The analysis of problems and prediction of group problem solving.

Abstract

The time to solve a multiple-stage problem is the sum of the times consumed by separate stages. If each stage is a random all-or-none process, then time to complete a stage is an exponentially distributed random variable, and time to complete the problem will have a gamma distribution (under suitable simplifying assumptions). From this theory, the number of stages in a problem can be estimated and goodness-of-fit tested. Results on 3 word puzzles, administered to 178 individual college students, agreed with theory, and the estimates of number of stages in each problem agreed well with independent judgments of number of stages. The same problems also were administered to groups of 4 Ss. The data suggest that all Ss progress at their usual pace toward solution, except that an S who made a mistake in interpreting the problem consumes and wastes his share of the group's time. This result agrees with the additional observation that the apparent social structure of the groups, as determined from analysis of sociometric choices, was equalitarian. (18 ref.) (PsycINFO Database Record (c) 2006 APA, all rights reserved)