Influence du mode de présentation de séries aléatoires sur le choix d'une stratégie : I. Situations hypothétiques
Maryla Zaleska · L’Année psychologique · 1974
Summary The purpose of the three experiments is to study subjects' beliefs concerning the best mode of response when predicting random events. Individuals are requested to describe, in writing, the strategy they would use in several hypothetical situations, each of which is a different transposition of the same basic problem. The problem consists in finding what strategy would maximize the number of elements x drawn with replacement from two separate sets A and B, knowing that the former contains 2/3 of elements x and 1/3 of elements y, and the latter, 1/3 of x and 2/3 of y. The « rational » strategy, which is the safest, consists in drawing systematically from set A since it contains always more elements x. Results of the three experiments show that the choice of this strategy depends on several factors, among others, on the extent to which the situation is artificial and unfamiliar to the individuals. The main obstacle to its adoption seems to be the difficulty of comprehending the independence of random events. It is also shown that in otherwise identical situations, simultaneous drawing from the two sets by an external agent, and substituting the action of choosing by that of predicting, lead both to significant decrease in the proportion of individuals in favor of the rational strategy. It is concluded that most people do not consider that the best strategy is to predict systematically the same event, while knowing that one third of their predictions would be incorrect. Rational strategy seems probably more justifiable to the subjects in choice situations in which lack of accuracy appears more acceptable.