We know that “average” faces are judged to be more attractive than the faces of the individuals making up the average. But this doesn’t tell us what the mechanism for judging attractiveness is. Do we judge faces to be attractive because they are potential mates, or is there some other reason for perceiving attractiveness?
Jamin Halberstadt and Gillian Rhodes came up with a novel way to try to answer that question: instead of faces, they asked participants to rate other things. If we rate average birds as more attractive than actual examples of birds, then this could suggest that we have a general mechanism for judging attractiveness—after all, we can be reasonably certain that most people aren’t seeking out birds as potential mates.
Halberstadt and Rhodes actually tested three different kinds of objects: birds, fish, and cars. For the bird experiment, they took drawings of 14 different passerines—robins, sparrows, and their cousins—and used software to generate a picture of an “average” passerine. Next, they took the drawings of the individual birds and distorted them in two different directions. They made the drawings more average by distorting them towards the average passerine. Next, they made exaggerated “caricatures” of the birds by emphasizing the differences between the original drawing and the average drawing. Thus, for each bird, a set of seven drawings was generated:
Participants rated each of these images on a scale of 1 to 10 for familiarity, averageness, or attractiveness. In every case, the more average birds were rated as more familiar, more average, and more attractive.
For the fish and car experiments, Halberstadt and Rhodes used a different technique. Starting with 32 fish drawings, they took two fish and averaged them together. Then they took two of these drawings and averaged them to create an average of four fish, continuing until they made one overall average fish comprising all 32 species. Again, participants rated each of these fish for attractiveness, familiarity, or averageness. Individual fish were rated as less attractive than two fish averaged together, and two-fish averages were less attractive than four-fish averages, but after that the pattern flattened out. 32-fish averages were no more attractive than 2-fish averages. There were similar results in the car experiment. Halberstadt and Rhodes reason that these larger groups of averages aren’t different enough to show different levels of attractiveness, and that the difference between individuals and 2-item averages shows that averageness is a component of attractiveness not only for faces, but for other items as well.
Halberstadt and Rhodes conducted one more analysis of their data—they factored out the effect of familiarity. When this was done, for birds and fish, there was still a relationship between averageness and attractiveness. But there was no link in the case of cars, suggesting that we may judge attractiveness in living organisms differently than we do for artifacts.
Halberstadt, J. & Rhodes, G. (2003). It’s not just average faces that are attractive: Computer-manipulated averageness makes birds, fish, and automobiles attractive. Psychonomic Bulletin and Review, 10(1), 149-156.