Proceedings of the Annual Meeting of the Cognitive Science Society

In a now classic experiment Ross, Amabile & Steinmetz (1977) showed that observers think that a participant who israndomly assigned to invent questions has more general knowledge than a participant assigned to answer these questions.This is taken to be an error arising from a reasoning process in which observers ignore social roles, and instead rely onsurface behavior to make social judgments. Here we test two potential explanations for this observation: (1) observers areusing a flawed reasoning process in which they do not consider the advantages and disadvantages that different social rolesmay confer, or (2) observers are using an unbiased reasoning process in which they do consider the influence of socialrole, but they are simply operating with an imperfect estimate of the advantage afforded the questioner. In a series of fivestudies, we show that not only is reasoning in this task consistent with an unbiased inference account, but, that observersare also surprisingly well calibrated to the influence of the social roles used in this paradigm.