Base Rates, Representativeness, and the Logic of Conversation: The Contextual Relevance of “Irrelevant” Information

Abstract

According to the cooperative principle of conversation that governs social discourse in everyday life, listeners expect speakers to be relevant, truthful, and informative. In studies on judgmental biases, researchers frequently violate this principle by presenting information that is neither informative nor relevant in a communicative context that suggests otherwise. However, subjects have no reason to doubt the relevance of the presented information and try to make sense of it, as they would be expected to do in everyday life. In Experiment 1, the applicability of the cooperative principle was varied to explore the impact of conversational principles on the apparent overreliance of individuals on nondiagnostic person information at the expense of base-rate information. Nondiagnostic person information was presented either as a statement written by a psychologist or as a random sample of information drawn by a computer. As predicted, subjects relied on the personality information rather than on base-rate information to a greater extent in the former than in the latter case, presumably because a human communicator (but not a computer) is supposed to conform to conversational norms and to provide in formation that is informative, truthful, and relevant. In addition, subjects relied more on individuating information when the framing of the task implied that psychologists provided correct estimates than when it implied that statisticians provided correct estimates and when the individuating rather than the base-rate information was varied as a within-subjects factor (Experiment 2).