Testing is essential in education and other social science fields because many assessments, decisions, and policies are made according to the results of testing. The purpose of testing is to estimate a person's ability, that is, latent trait or construct. In a test setting, responses to a set of test items by each individual are recorded. Through a scoring scheme, test scores are assigned to individuals according to their item responses. Test scores therefore provide information from which we infer a person's ability. In this study, combinations of different item response theory (IRT) models and dichotomous versus polytomous scoring models were compared. The different IRT models yielded different ability estimates. The polytomous models, in which each of the categories of response is evaluated and scored according to its degree of correctness or the amount of information provided toward the full answer, provided more accurate ability estimates than items that were dichotomously scored.