Date of Award
Doctor of Philosophy (PhD)
Past research on the nature of small group problem solving has led to contradictory findings. Using complex problems, Shaw (1954b) found a decentralized interaction structure, i.e.,a group arrangement in which group members can communicate freely, to be most effective in reaching accurate solutions. Mulder (1960), using Shaw's complex problems, found a centralized decision structure, i.e., a group arrangement in which decisions were made by a small proportion of group members, to be most effective in reaching accurate solutions. Hutte (1965), suggested greater accuracy was found when a decentralized decision structure was used in complex problem solving. In none of these studies was there an independent and systematic treatment of the effects of decision and interaction structures.
The present study was conducted to test a proposed structural model of small group problem solving. The model, derived largely from the research of Bales (1950), and the theoretical work of Homans (1961), suggests that accuracy of problem solving and indices of group process will vary as a function of the interaction and decision structures. It was hypothesized that the degree of congruence between the centrality of the decision and interaction structures would be reflected in indices of group process and problem solving outcomes.
There were 4 conditions of structural congruence: 1) decision centrality high, interaction centrality high; 2) decision centrality high, interaction centrality low; 3) decision centrality low, interaction high; 4) decision centrality low, interaction centrality low. Conditions 1 and 4 were described as structurally congruent, and conditions 2 and 3 were described as structurally incongruent. Under each of these conditions, 6 groups, each composed of 4 Ss, solved 3 complex problems.
Statistically significant interactions for the conditions of decision and interaction centrality supported the proposed structural model. The congruence of the decision and interaction structures was reflected in both time and number of information transactions required to reach the solution to each problem.
Support for the model was not found in the problem solving outcomes of accuracy of problem solving, perceived effectiveness of problem solving, and subjective satisfaction. Contrary to both Shaw and Mulder's research, groups with decentralized decision structures were found to be more accurate in their problem solutions. However, Ss perceived a centralized decision structure as more effective in solving problems. Subjective satisfaction was highest within a centralized interaction structure.
Based on the findings of this study, it might be generally predicted that a group which contains a decentralized decision structure will be more accurate in solving complex problems than a group with a centralized decision structure. If the dimension of a decentralized interaction structure is added to this group, the participants will, in all probability, like their task more than if the interaction structure is relatively centralized. Predictions regarding time and number of information transactions to solution will be most accurately made in terms of the combinations of conditions of decision and interaction structures. The Ss will bring into the group setting certain pressures and preferences for a centralized decision structure. If allowed a choice, a centralized decision structure will probably be the result, with such a choice leading to a relatively inaccurate performance.
Harshbarger, David Dwight, "An Investigation of a Structural Model of Small Group Problem Solving" (1969). Theses and Dissertations. 3698.