I am running a latent growth curve analysis. I have noticed that I get somewhat different results when I use a dataset with predictors versus a dataset without predictors. I am not including the predictors in the model (i.e., they are not in the usevariables statement), and I am running the exact same model statement. I am using MLR estimation. Is it correct to assume that this estimation technique is using information from the other (predictor) variables in the dataset, even though they are not named in the analysis command?
Only the variables in the USEVARIABLES list are included in the analysis. I would need to see the two outputs to understand what is happening. Please send them and your license number to firstname.lastname@example.org.