Applying the Rasch Model as a Diagnostic Tool for Rating Scale Refinement

Abstract

The Rasch model has been widely used in educational, psychological, and health research to evaluate the measurement quality of instruments. In many applications, however, Rasch analysis is primarily reported to support validation or confirm the adequacy of a scale. Although diagnostic analyses may be conducted during instrument development, their role in informing substantive instrument refinement is less explicitly documented and therefore less visible in the literature. This study aims to demonstrate how the Rasch model can be applied as a diagnostic tool to support the refinement of rating scales. Using empirical data from an attitude scale, the study illustrates how detailed Rasch outputs—such as item fit, response category functioning, and threshold ordering—can be interpreted to identify specific sources of measurement problems. These insights provide a basis for targeted revisions, demonstrating how Rasch analysis can contribute not only to validation but also to iterative instrument refinement. This study contributes to the methodological literature by highlighting a more comprehensive use of the Rasch model that integrates validation and diagnostic purposes. It also provides practical guidance for researchers, particularly those less familiar with Rasch analysis, on how to use model outputs to improve measurement instruments, especially rating scales.