Use of Mental Frequency Distributions to Represent Variability Among Members of Social Categories

We propose an alternative model of social category representation that suggests people use mental frequency distributions to remember and organize instances along an attribute dimension. We then explore the implications of this model for explaining differences in perceived group variability, including the out-group homogeneity effect. In Experiment 1, we demonstrate that using a greater number of categories to represent a dimension results in greater perceived variability of a group along that dimension. Experiment 2 shows that subjects spontaneously generate more subgroups (that is, a greater number of categories) along an attribute dimension when judging the in-group as compared to the out-group. Taken together, these experiments demonstrate how the use of a frequency distribution model might account for out-group homogeneity. If subjects use a frequency distribution to represent group members along a dimension, and they divide out-groups into a fewer number of categories than in-groups, then out-groups will be perceived to be less variable than in-groups.