The findings reported here indicate support for our hypotheses: restraints may have definite consequences for the patterns of differentiation in proactivity (hypothesis 1), formally imposed restraints have definite effects in reducing the amount of asymmetry found in power structures (hypothesis 2), and differentiation of proactivity is positively and strongly related to asymmetry in dyadic power relations (hypothesis 3). We were able to explain 93.5 and 94.1 percent of the variation in asymmetry for males and females, respectively. Two lines of relationship are apparent in these data. Restrictions upon the behavior of groups, as in the cyclical pattern, have an effect directly upon the differentiation in communication of proactivity, though it need not completely determine this differentiation. The differentiation which develops, in turn, has a direct effect upon the asymmetry of power relations as they develop among the dyads within a group. In the present research the effect of cyclical restraint seems to operate almost exclusively through its effect upon differentiation, though this need not hold for all situations; nevertheless, the pattern of restraint has an effect upon productivity which cannot be explained simply in terms of the amounts of differentiation and asymmetry which may be present. The apparent effect of restraint on productivity would seem to be due to the variations in skill which may be present in a group. When the group is restrained, as in the cyclical pattern, it may be more difficult for highly skilled members of groups to make their skill evident to others, or to contribute in a manner concordant with their ability. If the behavior which occurs is spontaneous, it may be easier to contribute in direct proportion to one's ability. This should lead to differentiation and asymmetrical power relations if skill is unequally distributed in a group, or to lack of differentiation and symmetrical power relations if it is equally distributed. The important factor is that the spontaneous setting leaves a group free to develop structures in accord with the distribution of ability, presumably raising their performance level. The reader should not mistakenly draw the conclusion that unrestrained situations will always result in more effective performance; the nature of the task seems important also. The task used here requires some degree of creativity and abstraction, probably making the spontaneous situation more efficient if skill is not equally distributed. Had the task been such that skill was more equally distributed, as might be the case with simple mechanical tasks such as sorting objects according to size or color, then we might find that asymmetry would disappear in both patterns of restraint, but that the cyclical groups were more efficient. This could occur because the imposed pattern of behavior might serve to eliminate or reduce the time-consuming development of norms and traditions with regard to the initiation of proactivity. Several lines of future research are suggested from the present research. Notably of interest would be variations of the present experimental situation utilizing alternative reward structures, e.g., piece rate pay, etc. It is under such circumstances that groups operating under conditions of restraint might be most likely to develop asymmetrical power structures. Other extensions of the present work might involve some division of labor in the task itself, alternative patterns of restraint, analysis of intergroup power relations, and variation in group size or communication nets. While additions of such variations certainly complicate the basic situation, they appear to promise the development of more precise and solidly grounded theory in the area of power relations and communication structures. © 1967, Sage Publications. All rights reserved.
