Twenty questions: efficiency in problem solving as a function of size of group.

Abstract

The game of "Twenty Questions" was employed in an expt. on problem solving. A total of 105 subjects were assigned by chance to solve problems working either alone, in pairs, or in groups of 4. Each individual or group was given 4 problems a day for 4 successive days. On the 5th day, all subjects worked alone, each being given 4 problems. In terms of number of questions, rapid improvement occurred in the performance of both individuals and of groups. By the 4th day, the curves appeared to be flattening out. Similar results were obtained in terms both of number of failures and of time per problem. Group performances were superior tc individual performance in terms of number of questins, number of failures, and elapsed time per problem; but the performance of groups of 4 was not superior to that of groups of 2, except in terms of the number of failures. In terms of man-minutes required for solution, the performance of individuals was superior to that of groups; the performance of groups of 2 was superior to that of groups of 4. Improvement in individual performance occurred as rapidly with individual practice as with practice as a member of a group.