Behavioral and Perceptual Differences Between Structurally Equivalent, Two-Person Games

This research primarily examines the effects of “naturalness†of context upon (a) game-related trusting behavior, (b) the cognitive predictors of trusting behavior such as the other's trustworthiness and perceived intentions, and (c) subjects“ attributions of causality within the experimental gaming situation itself. Since reward level and trust-worthy first impressions could possibly moderate the effects of context, a 2 × 2 × 2 (rich versus poor context x 83% versus 50% reward level x good versus bad first impression) factorial design was used in which 60 male Yale undergraduates responded to a series of 12 games (n = 40 for the rich; n = 20 for the poor). The †rich“ context was created with 12 verbal communication choice dilemmas. The †poor“ context consisted of 12 numeric outcome games, verified to be structurally equivalent to the verbal games. More trusting behavior was observed in the poor context (83% versus 62%). Negative first impression in the rich context only was a powerful inhibitor of initial, first-trial trusting behavior (5% versus 63%). A rich context produced more overall perceived trustworthiness and less bad intentions. High versus moderate reward and good versus bad first impressions produced more overall trustworthiness and more perceived good and less bad intentions. Subjects in the rich context perceived the experimenter to be biased against competitive subject behavior; subjects in the poor context perceived the experimenter to be biased for cooperative behavior, in particular, and for any activity in general. Additionally, a context by reward interaction indicated that subjects alter their perceptions of the appropriateness of reward/betray responses depending upon the type of context. When the response was appropriate for the context, causality tended to be attributed to the experimenter; when inappropriate, to the hypothetical other with whom the subjects were †interacting.“ The function of role expectations is discussed. Seven problems to monitor or control in future research are derived.