An Alternative To Classic Test Theory? An Exemplary Application Of Generalizability Theory-Based On Secondary Data

Background Applying repeated measurements, errors can arise from multiple sources, which affects the outcome. In order to determine these sources of error, an expanded statistical model like the generalizability theory framework is required. In contrast to the classic test theory, it enables the simultaneous investigation of several sources of error.Objective To demonstrate an exemplary application of generalizability theory in a physiotherapy setting based on an isometric muscle strength test as well as the comparison of two different approaches, the classical test theory and the generalizability theory.Method Examinizing the clinimetric property Reliability" based on data of a previous reliability study in which classic test theory was used. For this purpose, an exemplary generalizability study was conducted.Results Variances resulted mainly from subjects and interactions between subjects and raters. Generalizability coefficients were excellent for all conditions (overall, inter-rater, inter-time point), regardless of the type of facets (fix or random facets).Conclusion The results of the exemplary generalizability study shows that generalizability theory has advantages over classic test theory. Generalizability theory offers an approach to determine and implement effective and efficient use of measurement protocols for physiotherapy practice.