Abstract
Influence of training on deductive reasoning: Importance of content and transfer between domains. This research examines the type of teaching material more adequate for the students to bypass the "affirmation of the consequent fallacy" (AC), frequently underlying the erroneous causal conclusions among theories and observed phenomena. We present two experiments in which type of domain: easy and difficult, training in AC (yes/no), and domain in which the test is carried out (same/different) are manipulated. Results indicate that content difficulty worsen performance, but that the specific training, particularly in difficult domains, is better than the general training. It appears that transfer occurs only when the test task is in a domain easier than the one used in training. The data contribute with new evidence in favour of theories of mental models.
