Osman Aksit posted on Friday, February 26, 2016 - 9:25 am
I have a question related to the class counts in Latent Profile analysis. I'm using MPlus v7. I started running the analysis with 2 classes on my data and continued until 9. Based on AIC, BIC, Adj-BIC and entropy values, 7-class model seem to be the best fit. However, when I look at the class counts and proportions, it looks like one of the classes has zero membership count. I pasted the output below.
CLASSIFICATION OF INDIVIDUALS BASED ON THEIR MOST LIKELY LATENT CLASS MEMBERSHIP
If you look at the class proportions based on the estimated model you see that class 1 does not have frequency zero. You are looking at Most Likely Class - nobody is more likely to be in class 1 than in other classes.