Collective induction: Mutual group and individual influence by exchange of hypotheses and evidence

Abstract Yoked four-person groups and individuals induced a rule under four conditions of information exchange. The yoked groups and individuals exchanged both hypotheses and evidence on each trial, hypotheses only, evidence only, or neither. Both the exchange of hypotheses and the exchange of evidence improved the aggregate performance of the yoked group and individual. Group performance was superior to individual performance. Significant three-way interactions indicated that this superiority was greatest when evidence, but not hypotheses, was exchanged. Alternatively, the interactions indicated that exchange of evidence had relatively more effect than exchange of hypotheses for groups, whereas exchange of hypotheses and exchange of evidence had comparable effects for individuals. The patterns of same and different individual and group hypotheses on trials t and t + 1 indicated more group influence on the individual than individual influence on the group. Social combination analyses of the group hypotheses for distributions of member hypotheses indicated that the same basic social combination process applied in all four conditions. Sequential transition analyses of the transition probabilities from distributions of member hypotheses on trial t to distributions on trial t + 1 indicated that once the correct hypothesis was proposed by one or more group members it was highly likely to be proposed again by one or more members on the subsequent trial. We propose that the inductive task and yoking methodology will be useful for research on majority and minority influence.