Psicologia: Reflexão e Crítica
On-line version ISSN 1678-7153
ROAZZI, Antonio; O'BRIEN, David P. and DIAS, Maria da Graça B. B.. On the frequentist and probabilistic debate: "dumb-luck" turns out to be a plausible explanation. Psicol. Reflex. Crit. [online]. 2003, vol.16, n.2, pp. 389-402. ISSN 1678-7153. http://dx.doi.org/10.1590/S0102-79722003000200019.
A debate between Kahneman and Tversky and their associates, on the one hand, and Gigerenzer and his associates and Cosmides and her associates, on the other hand, has been fought in the area of reasoning about conditional probabilities. Kahneman and Tversky proposed that people typically represent information in terms of individual exemplars, and thus people make judgments using reasoning processes that are based on such individual exemplars. Cosmides and Gigerenzer and their associates, however, proposed that people typically represent information about population frequencies. A series of confounds in the comparisons between frequentist and probabilist problems by Gigerenzer and Hoffrage, and by Cosmides and Tooby are discussed.. Finally, we discuss a possible strategy, labeled by O'Brien, Roazzi and Dias as "the dumb-luck theory". This proposal holds that frequentist-formatted problems made available a successful guessing strategy that was not available on the probabilist-formatted problems, and the proposal opens the possibility that the entire research literature in this area has falsely assumed that correct answers stem from appropriate lines of reasoning, whereas incorrect answers do not, indicating in such a way that neither the correct nor the incorrect answers have stemmed from Bayesian lines of reasoning at all.
Keywords : Conditional reasoning; Bayesian reasoning; frequentist formatting; probabilistic formatting.