Structuring as an Aid to Performance in Base-Rate Problems (Report No. 84-16)

Abstract : Four groups of college students were each given two base-rate problems. Three of the groups were given an aid with the first problem: 1) An instruction to list factors or aspects that were relevant to solving the problem; 2) a fill-in-the-blank algorithm that provided the correct solution; or 3) a seven-page tutorial that explained base-rate problems and showed how to solve them using a 2 x 2 table. No aid was provided for the second problem. The control group replicated previous findings in disregarding the base-rate information. The list factors group showed no improvement over the control group. The algorithm group showed distinctly better performance for the first problem but were the same as the control groups for the second problem. The tutorial group did best: 42% of answers to the first problem and 31% of answers to the second problem were within + .10 of the correct answer. An error analysis identified a conceptual weakness in the tutorial; a high rate of arithmetic errors was also found. College students appear to lack the knowledge needed to solve base-rate problems but they can be taught this knowledge relatively easily. Keywords: Statistical reasoning, Cognitive psychology, Decision analysis, Subjective probability estimation. (KR)