Reliability and validity of the spielberger's state-trait anxiety inventory (STAI) in serbian university student and psychiatric non-psychotic outpatient populations

Abstract

Anxiety is an inborn emotional and cognitive reaction that enables humans to cope with everyday situations, however, it can be pathological as well. Although Spielberger's State-Trait Anxiety Inventory (STAI) scales are most frequently used to study the presence/absence of anxiety, there is a lack of empirical studies in Serbian healthy and/or diseased population that would indicate its reliability and validity. The aim of the present study was to determine some normative values for STAI scales (state and trait) and to determine its reliability and validity in the studied population of university students and psychiatric patients. Score values obtained from the two analyzed samples, which did not show a normal distribution, showed to be statistically significantly different. Receiver operating characteristic curve analysis revealed that values of area under the curve for the two scales were > 0.85 (p < 0.001) and cut-off values suggested that the presence of anxiety was determined to be 38.5 for the S-scale and 45.5 for the T-scale. Data analysis revealed, according to the intraclass correlation coefficient, that S-scale had a moderate reliability (Cronbach alpha = 0.713) and that T-scale had a good reliability (Cronbach alpha = 0.858). Also, STAI scales given to patients suffering from anxiety disorders were determined to measure the presence of anxiety with coefficient alpha higher than 0.9. The calculated validity, internal consistency, and reliability for the translated STAI version are satisfactory, meaning that the usage of the translated STAI can clearly measure the changes in the level of anxiety in patients suffering from neurotic, stress-related, and somatoform disorders.