Gary Chan posted on Monday, June 02, 2014 - 4:53 pm
I am very new to LCA and I would be very grateful if I can get some directions on performing LCA.
I would like to run a LCA on 6 items and would like to constrain the item endorsement probabilities to be all 0 for one class. Can anyone please tell me or direct me to an example on how this can be done?
See our UG ex 7.13 for examples of how you add constraints. You also want to read the reference we give there.
Gary Chan posted on Wednesday, June 04, 2014 - 6:22 am
Thank you very much for your direction. I have set up a LCA and I got very strange results from the Lo-Mendell-Rubin test when I run a 4-class model with one class constraint to have 0 item endorsement probabilities for all items. The mean and standard deviation was not shown. The followings are the output.
Number of initial stage random starts 500 Number of final stage optimizations 250
VUONG-LO-MENDELL-RUBIN LIKELIHOOD RATIO TEST FOR 3 (H0) VERSUS 4 CLASSES
H0 Loglikelihood Value -6232.377 2 Times the Loglikelihood Difference 35.222 Difference in the Number of Parameters 2 Mean ***************** Standard Deviation ***************** P-Value 1.0000
I also run a unconstraint model and I got an extremely large mean and standard deviation. Can you please advise why this happens? Any help would be greatly appreciated. Thank you very much.