Influence and performance in simultaneous collective and individual induction

Abstract

A four-person cooperative group (majority) and four separated individuals (distributed minorities) solved the same rule induction problem simultaneously in different rooms. The group and individuals exchanged hypotheses and evidence on each trial (HE), hypotheses only (He), evidence only (hE), or neither (he). The individuals did not exchange with each other. A formal analysis distinguishes (1) influence by correct, plausible, and nonplausible hypotheses; (2) instigative influence (when individual and group disagree on trial t); and supportive influence (when individual and group agree on trial t); and (3) individual influence on the group, group influence on the individual, and mutual individual and group influence. The results supported five predictions. First, the order of the four exchange conditions was HE > hE > He > he for group instigative influence on the individuals by correct hypotheses and HE = He = hE = he for individual instigative influence on the group by correct hypotheses. Second, there were higher probabilities of instigative influence by correct hypotheses than by plausible hypotheses and very low probabilities of instigative influence by nonplausible hypotheses. Third, mutual supportive influence by correct hypotheses was a virtual absorbing state, but there were very low probabilities of mutual supportive influence by plausible and nonplausible hypotheses. Fourth, the aggregate proportion of correct hypotheses of the group and four yoked individuals was increased by the exchange of evidence but not by the exchange of hypotheses. Fifth, groups and individuals followed comparable strategies of confirmation and disconfirmation of hypotheses.