Application of Generalizability Theory in Estimating the Reliability of a Motor Performance Test

Abstract

This paper discusses the use of generalizability theory, a flexible intraclass approach, for estimating the reliability of a motor performance task. The generalizability technique provides variance component estimates for each factor of an analysis of variance design. These variance estimates can be utilized to define a number of reliability coefficients which are dependent upon the logical or theoretical rationale for assigning variance component estimates as true score or error score variance. The second section of this paper presents a study illustrating the use of generalizability theory in estimating the reliability of a motor performance task. The purpose of this study was to determine the reliability of first-ball scores for beginning college bowlers. The facets of the generalizability study were identified as sex of the bowler, trials, and days. Three reliability coefficients, R ₁ (.93), R ₂ (.92), and R ₃ (.84), were computed by assigning the different factors to either true score or error score variance. The performance of beginning college bowlers as measured by first-ball scores is reliable. While the information might be useful for the teacher or researcher who is interested in detecting between-subject differences in coeducational bowling classes, the primary purpose of this paper was to demonstrate the use of generalizability theory to estimate reliability coefficients for a set of motor performance test scores.