CFA Covariance Question

Frequently asked questions about PLS path modeling.
Post Reply
PLS Junior User
Posts: 1
Joined: Wed Sep 18, 2019 3:42 pm
Real name and title: RIck Wolford, Ph.D Candidate in Sociology

CFA Covariance Question

Post by RickWolford » Wed Sep 18, 2019 3:47 pm

I'm constructing a measurement model that is a multilevel CFA. I have three first-order latent variables load on one second-order latent variable.
I'm adding three covariances between all first-order latent variables. All their p-values all end up coming out as 0.999 (no relationship).
HOWEVER, in the entire structural model, which includes my independent variable and my control variables, they (the first-order covariances) all end up being significant at 0.001***
ALSO, when I run just the measurement model with only two covariances, regardless of which of the two first-order latent variables I'm covarying, they also end up being significant.
I'm not sure what is going on. I suppose it wouldn't be a problem if I was only concerned about the structural model, but I have to validate the measurement first because it is a new way of operationalizing the measure.
Thanks in advance!!!

SmartPLS Developer
Posts: 1045
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: CFA Covariance Question

Post by jmbecker » Thu Oct 10, 2019 9:36 am

Are you using PLS? How do you ad covariances there? Could you show us your model?
Dr. Jan-Michael Becker, University of Cologne, SmartPLS Developer
GoogleScholar: ... AAAJ&hl=de

Post Reply