Uma abordagem empírica de um modelo matemático para avaliação de comportamentos
Título da Revista
ISSN da Revista
Título de Volume
This dissertation intends to be an empirical an aproach to test the validity of a mathematical model for evaluation of behavior, proposed by Ziviani (1974), in its use for evaluating and quantifying answers to multiple choice questions, the traditional model of classic psychometry standing for a criterium. It is, therefore, a comparative study between the results (scores) obtain ed by this model, and by the traditional method, in terms of the right-wrong dichotomy. It is shown how variation in the measuring instrument (test), or in measurement instruments with different characteristics, deter mines variations in scores offered by both models and that these scores are the more equivalent as the 'better' is the test from which they carne.