Author:

Abstract

Why do participants commit the conjunction fallacy, for instance by judging it more probable that Linda is a feminist bank teller than a bank teller? The conversational-implicature hypothesis (CIH) suggests that “bank teller” is interpreted as “non-feminist bank teller”. The nested-sets hypothesis (NSH) suggests that participants overlook that the set of bank tellers includes all feminist bank tellers. Both hypotheses were tested in an experiment with 157 participants. The results, analyzed using Bayes factors, indicated that the CIH manipulation does not robustly decrease the fallacy rate (B_{H(0, 1.52)} = 0.14, OR = 0.84). Furthermore, the effect of the NSH manipulation was substantially smaller than predicted (B_{N(3.04, 1.52)} = 0.1, OR = 1.47), suggesting that NSH does not explain the fallacy.

Keywords: conjunction fallacy, subjective probability judgment, conversational implicature, nested sets