2026-04-13T18:26:54Zhttps://recercat.cat/oai/request

oai:recercat.cat:2445/2208222025-11-20T14:49:55Zcom_2072_1057col_2072_478801col_2072_478906col_2072_478917

00925njm 22002777a 4500 dc Tubau Sala, Elisabet author Colomé, Àngels author Rodríguez-Ferreiro, Javier author 2025-05-05T16:09:36Z 2025-05-05T16:09:36Z 2023-06-02 2025-05-05T16:09:36Z It has been shown that Bayesian reasoning is afected by the believability of the data, but it is unknown which conditions could potentiate or reduce such belief efect. Here, we tested the hypothesis that the belief efect would mainly be observed in conditions fostering a gist comprehension of the data. Accordingly, we expected to observe a signifcant belief efect in iconic rather than in textual presentations and, in general, when nonnumerical estimates were requested. The results of three studies showed more accurate Bayesian estimates, either expressed numerically or nonnumerically, for icons than for text descriptions of natural frequencies. Moreover, in line with our expectations, nonnumerical estimates were, in general, more accurate for believable rather than for unbelievable scenarios. In contrast, the belief efect on the accuracy of the numerical estimates depended on the format and on the complexity of the calculation. The present fndings also showed that single-event posterior probability estimates based on described frequencies were more accurate when expressed nonnumerically rather than numerically, opening new avenues for the development of interventions to improve Bayesian reasoning. Raonament Teoria de l'estimació Probabilitats Reasoning Estimation theory Probabilities Previous beliefs affect Bayesian reasoning in conditions fostering gist comprehension