|
|
Covariates vs. DVs in latent class an... |
|
Message/Author |
|
Jon Elhai posted on Saturday, October 23, 2010 - 12:56 pm
|
|
|
I tried to post this question yesterday, but it didn't seem to post correctly. I'm wondering what the difference is between using y variables to estimate latent classes in an LCA, vs. using covariates to estimate the classes. In other words, is there a practical difference if you take a covariate in LCA and instead of using it as a covariate, you use it as one of your y variables? |
|
|
The difference is that among the covariates there is no assumption of independence given class membership. |
|
Back to top |
|
|