Why Using the Rasch Model to Analyze Educational Tests and Psychological Inventories?

Abstract

The scales in human sciences, usually linear transformations of raw scores for educational tests or psychological inventories, have been criticized as qualitative and subjective rather than quantitative and objective. The lack of interval scales makes parametric statistics inappropriate. To achieve objective interval scales, Rasch (1960) proposed a measurement model, which has been extended by several researchers to more complicated testing situations. In this talk, I will introduce why the derived scales from the Rasch model and its extensions are objective and interval, whereas those from other multi-parameter item response models are not. The recent trends and applications of the growing Rasch family and my research interests will be briefly addressed.