DISTRIBUTIONAL VERSUS SINGULAR APPROACHES TO PROBABILITY AND ERRORS IN PROBABILISTIC REASONING

Four experiments examined differences in probabilistic reasoning as a function of whether problems were presented in a frequentist or case-s pecific form. The experiments demonstrated that these different forms influence the likelihood of Ss committing the conjunction and disjunct ion fallacies. The authors contend that these 2 forms elicit different approaches to probability. Frequency problems, it is argued, elicit a distributional approach in which probabilities are equated with relat ive frequencies, whereas case-specific problems elicit a singular appr oach in which probabilities are equated with the propensities or causa l forces operating in an individual case. According to this account, d istributional and singular approaches evoke different kinds of inferen tial rules and heuristic procedures, some of which are more closely al igned with extensional principles than others.