Choose Language Hide Translation Bar
Level I

Comparing LCA results

I recently finished a LCA analysis as a replication project. My question is regarding the two different LCA (and there for my sample) and the ability to compare the two.
While the original project had around 2000 entries and resulted in the LCA with 3 groups, my project consisted of around 500 entries, and the LCA resulted with 2 groups. In both cases the variables used for the LCA are identical and the main difference (other than group size) is the cultural variance.
Can I somehow compare the two LCA results (BIC/AIC/etc) in such a way so I can add something to the difference between the groups and to make the argument of variability between the group stronger?
Or as the two LCA are originated from different samples there is not much to compare other than the number of latent groups results in the analysis?
(if anyone wants to see the results/original project, its in the OSF)

Re: Comparing LCA results

You cannot use information criteria based on likelihood, such as AICc or BIC, to compare models with a different set of responses, such as different samples or due to a transformation.

Learn it once, use it forever!
Article Labels

    There are no labels assigned to this post.