Rules Comm posted on Tuesday, May 04, 2010 - 8:24 pm
I know many people have asked about this topic. My question is different.
In one related discussion, you suggest a student to compare the models with and without outliers. What is the purpose to do this? If both models work well, at least at the facial level, what can we do in the comparison?
The question was how to run the model with and without outliers. The answer was that you would have to do the analysis twice using the USEOBSERVATIONS option to select cases that are not outliers. This was not a recommendation.