Re: Latent mean change from CFA to SEM

[Jeremy Miles]

On Thu, 26 Jul 2018 at 12:41 Nicolas <nicolas....@mail.dcu.ie> wrote:

Dear all,

I've been playing around with a very simple model featuring latent means. What is surprising (perhaps not so much?) is that the latent mean of one of my two factors changes from the CFA to the SEM. More precisely, the change pertains to my endogenous factor.

This is not surprising.

Any time you predict a variable with another in some sort of regression (which is what you are doing) or add another predictor, you expect the intercept to change.

The intercept of an endogenous variable is the expected value given that all the predictors are equal to zero.

Jeremy

 

--

--

My employer has nothing to do with this email.