Anonymous posted on Thursday, June 16, 2005 - 12:46 pm
I am looking for an example program in the Mplus user's guide. The research issue that I have is that I have a continuous latent factor that I want to divide into different classes that is made up of 6 continuous observed variables. For example, this variable is represented as f by x1 x2 x3 x4 x5 x6. Then I have a logistic regression model that I want to run within the classes of the continuous latent measure. That is represented by u on v1 v2 ...v10. The data are cross-sectional. What example in the user's guide will point me in the correct direction?
See Chapter 7. I think the example is called SEM Mixture. I don't have a user's guide available right now.
Chanes posted on Thursday, October 13, 2005 - 10:56 am
Dear DR Muthen When I fitted a mixture model i requested to save the posterior probabilities using the option SAVE=CPROB and specifying the file name. I want to know what is the difference between the "FINAL COUNTS... FOR THE LATENT CLASSES ON THE ESTIMATED MODEL" and the "CLASSIFICATION OF INDIVIDUALS BASED ON THEIR MOST LIKELY LATENT CLASS MEMBERSHIP". When I save the probabilities and the class membership in a file, it seems that what I get is the "MOST LIKELY LATENT CLASS MEMBERSHIP" as opposed to the "LATENT CLASSES BASED ON THE ESTIMATED MODEL". I don't understand why i have two percentages reported in the output and which of these i can use to report or to make future analyses.
When you save the CPROBABILITIES, posterior probabilities for each person for each class and each person's most likely class membership are saved. So more than one value is saved for each individual. See the summary of what is saved at the end of the output.
Besides the model estimated posterior probabilities by class, Mplus prints the classification of individuals based on their most likely latent class membership. Each row of the table represents individuals whose most likely class membership is class 1, class 2, etc. The entries in each row represent the average posterior probabilities for this set of individauls.
I'm new to Mplus. I have 3 categorical latent variables which are ordinal, thus going to run LCA. Can you direct me to an example, either in print or other that uses MPlus to determine how many possible latent classes?
See the following paper which is available on the website:
Nylund, K.L., Asparouhov, T., & Muthén, B. (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling. A Monte Carlo simulation study. Structural Equation Modeling, 14, 535-569.
See also the Topic 5 course video and handout on the website.