Outcome measurement and the use of Rasch analysis, a statistics-free introduction
Kersten, P; Kayes, NM,
MetadataShow full metadata
Outcome measures, which use questions or assess performance on tasks are increasingly used in clinical practice. A key principle of such measures is that of internal construct validity. This is a characteristic that is best evaluated with Rasch analysis. This approach to analysis tends to be described in the literature in a statistical and technical manner, not easily accessible to people new to measurement research. This paper focuses on concepts and interpretation of key messages in an attempt to de-mystify Rasch analysis for the practicing clinician. The paper first explains the basic tenets of the Rasch model. This is followed by explanation of the principles of the key analytical stages involved in a Rasch analysis. The paper demonstrates that the examination of internal construct validity, using Rasch analysis, involves various qualitative and quantitative judgments. The main strength of the Rasch model lies in its theoretical and scientific underpinning. Outcome measures that fit the Rasch model are more robust than those that don’t and arguably, the latter should not be used in practice or research.