Anonymous posted on Monday, September 06, 2004 - 8:49 am
I am dissertating doctoral student who's fairly new to MPLUS and am in need of help. I am trying to run a CFA on a two factor model (one with 4 observed variables, the other with 6). My data is quite nonnormal and I have been advised to use the Satorra Bentler chi-square correction for non-normality. I used the Estimator = MLM in the analysis command, but I get an error that says " Estimator MLM is not available for analysis with non continuous variables." I am using MPLUS version 2.13.
Do I have the correct command? I appreciate any help you can provide.
Anonymous posted on Monday, September 06, 2004 - 8:54 am
And I forgot to mention in my previous message that not only is my data non-normal, each observed variable is also categorial (responses were on a 4 point Likert scale).
The best estimator for you is WLSMV. The is the Mplus default for categorical dependent variables.
Anonymous posted on Monday, September 06, 2004 - 10:04 am
Thanks for your reply. I am sorry to belabor this (I am very new to CFA as well as MPLUS), but I don't understand why the WLSMV is a better estimator for my nonnormal categorical data instead of Satorra Bentler as I was told. Can you explain why or point me in the direction of an article that would explain this?
Thanks very much again for your help.
bmuthen posted on Monday, September 06, 2004 - 10:30 am
The Satorra-Bentler chi-square is defined for continuous outcomes, not categorical ones. You can read more about this from the viewpoint of WLS estimators in my "paper #75" (see Mplus Reference list) which you can request from email@example.com.
Anonymous posted on Tuesday, September 07, 2004 - 6:49 am
Is there another email address where I may request this article? The email address above was undeliverable.