Measurement error and reliability testing: Application to rehabilitation

Abstract

Aims: Effective measurement of clinical outcomes is dependent on reliable outcome instruments. Measurement error and reliability testing are fundamental underpinnings of reliability. This article defi nes and illustrates sources of measurement error, outlines strategies for error minimization, and gives an overview of the types of reliability studies. Content: The two main sources of measurement error: systematic bias and random error, are discussed. The three major types of reliability evaluation are then illustrated: test-retest; intra-rater and inter-rater reliability, and the relationship between reliability and validity is explained. Discussion and conclusions: Quantifi cation of measurement error is integral to determining the true effect of therapy, as quantifi ed by outcome measurement. Interpretation of reliability data involves consideration of many factors including demographic, diagnostic and clinical characteristics of the study sample.