

Output Re: Nominal/Categorical Variables 

Message/Author 


I have been trying to use the MIXTURE routine with nominal variables and the output generated lists "means" as opposed to "thresholds" with a corresponding set of conditional probabilities. From a statistical sense, I find this confusing. Other software I have used reports conditional probabilities regardless of level of measure and particularly when variables are nominal? Can you explain what "means" is supposed to refer to, why it is generated for a nominal set of variables, why probabilities are reported only for ordered categorical variables, and (most importantly) whether I can simply produce corresponding probabilities by plugging the reported MEANS into the formula 1/1+e^b (this works for converting thresholds into probabilities for categorical variables...). 


I've answered this via Mplus support so I will not repeat the answer here. 

Back to top 

