Simple but powerful models of stereotype formation

We propose simple models aiming at explaining the formation of stereotypes. A stereotype is an overall judgement brought by an observer over a group of individuals or objects. For each member of the group, we suppose that we observe a characteristic that belongs to a denumerable set. The formation of a stereotype about the group is governed by a perception function. Our basic model consists in decomposing a perception function into three steps: (i) characteristics are recoded numerically so that higher numbers mean a higher support for the stereotype, (ii) this vector of numbers is consistently aggregated into a single number, and (iii) this number is compared to a threshold and the stereotype is accepted if the threshold is exceeded. We characterize perception functions that can be explained using such a model. We then study various extensions of our basic model.