The Student–Teacher Relationship Scale revisited: Testing factorial structure, measurement invariance and validity criteria in German-speaking samples

• CFA for the German STRS did not confirm the original US factor structure.
• EFAs and CFAs suggested a parsimonious 3-factor model for the German STRS.
• The dependency scale had to be adjusted to the German culture.
• The STRS showed invariance across gender and context (kindergarten and elementary school).
• Using SEMs, construct and criterion validity were shown to be satisfactory.

The Student–Teacher Relationship Scale (STRS) is widely used for research in kindergarten and school. The increasing number of applications inside and outside of the U.S. stresses the need to investigate STRS properties, accordingly. The present study used the STRS in German-speaking countries, examining whether (a) the original factor structure is appropriate for a German version, (b) whether applications of a German STRS are invariant across contexts (kindergarten, first and second grade) as well as gender, and (c) whether construct and criterion validity are met. The original STRS was translated into German and filled out by 368 kindergarten and 503 elementary school teachers in Germany and Austria. Observations in kindergartens, student reports in schools, and teacher reports of students’ characteristics served as validity criteria. Results of confirmatory factor analyses (CFAs) did not confirm the original STRS factor structure. Subsequent exploratory factor analyses on training samples resulted in significant item reductions, followed by further CFAs on validation samples. The bootstrapped results yielded an adjusted three-factor model with subscales indicating satisfying alphas and invariance across context and gender. Construct and criterion validity were met for all subscales of the German STRS based on various criteria from both, observations and reports.