The effectiveness of using Rasch Model in calibration of Test Descriptors for the Learning and Thinking skills Course Applied to a sample of undergraduate students at King Faisal University

Comparative Study with Classical Test Theory

Abstract

The aim of the study is to identify the psychometric characteristics of the Learning and Thinking Skills Test on a sample of 162 students of the college of Education, King Faisal University, using Classical Theory, as well as determining the psychometric characteristics of the same test using the Rasch model to verify the possibility of comparing students' grades in the test within both theories.

The results indicate that the number of items used to test the skills of learning and thinking differ according to the theory used in analysis. The classical theory showed a thirty-item test, while the Rasch model produced a thirty-two -item test. The calibration of the test items does not change according to changes in sample achieving ability (high – low). Seven items were omitted according to Rasch analysis. The ability parameters for persons ranged between (-4.03) logit and (4.04) logit.