Limited temporal stability of the Spielberger State-Trait Inventory over 3.5 years

The Spielberger State Trait Anxiety Inventory (STAI), which aims to assess state (STAI-S) and trait anxiety (STAI-T), is the most cited measure of trait anxiety and employed in a wide variety of research contexts over 50 years. Psychometric properties are key criteria for the assessment of questionnaires. For the STAI-T, data on reliability are currently limited to short time-intervals calling for investigation on the temporal stability over longer time-intervals as the questionnaire is also applied in research on long-term prediction. Here, we investigate test-retest reliability and temporal stability of the STAI-T scale across multiple measurement time-points spanning 5 months to 3.5 years in two independent samples (sample 1: 105 female Caucasians, sample 2: 120 mixed sex Caucasians). We replicate previously reported excellent psychometric properties for shorter time-intervals (i.e., < 12 month). We extend this previous work by demonstrating decreasing test-retest coefficients with increasing time-intervals pointing to limited temporal stability of the STAI-T for intervals exceeding 12 months. Of note, coefficients for the STAI-T subscale for the longest time-intervals (i.e., 41 and 32 months) approached reliability scores of the state scale, which is not intended to measure a temporally stable construct but only assesses a current state. Explorative analyses did not suggest an impact of self-reported life events between time points on temporal stability. These results relate to the debate to what extent trait constructs are expected to be stable across longer time-intervals and are of great relevance for researchers aiming to use the STAI-T scale in longitudinal investigations.