Latent Profile Analysis and Confounders

#1
Hi!

I need help finding resources that support the idea that we do not need to control for confounders - e.g., gender - when trying to conduct an LPA to identify latent classes of patients with distinct symptom profiles based on the measure scores over 2 months (total of 4 assessments).

Is the propensity score method necessary? We are not looking for causality.

Thanks a lot!
 

spunky

Can't make spagetti
#2
Hi!

I need help finding resources that support the idea that we do not need to control for confounders - e.g., gender - when trying to conduct an LPA to identify latent classes of patients with distinct symptom profiles based on the measure scores over 2 months (total of 4 assessments).

Is the propensity score method necessary? We are not looking for causality.

Thanks a lot!
Ok... So, why would you not want account for confounders? What if there are no profiles and the algorithm is just catching on a potential confounder/covariate that you could account for?