Collective induction: Effects of multiple hypotheses and multiple evidence in two problem domains.

Abstract

This article proposes that collective induction is improved more by multiple evidence than by multiple hypotheses. In Experiment 1, 4-person cooperative groups solved rule induction problems. In 9 conditions, they proposed 1, 2, or 4 hypotheses and selected evidence on 1, 2, or 4 arrays, on each trial. Correct hypotheses increased with multiple evidence but not with multiple hypotheses. Conversely, nonplausible hypotheses increased with multiple hypotheses but not with multiple evidence. In Experiment 2, 4-person cooperative groups solved a random mapping of the letters A-J to the numbers 0-9. On each trial, they proposed 1 or 2 equations in letters (e.g., A + D = ?), learned the answer in letters (e.g., A + D = G), and proposed 1 or 2 hypotheses (e.g., A = 5). Performance was improved by multiple equations (evidence) but not by multiple hypotheses.