What's the robust WLS variants?

40 views

Skip to first unread message

ss xx

unread,

Jun 29, 2020, 1:32:41 AM6/29/20

to lavaan

Note that for the robust WLS variants, we use the diagonal of the weight matrix for estimation, but we use the full weight matrix to correct the standard errors and to compute the test statistic.

I don't understand this sentence.

I can use estimator="MLR" if continuous data is not multivariate normal.

But,besides add `ordered=`,what can I do when data is endogenous ordinal data and is not multivariate normal?

Terrence Jorgensen

unread,

Jun 30, 2020, 6:12:07 PM6/30/20

to lavaan

what can I do when data is endogenous ordinal data and is not multivariate normal?

ordered data already implies data are not normal, because normal data are continuous, not discrete. The robust test in the case of DWLS for ordinal outcomes is not correcting for nonnormality. It assumes normality of the latent item response(s), but it needs to be robust to the multistage estimation.

https://www.statmodel.com/download/webnotes/CatMGLong.pdf

Terrence D. Jorgensen

Assistant Professor, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam