Reciprocal strategies for large groups.

Reciprocal strategies have been shown to be effective in inducing long-term cooperation in social dilemmas involving small, but not large, groups. In this article, we argue that group-based reciprocal strategies (GBRS)--under which one reciprocates cooperative factions, not individuals, within the group--can induce cooperation in large groups. Results of two computer simulations of 100-member groups show that a GBRS accumulates a significantly larger payoff than do other commonly advocated mixed-motive strategies. A third simulation showed that the size of the cooperative faction is influenced by the nature of the payoff matrix; when there is considerable temptation to defect, one should cooperate only after a large number of others have done so, but if there is little temptation to defect, one should reciprocate the cooperative actions of a smaller faction. We also found that the use of a "win-cooperate, lose-defect" heuristic led to substantial total outcomes. Our results suggest several psychological factors are important in facilitating cooperation, yet need empirical study.

Components of reward in social dilemmas.

A social dilemma (Dawes, 1980) may be defined as a situation in which a collection of individuals is faced with a conflict between maximizing selfish interests and maximizing collective interests. The dilemma is based on the fact that if all choose to maximize selfish interests, all are worse off than if all choose to maximize collective interests. In a decomposed social dilemma (Pruitt, 1967), the outcomes are divided into two components: one component is based on one's own choice and the second component is based on the choices made by the others. Using 3-person decomposed games, two types of incentives were contrasted: a positive incentive (bonus) for cooperative choices and a negative incentive (penalty) for noncooperative choices. Two experiments were conducted using male college students. The results of both experiments showed that the positive incentive evoked a higher level of cooperation than the negative incentive. The results are discussed in terms of nonadditive utility components, Pruitt's motivational interpretation, and Kelley and Thibaut's (1978) theory of interdependence.

An equal excess model of coalition formation.

Current theories of coalition formation in decision processes of systems at the level of the group lack generality in that some theories do not predict which coalitions are likely to form while others make predictions in only a narrow range of conditions. A model of coalition formation is proposed which predicts coalition formation in a variety of situations. Comparisons between predicted and observed results of studies using different coalition paradigms show considerable empirical support for the model. Despite some ambiguity in specifying the effects of some situational variables, the model provides a framework with which to study the processes of coalition formation and to examine the effects of such situational factors.