I would like to do a confirmatory latent class cluster analyses (finite mixtures) with a continuous and several categorical variables. I know how I can constrain binary variables (such as Cluster A has got more males than B) using Mplus (Thanks to "Conducting Confirmatory Latent Class Analysis Using Mplus" by W. Holmes Finch) but I do not know how to do this with a continuous outcome variable (e.g. Age in cluster A is < B < C = D). I checked the Mplus website and their FAQ/discussion sites but could not find anything.

  • $\begingroup$ The Mplus manual and website and forums are great. Did you look there? $\endgroup$ – Behacad Jan 31 '15 at 17:52
  • $\begingroup$ Can you post the code for your model without constraints? $\endgroup$ – D L Dahly Jan 31 '15 at 17:57
  • $\begingroup$ Yes, I checked the Mplus website and the FAQ/discussion sites as well as the manual (which in my opinion is rather useless if you have a very specific problem). $\endgroup$ – Stats_Monkey Feb 1 '15 at 15:43
  • $\begingroup$ My model is currently very simple: Data: File is 'c:\test.csv'; Variable: Names are a b c d e age MISSING ARE all (-999) ; USEVARIABLES ARE a b c d e age ; NOMINAL = a b c d e ; CLASSES = c(5); Analysis: Type=mixture; STARTS = 500 10 MODEL: $\endgroup$ – Stats_Monkey Feb 1 '15 at 15:46
  • $\begingroup$ See above my mplus code. I read "Conducting Confirmatory Latent Class Analysis Using Mplus" by W. Holmes Finch who describes to constrain inequalities to nominal data but I need to put an inequality constraint on my continous age variable, such as cluster a < cluster B = Cluster C <cluster D = cluster E. Hope this explains my problem better. $\endgroup$ – Stats_Monkey Feb 1 '15 at 15:49


You could include the mean in each cluster:

[age*] (age_c1) ;

[age*] (age_c2) ;

And than include a constraint:

Model Constraint:
age_c1 < age_c2 ;

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.