Democracy under uncertainty: the wisdom of crowds and the free-rider problem in group decision making.

We introduce a game theory model of individual decisions to cooperate by contributing personal resources to group decisions versus by free riding on the contributions of other members. In contrast to most public-goods games that assume group returns are linear in individual contributions, the present model assumes decreasing marginal group production as a function of aggregate individual contributions. This diminishing marginal returns assumption is more realistic and generates starkly different predictions compared to the linear model. One important implication is that, under most conditions, there exist equilibria where some, but not all, members of a group contribute, even with completely self-interested motives. An agent-based simulation confirmed the individual and group advantages of the equilibria in which behavioral asymmetry emerges from a game structure that is a priori perfectly symmetric for all agents (all agents have the same payoff function and action space but take different actions in equilibria). A behavioral experiment demonstrated that cooperators and free riders coexist in a stable manner in groups performing with the nonlinear production function. A collateral result demonstrated that, compared to a dictatorial decision scheme guided by the best member in a group, the majority/plurality decision rules can pool information effectively and produce greater individual net welfare at equilibrium, even if free riding is not sanctioned. This is an original proof that cooperation in ad hoc decision-making groups can be understood in terms of self-interested motivations and that, despite the free-rider problem, majority/plurality decision rules can function robustly as simple, efficient social decision heuristics.

Anxiety, reactivity, and social stress-induced cortisol elevation in humans.

OBJECTIVES: Traditionally, it has been hypothesized that highly anxious/emotionally reactive subjects may have exaggerated social stress response. We examined the relationship between self-reported anxiety, emotional reactivity, and social stress response. METHODS: We investigated the relationship between personality scales of trait-state anxiety, subjective autonomic reactivity, and salivary cortisol levels before and after social stress exposure (Trier Social Stress Test) in 20 men. RESULTS: Significant positive correlations between anxiety, subjective autonomic reactivity, and basal cortisol levels were observed, while neither anxiety nor subjective autonomic reactivity was correlated with social stress-induced cortisol elevation. CONCLUSIONS: The present results indicate (i) subjects with higher degrees of trait anxiety/subjective autonomic reactivity have higher basal cortisol levels, and (ii) in contrast to the traditional view, anxious personality is not strongly associated with exaggerated cortisol response to social stress.

Interpersonal trust and social stress-induced cortisol elevation.

A neuroendocrine correlate of interpersonal trust is relatively unknown. We investigated the relationship between an interpersonal trust-related personality (General Trust Scale) and cortisol elevation induced by social stress in 20 men. Spearman's rank order correlation analysis revealed a significant negative correlation between social stress-induced cortisol elevation and General Trust Scale. The present results indicate subjects with higher degrees of interpersonal trust have lower levels of neuroendocrine response to social stress.

Social stress-induced cortisol elevation acutely impairs social memory in humans.

Although stress is known to affect the memory process, little has been elucidated regarding its effect on social memory. In this study, 30 subjects' social memory (face-name association memory) performance was tested. Twenty subjects were exposed to social stress with a subsequent test of social memory, while the remaining ten subjects served as controls. Free cortisol was determined via saliva samples taken before and after the social stress exposure. Because stress hormones might have biphasic actions on neurons, we separated the subjects with cortisol elevations between high and low responders. The high responders had significantly impaired social memory. Furthermore, we observed a negative relation between social stress-induced cortisol elevation and social memory. These results indicate that social stress acutely impairs social memory.