information about genetic relatedness as did the actual judgments of genetic relatedness by the second group. These ratings could also be transformed into accurate estimates of the probability that a given pair of children portrayed siblings. The study was conducted by researchers Laurence T. Maloney, of New York University's Psychology Department and Center for Neural Science, and Maria F. Dal Martello, of the Department of General Psychology at the University of Padova in Italy.
"We have shown that a complex, difficult to characterize perceptual judgment ('facial similarity') can be simply defined in evolutionary terms: judged facial similarity of children appears to be little more than a visual assessment of genetic relatedness," said Maloney.