A Three-Dimensional Theory of Group Process in Adolescent Dyads

Abstract

This dissertation tests a three-dimensional theory of group process originally proposed by William Schutz (1958) . His theory is that three process variables can account for group interaction: Inclusion, the degree to which persons in a group feel "in," "a part of" the group; Control, the degree to which persons can command and direct the group's resources, means, and goals; and Affection, the degree of relatedness that persons in the group feel for one another. Eighty-nine tenth grade suburban high school students completed a sociometric rating of their intact homeroom classes, and twenty-four pairs of students were randomly selected to participate in the experimental portion of the study. The dyads were selected along the Inclusion and Affection dimensions, each at two levels. Each pair played eight ten-choice games of "Prisoner's Dilemma, " a two-person, two-choice nonzero sum game, under an experimental instruction set of "trust and cooperation." The eight payoff matrices were systematically varied to provide two levels of Asymmetry and two levels of Fate Control, which are taken as the operational equivalent of the Control dimension. The matrices were randomly ordered for each pair. The design is a 2^4 factorial with repeated measures over two dimensions, analyzed as analysis of variance. The data is analyzed only for those matrices which give less payoff ("go against") the first player in the dyad to make a choice, since these matrices alone offer an incentive to trust the partner. There are six dependent variables in this study: (1) one's own number of trusting choices in each ten-choice game; (2) the partner's number of trusting choices; (3) one's total estimate of the partner's trustworthiness; (4) one's total number of years in jail; (5) the partner's total number of years in jail; and (6) the combined number of years in jail for both players. The results show a significant effect only for Fate Control, and only on three dependent variables: (1) total estimate of the partner's trustworthiness; (2) one's total number of years in jail, and (3) the partner's total number of years in jail. In general, the level of trusting behavior was high across all experimental conditions. The results are only partial support for the theory of group interaction. Fate Control is the one operational dimension most clearly linked with the experimental task demands, and therefore cannot be seen as strong support of Schutz's theory, especially in view of the lack of significant results on any other dimension. Affection, Inclusion, and Asymmetry of the payoff matrix were not significantly associated with any dependent variable. Second, factors beyond the experimenter's control may have contributed to the null results. For instance, students may have been "loyally" trusting to other students at a very high level perhaps because of their role vis a vis adult authority as manifested by the experimenter. Moreover, an overall lack of interpersonal interaction in the homeroom setting may have attenuated the results. Third, there is wide variance for each of the dependent variables, small effect size, and, consequently, the heightened chance of a Type II error. Moreover, the dependent variables are highly correlated, further limiting the potency of this experimental test. Finally, Schultz's theory is one of process, and the variables used in this study can capture this process only insofar as the dyad's structure reflects the process. To the extent that the structural measurements used in this study may not fully reflect palpable interpersonal process, the experiment, not the theory, may be held deficient. In summary, this attempt to empirically assess this three-dimensional theory of group process is not wholely successful. The experimental analogue situation (the Prisoner's Dilemma) gives only partial support to the theory.