Brainstorming and task performance in groups constrained by evidence

Abstract

Group brainstorming is usually considered a task of divergent thinking, and the ideas produced in most research on brainstorming are counted and scored for creativity but put to no further use. We studied brainstorming by embedding it in a rule induction task that initially requires divergent thinking but increasingly requires convergent thinking as evidence accumulates across trials. We also tested whether brainstorming facilitated performance on the induction task itself. The experimental design was a 2 (nominal or interacting groups) × 3 (brainstorming early in the task, late in the task, or none) factorial. For brainstorming performance, nominal groups of 4 individuals outperformed face-to-face groups of 4 individuals. But as predicted from an analysis of the effects of constraining hypotheses by evidence, the advantage for nominal groups declined when brainstorming took place late in the task where there was a large amount of accumulated evidence to consider. Brainstorming did not generally affect performance on the induction task, although early group brainstorming resulted in more correct hypotheses than late group brainstorming. Group brainstorming was perceived as more effective than individual brainstorming by both interacting and nominal group members, a finding that extends the illusion of group productivity in brainstorming to tasks of convergent thinking.