Estimating school climate traits across multiple informants: an Illustration of a multitrait–multimethod validation through latent variable modeling

Abstract

The use of multiple informants is common in assessments that rely on the judgments of others. However, ratings obtained from different informants often vary as a function of their perspectives and roles in relation to the target of measurement, and causes unrelated to the trait being measured. We illustrate the usefulness of a latent variable multilevel multitrait–multimethod measurement model for extracting trait factors from reports of school climate obtained by students (N = 45,641) and teachers (N = 12,808) residing within 302 high schools. We then extend this framework to include assessments of linkages between the resulting trait factors and potential outcomes that might be used for addressing questions of substantive interest or providing evidence of concurrent validity. The approach is illustrated with data obtained from student and teacher reports of two dimensions of school climate, student engagement, and the prevalence of teasing and bullying in their schools.