Wasonov izborni zadatak: Utjecaj upute, tipova kondicionala i tematskog materijala

Abstract

Questions that were opened in previous research (Valerjev, 2000) are the aims of this one: First, to investigate instruction influence on results in Wason selection tasks (instruction varied from checking two cards to checking arbitrary minimum of cards). Furthermore, to investigate possible differences in answers due to the use of conditional or biconditional form of hypothesis in tasks, taking sequence effect into account. And finaly, to investigate differences in answers in respect to the material used in the task: abstract, neutral concrete, and concrete material that include the possibility of cheating (getting the benefit, without paying the price). The experiment was carried out on 64 participants that were randomly assigned to corresponding experimental situations in order to investigate the influence of four mentioned factors. Every participant solved six adequate selection tasks. Results showed that: a) there is no significant influence on results of solving sequence (conditional- biconditional and vice versa)', b) there is no (except in one case) significant difference in results between conditional and biconditional tasks, which supports the hypothesis about bias of perceiving conditionals as biconditionals; c) instruction has significant influence on results in a way that results were more stable and biases were more emphasized in the situation where instruction was to check two cards. An interesting finding was that this effect was not significant when biconditional rule was used, which can implicate that there was some interaction effect between instruction and type of conditional. Finally, d) it showed up that there was no significant difference in answers between abstract and neutral concrete tasks, but that there was significant difference in answers between concrete tasks that include possibility of cheating and other two types of used material (abstract and neutral concrete). These results are very interesting because they support Leda Cosmides’ evolution theory (1989) which claims that there exists a specific module of reasoning, which is sensitive to cheating. Results do not support pragmatic reasoning schema theory (Cheng and Holyoak, 1985) and mental model theory (Johnson- Laird and Byrne, 1991).