Causal models of measurement: some problems for theory construction

The use of causal diagrams and interpretations to elaborate the interplay between theory and research has stimulated much interest in the application of theory construction procedures to research practice. However, when applied to efforts to combine structural models and measurement models within a single framework, the application of causal theoretical interpretations can be shown to be flawed in a way that undermines rather than enhances our understanding of the interplay between theory and empirical research. At the heart of the problem is the practice of equating the relationships in the structural model with those in the measurement model through the invocation of “causality” relationships. Abandoning this allows us to avoid the reification of theoretical concepts and provides a basis for understanding multivariate research as a process of theoretical model-building.