A non-parametric item analysis of the motivated strategies for learning questionnaire

Abstract

The quality of the Motivational Strategies for Learning Questionnaire - Chinese Version (MSLQ-CV) items was assessed with a nonparametric item analysis on a sample of 1292 Hong Kong primary and secondary students. Although a number of items within each scale needed options combined, the overall quality of these scales in discriminating between trait levels was good. Multivariate analysis on the maximum likelihood estimates of the scale scores showed significant grade and gender differences for some scales, but the effect sizes were small. Correlational analysis of these scales supported earlier research results, and a regression analysis found that four scales significantly predicted participants' scores on a standardized measure of Chinese language achievement, although the amount of variance explained was less than 5 percent. The advantages of nonparametric item analysis over traditional sample dependent omnibus measures of item and scale quality are discussed.