On the (dis-) confirmability of stereotypic attributes.

Two experiments with 154 17-82-yr-olds investigated the linguistic abstractness and confirmability of elements contained in ingroup and outgroup stereotypes. The first experiment shows that positive elements of the ingroup stereotype (Italians) and negative elements of the outgroup stereotype (Jews, Germans) tended to be particularly abstract. Also, negative elements contained in the outgroup stereotypes required relatively little evidence to be considered "true" but much disconfirming evidence to be rejected as "false." No such bias emerged for ingroup stereotypes. The second experiment compared the abstraction of four outgroup stereotypes (Jews, Blacks, homosexuals, career women) finding the greatest abstraction for the oldest stereotype (Jews), and least abstraction for the most recent stereotype (career women) with the remaining two groups (Blacks, homosexuals) occupying an intermediate position. Results are interpreted as suggesting that stereotypes may become more abstract over time as they lose the concrete elements that are easier to disconfirm while maintaining the abstract elements that are more resistant to chan