Effects of Explaining Anomalies on the Generation and Evaluation of Hypotheses

Abstract

We investigate the effects of explaining anomalies (i.e., observations that conflict with current beliefs) on belief revision, and in particular how explaining contributes to the rejection of incorrect hypotheses, the generation of alternative hypotheses, and the selection of a hypothesis that can account for anomalous observations. Participants learned how to rank students across courses using statistical concepts of deviation, and did so while either explaining sample rankings or writing their thoughts during study. We additionally varied whether or not candidate hypotheses about the basis for ranking were presented to participants prior to learning, and the number of sample rankings that violated intuitive misconceptions about ranking. Measures of learning and coded responses suggest that prompting people to explain can increase the rate at which they entertain both correct and incorrect hypotheses, but that explaining promotes the selection of a hypothesis that can account for anomalous observations.


Back to Table of Contents