Results of previous likelihood estimation study (Valerjev, 2005) showed typical biases in conditional reasoning, as well as "purification effect" which manifested in more logical answers caused by concretization of task's content. The aim of this study is the application of the same paradigm on disjunctive and conjunctive propositions. According to mental model theory (Johnson-Laird & Byrne, 1991), inclusive disjunction of P or Q form can be problematic to solvers because it requires greater number of mental models. Solvers often restrict it to exclusive disjunction of Either P or Q form because of their bias to use working memory economically. Fifty psychology students estimated occurrence likelihood of four possible cases for two disjunction and one conjunction type, each for three levels of content (abstract, concrete neutral and concrete with social contract) in 3 x 3 factorial experiment. Results showed great similarity in answers for both disjunction types which confirmed initial hypothesis. Content effect did not fully separate those two types of disjunction, but increased the difference of estimation magnitude of exclusive cases PnotQ and notPQ related to PQ and notPnotQ cases which produced certain purification effect in answers. Content effect was studied on P and Q form conjunction propositions as well. The results were surprising. While purification effect took place as a consequence of content concretization for PnotQ and notPQ cases appearing as estimation magnitude decrease, quite unexpectedly, utterly illogical significant increase of estimation magnitude occurred for notPnotQ case.

Psihologija