Group membership and communication in a prisoner's dilemma setting



Journal Title

Journal ISSN

Volume Title



This study was concerned with the effects of group related and communication measures on cooperative choice, expected cooperative choice, ratings of trust, and attitude ratings in a 20 trial Prisoner's Dilemma game allowing simultaneous written communication over half the trials. Forty-eight volunteers in nine groups from a single psychology class at the University of Houston interacted, either with members of their own group (Intra), or with members of other groups (Inter). Half were allowed to communicate by simultaneous note the first 10 trials, the other half were unable to communicate until the last 10 trials. The game was cast in terms of disarmament proceedings in an arms race. Measures of cohesiveness and group functioning were obtained from all class members prior to the game. Factor Analysis of the sociometric produced three factors bearing some resemblance to the basic needs (control, affection, and inclusion) postulated by Schutz. Analysis of the data revealed the following. 1) The Intra-Inter group variable was too gross a measure to be meaningful in a situation where membership in that group had no particular validity, but 2) when one's cohesiveness with his own group was included, the expectation of cooperative play was significantly influenced. High cohesive Intra and low cohesive Inter dyads tended to expect more cooperation than high cohesive Inter and low cohesive Intra dyads. Furthermore, considering only the very high and very low cohesive scores, even Play seemed to be influenced. 3) Individual scores on the three sociometric factors, when ranked by dyads, provided significant distinctions in game play. Factor I (control) interacted with Group, and Factor III (inclusion) was itself, highly significant. Factor II (affection) was unfortunately confounded with communication so its significant effect could not be interpreted. 4) Communication, in this study, led to lower levels of cooperation. 5) Communication failed to resolve this dilemma. 6) Gaining or losing communication after trial 10 failed to make other than temporary changes in game measures. Subjects apparently were unable to 'take advantage' of the tool of communication. 7) The combined play of the dyad on Trial 1 was significantly related to what happened on subsequent plays, even across the communication change. 8) Use of the three dependent variables. Play, Expectation, and Trust, provided useful tools for understanding much of game play. 9) Attitude ratings failed to reveal any meaningful data relating to 'mixed motive' games. 10) Balanced play shows promise of being a meaningful indicator of cooperative behavior in groups.



Prisoner's dilemma game