Hi everbody
I am doing a latent class analysis. Based on BIC, class size, entropy, and interpretability, I have identified three classes. I have the following code:
Now, I want to use these classes as explanatory variables to investigate different outcomes. I have read that it is best to use GSEM instead of modal class assignment (e.g. #6 in this thread). However, the "modal class"-approach is the only way I have figured out how to use the classes as explanatory variables in a subsequent regression. So, my question is: Is there a way to use the classes as explanatory variables within the GSEM framework? My dependent variable is called håbefuld_undervisning
Best,
Gustav
I am doing a latent class analysis. Based on BIC, class size, entropy, and interpretability, I have identified three classes. I have the following code:
Code:
gsem (natur tilid1_3_dum tilid1_4_dum optimistisk bange skamfuld magtesløs håb_personlig håb_miljøorg håb_Paris håb_teknologi løs_klima Tænker_på_egen_påv korrekt_viden <- ) if int_resultat == 1 [pweight=pervgt], logit lclass(C 3) nonrtolerance startvalues(randomid, draws(15) seed(123321))
Best,
Gustav
Comment