Evaluating the reliability of multi-item scales: A non-parametric approach to the ordered categorical structure of data collected with the Swedish version of the Tampa Scale for Kinesiophobia and the Self-Efficacy Scale

Abstract

To compare the ability of a rank-invariant non-parametric method with that of kappa statistics to evaluate the reliability of the Swedish version of the Tampa Scale for Kinesiophobia and the Self-Efficacy Scale by identifying systematic and random disagreement. The aim was, further, to compare 2 different statistical approaches to obtain a global value from multi-item scales. A test-retest study. A total of 46 patients with whiplash-associated disorders were enrolled and 39 (85%) completed the test-retest assessment. Data from the multi-item scales were summarized using both sum and median scores. Paired data were evaluated with a rank-invariant statistical method to identify systematic and random disagreement. Data were also evaluated with kappa statistics. The non-parametric approach demonstrated that the Swedish version of the Tampa Scale for Kinesiophobia and the Self-Efficacy Scale are reliable for patients with whiplash-associated disorders. In contrast to the rank-invariant method, kappa statistics provided no information on disagreement between the 2 test occasions. Median scoring improved reliability due to lack of disagreement while the sum scores method was characterized by random individual disagreement. This study has increased understanding of the advantages and limitations of 2 non-parametric statistical methods and, it is hoped, will contribute to the development of reliable measurements.