Font Size: 

In latent class (LC) analysis a standard assumption is conditional independence, that is the indicators of the LC are independent of the covariates given the LCs. We compare the likelihood ratio based MIMIC test to residual statistics (BVR and EPCinterest) for identifying nonuniform direct effects (DEs) of covariates on the indicators of the LC model. The simulation study results show that the LR test and EPCinterest correctly identifies direct effects more often than the BVR.