Parsimony in Intuitive Explanations for Behavior: Reconciling the Discounting Principle and Preference for Conjunctive Explanations

Abstract

How do people judge how much explanation a behavior requires? Attribution researchers have identified two pertinent phenomena: the discounting effect, in which participants attribute less to one cause of a behavior when informed of the presence of a strong alternative cause, and the conjunction effect, in which participants deem an explanation with two causes as more likely than an explanation with only one of these causes. Several researchers have asserted that these two effects reflect contrary modes of explanation and, consequently, that findings of widespread conjunction effects in explanation are grounds for doubting the lay person's concern for parsimony that Kelley's discounting principle implies. By contrast, we propose that discounting and conjunction effects reflect a common process of parsimonious reasoning from shared causal schemas. Experiment 1 tests our proposal concerning how these processes co-occur. Experiments 2 and 3 test unique predictions from our model about moderators of the two effects.