A Czech version of the TAI-G: Psychometric properties and factorial structure

In adapting the Test Anxiety Inventory (TAI; Spielberger, 1980) to German language, several modifications have been proposed. Special attention was given to problems that arise from situational descriptions. To avoid arbitrary combinations of situational and response characteristics, situational references in the items were eliminated and a modified instruction referring to feelings and thoughts in examination situations were introduced. Items were reformulated and several new items based on theoretical classification were added. The revised German Test Anxiety Inventory (TAI-G. Hodapp, 1996; Hodapp, Glanzman, Laux, 1995) was evaluated by means of factor analyses, item analyses, and Rash analyses and yielded four subscales: Worry, Emotionality, Lack of Confidence, and Interference. The results provide evidence of the psychometric adequancy of this instrument. A Czech version of the TAI-G was tested by a sample of 304 university students (121 males, 183 females). The Czech version of the TAI-G provides evidence for a clear four factor solution which corresponds to the orginal structure. Subscales of the Czech TAI-G show satisfactory alpha-coefficients. Confirmatory factor analyses (CFA) support the four structure of the TAI-G. With a reduced 20-item form, CFA primary and second-order factor models fit the data. Analyses accross the Czech and German samples of university students show similar subscale means, invariance for factor structure and loadings, and differences in factor variances and item residuals. The Czech TAI-G is an appropriate adaptation. Further research is needed to investigate sources for differences across national sample as well as construct validity and prediction of performance and other behavioral outcomes of test anxiety.