Divergent thinking tasks are a popular basis for research on group creative problem solving, or brainstorming. The brainstorming literature has been dominated by research that investigates group performance by measuring the total number of generated ideas using the original rules put forth by Osborn (1953). This review of empirical literature on brainstorming suggests that Osborn was right about many but not all of his intuitions. It discusses the potential advantages of cognitive stimulation, and potential disadvantages of conformity or fixation, due to exposure to others' ideas. Further, although Osborn recommended "withholding criticism," the potential advantages of conflict in interacting problem-solving groups are also discussed. In addition, this review suggests that incorporating performance measures of idea quality, in addition to the ubiquitous measures of idea quantity, can provide useful and unique insights into group brainstorming.