Four-person groups responded to a four-choice sequential decision task for three blocks of trials. Decision patterns were analyzed in terms of strategies (a plan for distributing choices across a 50-triaI block) and social decision schemes (majority, plurality, equiprobability, and highest expected value) for selecting among member strategy proposals. Predictive models, assuming each decision scheme in turn and using parameter estimates from an independent sample of individuals, were compared with group data on each block. By the third trial block, all of the social process models considered, except equiprobability, could be rejected as inaccurate. Such a finding, though consistent with other research and informal interaction data from this study, was unanticipated for a sequential response task.

No abstract

Dyads discussed and reached a decision about either of two social issues: the percentage of university control that should be invested with students and the percentage of the national budget that should be spent on pollution control. For each issue, three different group compositions were created by assigning persons who had been differently obligated by constituencies to whom they were responsible. The summary obligation was in the form of a distribution of preferences about the proportion of constituents favoring each of the decision alternatives. Several stochastic models, assuming different social decision schemes, were derived and tested against the dyadic decision distributions. The idea that members moved through equal distances to a midpoint between them received the most support, but some cases favored an alternative idea; namely, each member was equally likely to win the decision for his position. Both decision schemes could be regarded as variations on the notion of an equalitarian social process.