

TECH11  Replication of H0 Loglikelih... 

Message/Author 

Jen posted on Friday, December 20, 2013  1:14 pm



Hello, I am having issues replicating the H0 loglikelihood value for the K1 solution in the TECH11 output when comparing 5 and 6class solutions in a LCA. I have read "Using Mplus TECH11 and TECH14 to test the number of latent classes" and generally had good luck with the described method. I have no problems replicating the loglikelihood values for either the c(5) or c(6) solutions themselves (with a modest number of starts, such as STARTS = 400 80), yet no matter how high I put the K1STARTS for c(6) (I have gone up to 5000 1000), I cannot replicate the c(5) loglikelihood value in the TECH11 output (the loglikelihood for the c(5) = 5974.33, but in TECH11 it consistently = 5989.43). Should I worry about this, and is there anything else I might try? Thank you! 


I assume this is a regular "exploratory" LCA so that the first class  which is dropped when doing the k1 class model in the kclass run  doesn't have any parameter restrictions. You can try stscale=10 or 20 or 30. Mplus drops the first class of the 6th class solution – this gives the starting values for the 5th class solution. If there is a some kind of perfect indicator in the 6class solution that is not there for the 5class solution this can happen. If you can match the classes between the 5class and the 6class solution, put the extra class as class one (you can do that by using stvalues or different optseed for the 6 class solution) so that the extra class come up as class 1 in the 6class solution. 

Jen posted on Monday, January 06, 2014  8:50 pm



Thank you! stscale fixed the problem immediately, with a much lower k1starts. 

Pia H. posted on Friday, August 11, 2017  6:23 am



Hello, I am running a factor mixture analysis and am facing a similar problem as described above. This is my input for a 3 class model: ANALYSIS: TYPE = MIXTURE; STARTS = 8000 2000; ALGORITHM = INTEGRATION; PROCESSORS = 32(STARTS); MODEL: %OVERALL% f1 BY PCL1 PCL2 PCL3 PCL4 PCL5 PCL6; f2 BY PCL7 PCL8 PCL9 PCL10 PCL11 PCL12; [PCL1$1PCL12$2](124); %c#1% f1f2; f1 with f2; [f1f2@0]; %c#2% f1f2; f1 with f2; [f1f2*]; %c#3% f1f2; f1 with f2; [f1f2*]; The loglikelihood is replicated > 10 times. However, the loglikelihood I get for this model slightly differs from the k1 loglikelihood obtained with TECH11 in the next step (not only for 3 vs. 2 classes, but also for 4 vs. 3 and 2 vs.1). The differences are quite small (i.e. 3409.149 vs. 3409.151). Are these differences negligible? Thank you! 


Yes, with numerical integration involved I would say this difference is negligible. 

Back to top 

