A simple statistical method for assessing the reliability of data entered into sport performance analysis systems

Abstract

The purpose of this srudy was to present a method for Assessing a reliability based upon Bland and Altman’s (1999) suggestions for the non-parametric treatment of comparison data and the proposals of Nevill et al. (2001) who recommended that 95% of differences should be recorded within a reference value thought to be of ‘no practical importance’. In this study, it is shown how the analyst can apply the method when individual sport performance indicators are treated as individual variables. A simple notation analysis test-retest ‘experiment’ is described, and the intraobserver data collected by one subject was used for illustrative purposes to explain how the necessary graphical, point estimate and confidence interval computations are made. The notational analyses of a further three volunteer subjects are considered in order to investigate the sensitivity of the method. The method is a simple measure of absolute agreement that is sensitive enough to distinguish between the successes and errors made by expert, experienced and improving sport performance analysts. The method is also free from the assumptions of normality, does not depend upon high within-systems variance and it can be applied successfully to individual performance indicators in the development of sport performance analysts and analysis systems.