tidyLPA - including latent variables/constructs (like CFA)

[Dyck, Stefan]

I came across a question using tidyLPA r-package.

Is there a possibility to include latent variables/constructs (e.g., satisfaction measured by 4 items) for comparing and estimating the latent profiles (similar to latent variables used in structural equation modeling; a CFA measurement model)?

Example:

u11-u23 are the manifest/measured variables (items)
c1 + c2 are latent variables / constructs
f is the Latent Class/Latent Profile grouping variable

Best Regards
Stefan Dyck

[Joshua Rosenberg]

Stefan, please excuse my very late response - dissertation ended up sucking all of my time! But, it is done now.

So, I think this is a very good thing to include, at least for the functions that use MPlus. Do you have access to MPlus? I think this is something we can try to include int he next month or so, but it would only be available for the functions that interface to mplus - i.e., estimate_profiles_mplus(), but not estimate_profiles().

Thank you for this idea and for the question,

Josh

[ddis...@gmail.com]

Hi Josh,

I was wondering if this functionality was ever added (i.e., factor mixture models). I am looking to do LGMM with Mplus and am considering tidyLPA (+ MplusAutomation) for my workflow. Does tidyLPA currently allow for mixing on latent variables (e.g., intercept and slope factors)? I was browsing the package website (https://data-edu.github.io/tidyLPA/index.html) but could not find a clear answer.

Best,

David