The Conjunction and Disjunction Fallacies: Explanations of the Linda Problem by the Equate-to-Differentiate Model

Abstract

We propose the use of the equate-to-differentiate model (Li, S. (2004), Equate-to-differentiate approach, Central European Journal of Operations Research, 12) to explain the occurrence of both the conjunction and disjunction fallacies. To test this model, we asked participants to judge the likelihood of two multi-statements and their four constituents in two modified versions of the Linda problem in two experiments. The overall results underpin this pragmatic model's inference and also reveal that (1) single conjunction and disjunction fallacies are most prevalent, (2) the incidence of the conjunction fallacy is proportional to the distance between the constituent probabilities, and (3) some participants misinterpreted A ∧ B either as ¬ A ∧ B or A ∨ B. The findings were generally consistent with the configural weighted average model (Nilsson, H., Winman, A., Juslin, P., & Hansson, G. (2009), Linda is not a bearded lady, Journal of Experimental Psychology: General, 138) and the potential surprise conceptual framework (Fisk, J. E. (2002), Judgments under uncertainty, British Journal of Psychology, 93).