Introduction and application of Dichotomous Moken Scale Analysis (MSA) in educational and psychological measurements

This study aims to introduce this scale for Dichotomous Responses and shows the application of it in the field of educational measurements/assessments by providing evidence of fitting Mokken in study the level of knowledge of the fifth- and sixth-year students of Kerman dental school (academic year 95-96) about blood pressure (BP)measurement technique according the guideline of the American Heart Association.

Mokken scale analysis (MSA) is a probabilistic-nonparametric approach to Item Response Theory (IRT) that can be used to evaluate fundamental measurement properties in educational and psychological measurements with less strict assumptions than parametric IRT models.

The results of secondary data analysis of the test included 14 items and all of 131 students, showed (using R software), that, the scalability coefficients of each item and the scalability coefficients of the item pairs were positive with low standard error. The overall scalability coefficient was 0.490 (medium), and omission of item 5 increased the overall scalability coefficient to 0.54 (strong/high). Therefore, the results under the monotone homogeneity Model (MH) showed no violations of monotonicity and one-dimension was obtained. Invariant item-ordering Results from IIO analyses showed the number of violations from the assumption of non-intersecting response functions, bottom-up method for selecting items confirmed the deletion of item 5. Obtaining the HT scalability coefficient of 0.45, was an evidence of fitting double monotonicity (DM) model. Four reliability coefficients calculated for the set of 13 items ranged from 0.86 to 0.91.

This study showed that this scale can be used to rank both individuals and items.