Studies among children and adults showed that valid syllogistic problems (Modus Tollens and Modus Ponens) are easier to solve than invalid ones (Affirmation of Consequent and Negation of Antecedent) where subjects’ performance is very low. In this study we tried to diminish the great number of errors in the invalid problems using another alternative for the responses and counter-example. University subjects obtained a great number of correct responses in the invalid forms, however their performance was low in the valid forms. The adults with low education and the illiterates did not seem to be influenced by the control variables introduced, showing a better level of performance in the valid forms then in the invalid ones, as others studies had shown.
Syllogisms; valid and invalid form; invalid conversion
