Hi,
I have described my study in the related question and now will follow the suggested article to generate a cutoff RLH level for random respondents.
My question is in both the simulation analysis and the actual data, how could I decide if the best and worst results are suitable to be combined and generate a single scale or should I separate them and calculate each one alone?
guess that is relative to the homogeneity somehow but I am not sure what kind or how to test it.
Regards.