Hi SEM and Mplus community!
I think this question is rather basic but unfortunately I can not find a solution to this issue:
In Mplus 8 I estimated a linear second order growth curve model with the two latent growth curve factors INT and SLOPE which works fine so far. Now I want to introduce two independent/exogenous manifest variables (X1 and X2, both are dichotomies, coded 0 vs 1) in the model to explain the (now endogenous) latent INT and SLOPE factors (command lines are: INT on X1 X2; SLOPE on X1 X2;). This also works fine and I know that X1 and X2 empirically correlate r=.254 (p
According to the Mplus support team, this message is "benign" (does not signify a serious issue or problem) and is merely due to the fact that your IVs are binary (non-continuous). I would not worry about it as long as you get the same parameter estimates as before (i.e., in the model version in which you did not include the X1 with X2; command line).
Dear Christian,
thanks this is really a relief!
I compared the two outputs and they are numerical identical (model fit and estimations). The only difference is in the information criteria (AIC, BIC, adj BIC) which are much higher now than the model without the X1 with X2 command line.
greetings,
Gert
Hey Gert
Maybe the information criteria are different because X1 and X2 are now 'in the model' - are the number of parameters to be estimated the same for the models?
Mark
- Badges
- Science topic
- Similar topics
- Indexes