"My Questionnaire is Too Long!" : The assessments of motivational-affective constructs with three-item and single-item measures

Abstract

Because testing time in educational research is typically scarce, the use of long scales to assess motivational-affective constructs can be problematic. The goal of the present study was to scrutinize the psychometric properties of short scales (with three items) and single-item measures for two core motivational-affective constructs (i.e., academic anxiety and academic self-concept) by conducting systematic comparisons with corresponding long scales across school subjects and within different subject domains (i.e., mathematics, German, French). Statistical analyses were based on representative data from 3879 ninth-grade students. All short forms possessed satisfactory levels of reliability (range: .75–.89) and substantial correlations with the long scales (range: .88–.97); correlational patterns with educational student characteristics (e.g., achievement, school satisfaction, gender, academic track, and socioeconomic status) were comparable to those obtained with the corresponding long scales (all average differences in correlations below .07). The correlational patterns between all single-item measures and the external criteria were similar to those obtained with the corresponding long scales (all average differences in correlations below .08), yet the single-item measures demonstrated low to modest score reliabilities (estimated with the model-based omega coefficient; range: .22–.72) and correlations with full scales (range: .50–.88). When long scales are not applicable, short forms and perhaps even single-item measures may represent psychometrically sound alternatives for assessing academic anxiety and academic self-concept for educational research purposes.