Size, performance, and potential in brainstorming groups.

Abstract

Using a 3X2X2 factorial design, five-, seven-, and nine-man brainstorming groups were compared to groups, composed of individuals who brainstormed alone. The third factor was due to the use of two £s (a male and a female). The criterion was total number of nonoverlappin g ideas produced in each of the conditions. There was no main effect due to £s nor were there any interactions. As expected, there was a significant effect due to size and type of group. The larger groups produced more ideas and the nominal groups were more effective than the brainstorming groups. Contrary to our prediction that the nominal and brainstorming groups would converge as size increased, they diverged and the interaction was significant. The authors concluded that group brainstorming, over a wide range of group sizes, inhibits rather than facilitates creative thinking, and pooled individual effort is a far more productive procedure than group effort. A number of studies (Bouchard, 1969; Dunnette, Campbell, & Jaastad, 1963; Taylor, Berry, & Block, 1958) have compared the productivity of real brainstorming groups with or statistical groups of individuals who brainstormed alone and had their nonoverlapping ideas combined. The logic of this comparison is that the nominal groups represent the expected level of productivity if group brainstorming neither facilitates nor inhibits creative thinking. All of the studies have concluded that group participation inhibits rather than facilitates creative thinking, and that pooled individual effort, or some combination of individual and group effort (Dunnette, 1964), is a far more productive procedure than group effort alone. One of the drawbacks of all these studies is that none of them has systematically varied group size. Each of the studies has used an N of 4. Osborn (1957) has suggested that the optimum size of brainstorming groups is between 5 and 10. Bouchard (1969) has suggested that there may be more rapidly diminishing returns in nominal groups than real groups, that is, the growth curve for nominal