Structural Equation Modeling: Applications Using Mplus

3.1 MIMIC Model

MIMIC model stands for multiple indicator multiple cause model, in which multiple indicators reflect the underlying latent variables/factors, and the multiple causes (observed predictors) affect latent variables/factors. When covariance structure (COVS) is analyzed, the MIMIC model is described as (Instead of assuming only one latent variable in a MIMIC model (Bollen, 1989), we assume one measurement model with one or more latent variables in a MIMIC model):

(3.1)

where multiple endogenous indicators ys are used to measure the endogenous latent variables s; no causal effects, but the covariance/correlations, are specified among the s because all latent variables are in one measurement CFA model; and s are affected by observed predictors xs, which are assumed to be perfect measures of the exogenous latent variables s (e.g., respondent self-reported gender status is often treated as a measure of his/her sex identity without measurement error).

When the MACS is analyzed, the ...