теория игр (two person mixed-motive games и experiments with prisoner dilemma and related games )

And there are several types of behaviourof the subjects determined by their character traits: unconditional cooperation, unconditional defection and variants of combinations of C and В choices. Also there are several interesting patterns in the alternation of the combination. There two main interesting combinations, that were called reformed sinner (ALL D, then ALL C, then TFT choices) and lapsed saint (ALL C, then TFT choices). Interestingly enough that the reformed sinner subjects made more C choices than lapsed saint subjects.The new life into PDG experiments was breathed by the computer tournaments organized by Robert Axelrode (7.6). His experiment shows some more new strategies of behavior in such games. Axelrodу formulates several more efficient scenarios and strategies: do not envy, do not defect first, do not be too clever, reciprocate both defection and cooperation. According to 7.7 there were some unexpected results concerning the gender distribution of the subjects. Women showed much lower levels of cooperation in both the PDG and Chicken. It can be easily explained by the idea that В choice is less dangerous and women just behaved more cautious, but it can explain only the PDG experiment, not Chicken. There is another possible explanation for both games. Women chose D, because they’d rather lose the explicit payoff than seem foolish to the other participant. There are series of studies devoted to the attribution effects (7.8), i.e. on the effects of people’s beliefs about their co-player’s intentions during PDG. There is plenty of scope of attributional errors. For example, the C choice can be interpreted as cooperative, but in fact it is only a mean to lure the co-player into making a C choice, and the В choice not only can be competitive, but also self-protective. Also the behaviour of the subjects can be influenced by the frame of the game. A number of studies showed higher levels of cooperation in decomposed PDGs than in conventional PDGs with the equivalent payoffs. Decomposition of this type provide the frame that makes plain the players’ dependence on each other’s good will and generosity, and this evidently facilitates the development of mutual trust and trustworthiness. Several studies of framing effects are devoted to the comparison of subjects’ choices in matrix, abstract games and choices in more life-like games. The major finding was that subjects are significantly more cooperative in abstract games, though there are several peculiarities that make it unsafe to draw any conclusion from this result. Anyway the results of experiment designed to investigate framing effects have shown that frames make a difference. The interpretations that subjects are induced to make of the decision problems of the games have large and predictable effects on the choices that they make. This conclusion makes doubtful make earlier research findings made on the matrix games. But also it shows that there is still a rich seam of information about framing effects in experimental games waiting to be mined by future researchers.The concluding part of the chapter is devoted to summing up the contents of the chapter and to underline the main aspects of it.