Psychometric Evaluation and Experimental Validation of the Statistics Anxiety Rating Scale

Abstract

The Statistics Anxiety Rating Scale (STARS) was adapted into German to examine its psychometric properties (n = 400). Two validation studies (n = 66, n = 96) were conducted to examine its criterion-related validity. The psychometric properties of the questionnaire were very similar to those previously reported for the original English version in various countries and other language versions. Confirmatory factor analysis indicated 2 second-order factors: One was more closely related to anxiety and the other was more closely related to negative attitudes toward statistics. Predictive validity of the STARS was shown both in an experimental exam-like situation in the laboratory and during a real examination situation. Taken together, the findings indicate that statistics anxiety as assessed by the STARS is a useful construct that is more than just an expression of a more general disposition to anxiety.

Notes

The translated questionnaire is available for scientific purposes on request.

The average scale means of the subscales (rather than sum scores) are used to represent STARS-Anxiety and STARS-Attitudes as measures being unaffected by the different number of items in each subscale.

Some additional data were obtained that are not relevant to the research reported here.

The usefulness of considering STARS-Anxiety and STARS-Attitudes separately (instead of one global factor, which was not supported by the CFA) was further corroborated by the results of hierarchical regression analyses in which the total STARS score was entered in Step 2. The total score accounted, at best, for comparable or smaller amounts of criterion variance than the two separate scales in both validation studies.