Formative and Reflective Measurement Models in Workplace Learning Research: A Critical Review and Guidelines

Abstract

ABSTRACT Quantitative empirical research in workplace learning regularly makes use of measurement instruments that operationalize theoretical constructs. However, in domains related to workplace learning, high levels of measurement model misspecifications—up to 62% of all cases—have been identified. In most cases, formative measurement models, where indicators cause the construct, are erroneously specified as reflective, where the construct causes the indicators. This can significantly bias estimates, distort structural relationships, and, consequentially, compromise the validity of research findings. However, in workplace learning research there is currently no evidence available regarding the level of measurement model misspecification. To address this need for research, we systematically reassess 325 unique measurement instruments derived from 92 studies identified in a prior qualitative meta‐synthesis. Using two distinct coding strategies—a conservative coding and a majority coding—we evaluate the extent to which formative measurement models have been misspecified as reflective. Results indicate that between 34.8% and 50.2% of instruments are misspecified in this way, depending on the coding approach. Based on this, we provide guidelines to support researchers in selecting and evaluating measurement models. Our study advances the methodological rigor of workplace learning research and contributes to the broader discourse on improving construct validity in basic and applied research contexts.

The authors' abstract, as published at the source. Human Resource Development Quarterly, 2026 · DOI ↗