The measurement model is the part of an SEM that contains latent variables and their indicators. Educationists and psychologists familiar with tests and measurement will understand this intuitively. The measurement model is also known as the Confirmatory Factor Analysis Model and Null Model. Paths from latent variables to the indicators are modeled, as are the paths from the error terms to the measured variables. In these types of models, there are no paths (direct effects) between the latent variables specified.
The structural model is model Components connecting unobserved (latent) components with each other. Model specification errors refer to using the wrong or an inappropriate model. This can happen in several ways: Omitting relevant variables, including irrelevant variables, postulating linear model when nonlinear more appropriate, and introducing measurement error are among the most common.
Key Terms
Structural Equation Models, error terms, latent constructs, indicator variables, observed variables, structural regression coefficients, recursive, non-recursive, exogenous variable, direct effect, indirect effect, total effect, goodness of fit, fit indices, Goodness of Fit Index (GFI), Adjusted Goodness of Fit (AGFI), parsimony, Parsimony Goodness of Fit Index (PGFI), comparative fit indices, Normed Fit Index (NFI), Comparative Fit Index (CFI), independent model, saturated model, path coefficients, standardized structural coefficients, measurement model, model specification errors
Last Modified: 06/04/2021