by jhines » Wed Jan 22, 2020 6:53 pm
c-hat below zero??? Did you mean below 1.0?
The "-1.#IND00" for SE estimates is usually the result of 1) sparse data relative to model complexity, or 2) estimates at a boundary (zero or one). If it's #2, you can ignore the SE's for those parameters as it is undefined. If it's #1, then simplify the model.