
Message/Author 


This question is about analysis of formative indicators of an emergent construct (in this case, different indicators of childhood maltreatment). First, can such indicators be factor analyzed using MIMIC? Second, given the different handling of error terms for formative indicators, can factors scores from the MIMIC be used as one of a set of x variables in a SEM or growth mixture modeling analysis without adjusting the model? In this case, the other x variables are either observed demographics or other standard latent variables such as personality traits. The Y variable in this case is a latent variable representing alcohol dependence. 


You cannot regress a formative factor on a covariate. The covariate would become one more factor indicator. You can regress a regular factor on a formative factor. 

Back to top 

