Premium
Prediction and choice in the Prisoner's Dilemma
Author(s) -
Halpin Stanley M.,
Pilisuk Marc
Publication year - 1970
Publication title -
behavioral science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.371
H-Index - 45
eISSN - 1099-1743
pISSN - 0005-7940
DOI - 10.1002/bs.3830150205
Subject(s) - dilemma , prisoner's dilemma , psychology , task (project management) , social psychology , sequence (biology) , adversary , block (permutation group theory) , statistics , cognitive psychology , mathematics , economics , geometry , management , biology , genetics
This study utilized identical false feedback to subjects in five conditions. In two conditions the subjects played a standard two‐choice prisoner's Dilemma game (200 trials); in two conditions the subjects played the standard game modified to include a prediction of Other's behavior on each trial. In each of the above situations, subjects in one condition were told that they were faced with a preset sequence while subjects in the other condition thought they were playing with one of the other subjects present in the experiment. In a fifth condition the feedback was not given any meaning as part of the game task; subjects merely had to predict the appearance of one of two lights on each of the 200 trials. The 100 subjects were assigned randomly to conditions with ten males and ten females in each condition. Analysis of the frequency of cooperative choice in the game indicates that the addition of prediction to the game had no significant main effect on choice behavior. There was a three‐factor interaction effect with the description of the feedback and the trial block of the game. Subjects tended to decrease cooperation over trial blocks. The act of prediction reduced this effect for subjects who thought they faced a preset computer sequence but not for those attempting to predict the moves of a human adversary. The use of prediction data in combination with choice data produced four indices of behavior contingent upon expectation of opponent's behavior. These indices were useful in interpreting the higher cooperation of the females as an inability to find or maintain optimal strategies for maximizing their own rewards. An examination of the prediction data alone revealed that a probability matching hypothesis by itself will be inadequate to explain subjects' trial by trial expectations of Other's behavior.