WinLTA is a free-standing Windows application for conducting LCA and LTA. Although the Methodology Center no longer supports this program, it is still available free of charge. If your organization does not permit download and installation of files, please contact us with a shipping address and we will provide a CD containing all downloadable files.
latent class analysis or latent transition analysis
multiple-groups LCA or LTA
data augmentation (DA)
Data augmentation (DA), a Gibbs sampling-based procedure, is available in WinLTA for obtaining standard errors of parameter estimates. See the following article for more information on DA:
Lanza, S. T., Collins, L. M., Schafer, J. L., & Flaherty, B. P. (2005). Using data augmentation to obtain standard errors and conduct hypothesis tests in latent class and latent transition analysis. Psychological Methods, 10, 84-100.
WinLTA (Version 3.1) [Software]. (2002). University Park: The Methodology Center, Penn State. Retrieved from http://methodology.psu.edu
Collins, L. M., Lanza, S. T., Schafer, J. L., & Flaherty, B. P. (2002). WinLTA users' guide (Version 3.0). University Park: The Methodology Center, Penn State. Retrieved from http://methodology.psu.edu
Frequently Asked Questions
Can WinLTA run under different Windows operating systems?
Yes. It should run properly in different computer environments. WinLTA software has been tested in Win98, Win2000, WinNT and WinXP.
Can WinLTA read in raw data?
WinLTA reads data in a response pattern format. Basically, items need to be recoded as 1, 2, and so on, with missing data coded as 0. The last column represents the counts for each pattern of responses. How to create response patterns is described below.
How can I create response patterns?
Most people find it easiest to use statistical software (e.g., PROC FREQ in SAS) to aggregate data. You may also use our data aggregation program, available above.
What parameters does WinLTA estimate?
WinLTA can estimate up to five sets of parameters. Latent class models involve two sets of parameters. Little rho (ρ) parameters represent the probability of a particular response to a manifest variable, conditioned on latent class membership. Gamma (γ) parameters represent the proportion of the population of interest expected to be members of a particular latent class. Latent transition models involve three sets of parameters. Delta (δ) parameters represent the proportion of the population in each latent status at each measurement occasion. (A latent status is a stage in the model.) Tau (τ) parameters are transition probabilities. They represent the probability of being in a particular latent status at Time 2 (t+1), conditional on latent status membership at Time 1 (t). Big rho (Ρ) parameters have the same meaning in LTA as they do in LCA. They represent the probability of a particular item response, conditional on latent status membership.