1. the model selection procedure in real analyses should be replicated in simulation study to assess its validity.
2. the text seems inconsistent to title bcz dif assessment is not the focus but minor point. the author made a great deal of effort to tell model data fitting instead of dif stuff.