The prediction of item parameters based on classical test theory and latent trait theory


Anil D.

HACETTEPE UNIVERSITESI EGITIM FAKULTESI DERGISI-HACETTEPE UNIVERSITY JOURNAL OF EDUCATION, sa.34, ss.1-11, 2008 (SSCI) identifier identifier

Özet

In this study, the prediction power of the item characteristics based on the experts' predictions on conditions try-out practices cannot be applied was examined for item characteristics computed depending on classical test theory and two-parameters logistic model of latent trait theory. The study was carried out on 9914 randomly selected students who took the student selection and placement examinations applied in Turkey that is one the examinations on which try-out practices cannot be applied. In this study, the experts' prediction values were obtained from differently prepared metric scale applications according to classical and latent trait theories which were applied to 16 math teachers. In the study findings it was found that the item difficulty indexes of mathernatic sub-test of student selection and placement examinations based on experts' predictions have the capability of predicting the item difficulty indexes obtained from the classical test theory. Furthermore the predictions of bi parameters based on the experts'prediction have the capability of predicting bi parameters based on logistic model with two parameters of latent trait theory. But it is also found out that item discriminating indexes obtained from the classical test theory and ai parameters obtained from latent trait theory cannot be estimated through the experts' predictions.