Hello,
I would like to confirm a couple of points for structural equation modeling:
1. SmartPLS cannot be used for a model with a categorical (dichotomous) dependent variable, but I understand mPlus does.
2. Is there any precident for using PLS for the measurement model then using the output in a covariance-based structural model, for example in Mplus?
You may ask why I would not do it all in mPlus, but the model does not converge in this environment while it does in SmartPLS. Though it is probably treating the dependent variable as continuous which is not correct, I believe the measurement model should be fine.
Any thoughts?
Regards,
Michael Sheppard