James T. Tedeschi

Matrix indices and strategy choices in mixed-motive games

Rapoport (1964) reported the effect on proportion of cooperative strategy choices over iterations in the Prisoner‘s Dilemma gamc (PD) of varying single matrix payoffs. Oskamp and Perlman (1965) derived a payoff index which was thc average payoff per trio2 and found higher values of the index related to larger proportions of cooperation. They mentioned, but did not test, another index which measured the average adoaritage of competition. Rapoport and Orwant (1962) suggested that the difference bctwccn the payoff for successful defection and unsuccessful cooperation-called thc competitive advantag-might be related to cooperative strategy choices in PD. hlinas et al. (1960) manipulated payoffs so as to vary thc dominance of the cooperation and defection strategies. Althoudi no index was formally defined, it is possible that the average advantage of competition would describe changes in dominancc of the cooperation or defection strategies. Since the formal rules of PD require that the defection strategy be dominant, it is not technically correct to define dl of the hiinas ct al. (1960) matrices as PD matrices, since both equal-dominance and coopcrntive-dominant matrices were used. Hotvever, some

Original stakes and behavior in the prisoner’s dilemma game

A 5 by 2 by 2 design was used to investigate the effects in the PDG of giving an advantage to one of the players of an original stake of 0, 25, 50, 100, or 500 points under two levels of conflict intensity. The 216 Ss were assigned to like-sex dyads. The results showed that the 25 point OS Ss cooperated more and as dyads produced fewer DD outcomes than the other OS Ss. The high conflict condition produced more defections, more DD outcomes, fewer CC outcomes, and fewer CD outcomes than the low conflict condition. Finally, female dyads had more CC outcomes than males.

Behavior of a threatener: retaliation vs. fixed opportunity costs

Harsanyi (1962) has suggested that costs are usually incurred by the wielder of power in his attempts to enforce compliance in the social influence process. For example, should one country attempt to impose its will upon another country by an act of war, certain costs related to the loss of men and materials would be incurred. Harsanyi refers to these losses as opportunity costs. Opportunity costs constitute, in effect, the self-imposed price paid by source as a function of his own influence attempts; as such, these costs are fixed and relatively voluntary. On the other hand, if the target of the influence attempt possesses retaliatory power, additional costs may be imposed upon the source which are neither voluntarily incurred, nor certain of occurrence, from sources point of view. The behavior sequence when the target is recalcitrant in a unilateral threat situation is threat, noncompliance, and punishment; but in a bilateral threat situation the sequence is threat, noncompliance, punishment, and retaliation. The additional target behavior of retaliation in a bilateral threat situation should either exacerbate the conflict or deter the sending of threats and/or their enforcement, and thus provide opportunities for conciliatory behavior. In either case, the target possessing retaliatory capacity should be perceived as more potent than the target who does not possess such power. In a relevant study, Deutsch and Krauss (1960) had pairs of subjects play a bargaining game in which they were told that each was an operator of a trucking company and that operating costs would be assessed on the basis of the time taken to deliver goods to a fixed destination. Conflict between the subjects was generated from the layout of available delivery routes; the shortest and hence most profitable route was a one-lane road which only one subject might traverse at a time. Given the basic paradigm, three conditions of threat were employed by providing gates to one, both, or neither of the subjects, which could be lowered to prevent the other from reaching his destination. Thus the three conditions of no threat, unilateral threat, and bilateral threat were established. The ability of subjects to reach agreements, as reflected by the magnitude of their joint payoffs, was inversely related to the amount of threat present. The power of the adversary is not easily determinable by the threatener in the trucking game studies. There is the possibility, for instance, of retaliation by the subject not possessing gates; he can park his truck in the one-lane road and thereby prevent the use of the shorter route by the subject who possesses a gate. There was also the uncertainty of retaliatory actions and their associated costs in the bilateral threat condition. Thus the opportunity costs to the source of threats are

Aggressive behavior of delinquent, dependent, and “normal” white and black boys in social conflicts ☆

Abstract The Prisoners Dilemma Game was modified by giving an option to the subject after every seventh iteration of the game of taking

The effects of opportunity costs and target compliance on the behavior of a threatening source

10 from a dummy other than at a cost to himself of 2, 5, 8, or

The effects of conflict intensity and promisor credibility on a target’s behavior

11. A 4 × 3 × 2 orthogonal design was used, with one dimension representing the four opportunity cost levels. The second dimension consisted of adjudicated delinquent, adjudicated dependent, and “normal” preadolescent boys of 10–13 years of age. The third dimension was race. Each subject played 50 trials against a dummy, who played a preplanned 50% cooperative strategy. It was found that black subjects cooperated more and won less than white subjects. There were no differences related to frequency of exercising the option of punishing the “other.” Only analyses based on the cue of the dummys cooperation yielded differences in aggression. When the dummy had cooperated on the trial preceding the option, black subjects were more aggressive than white subjects, dependents were more aggressive than delinquents, and delinquents were more aggressive than “normals.” White subjects were more sensitive to the costs of using their power than were black subjects, and “normals” were more sensitive to costs than delinquents. Dependent children were not affected by opportunity costs. A cue-personality-aggression hypothesis was used to interpret the data.

Start effect and response bias in the prisoner's dilemma game

Abstract Subjects were given options to send threats to a simulated target in a modified Prisoners Dilemma. Manipulated variables included three levels of opportunity costs for threat enforcement and three levels of compliance by the target. Results showed that the sending of threats was deterred when the costs were greater than the anticipated gains from target compliance. However, threat enforcement was not affected by costs, indicating that once a commitment was made by sending a threat, enforcement costs had already been taken into account. No effects of target compliance were found.

Relative costs and reactions to threats

Noncontingent promises of varying credibility were made to Ss in conditions of high and low conflict intensity established by manipulating the payoff values in the Prisoner’s Dilemma Game (PDG) payoff matrix. Highly credible promises produced more cooperation on the message trials than did low credibility promises. Ss concealed their choice intentions more when credibility was low. Greater overall cooperation, trust, and repentence were displayed when the conflict intensity was low. Conflict intensity interacted with credibility on the proportion of forgiveness shown. Females were more repentent than males when promise conditions were compared to controls.

The ethical shift in the Prisoner’s Dilemma Game1,2

In a Prisoner’s Dilemma Game it was found that location of response switch did not bias the strategy selections of Ss. No DD outcome occurred for any pair of Ss on the first trial, but initial trial CC outcomes were followed by more cooperations than were initial trial CD outcomes. Females displayed more “trust” than did males.

Motivational Correlations of Strategy Choices in the Prisoner's Dilemma Game

A 3 by 3 by 2 design was employed in which 72 male Ss were assigned across three levels of a linearly transformed Prisoner’s Dilemma Game (PDG) pay off matrix, three levels of punishment magnitude, and two levels of threat credibility. Occasional threats of a loss of points were sent by a simulated player during 150 PDG trials. Unanticipated effects of both matrices and punishment magnitude were obtained on overall cooperation. An effect of matrices was obtained on compliance to threats. Postimpressions of the simulated source were that the 10% credible threatener was perceived as strongly potent, while the highly credible source was considered mildly impotent.