Reliability Generalization of Scores on the Spielberger State-Trait Anxiety Inventory

Abstract

A reliability generalization study for Spielberger’s State-Trait Anxiety Inventory (STAI) was conducted. A total of 816 research articles utilizing the STAI between 1990 and 2000 were reviewed and classified as having (a) ignored reliability (73%), (b) mentioned reliability or reported reliability coefficients from another source (21%), or (c) computed reliability for the data at hand (6%). Articles in medically oriented journals were shorter and somewhat less likely to mention or compute reliability than nonmedically oriented articles, perhaps due to paradigm differences. Average reliability coefficients were acceptable for both internal consistency and test-retest, but variation was present among the estimates. State test-retest coefficients were lower than internal consistency coefficients. Score variability was predictive of internal consistency reliability for scores on both scales. Other predictors were the age of research participants, the form of the STAI, and the type of research design.