RMSEA vs. SRMR / ULS vs. ML

[In-Sue Oh]

Dear SEMnetters,

I have two questions. I am conducting meta-analytic path analyses
based on a meta-analytically derived correlation matrix and testing
several alternative models using LISREL.

In this case,
which index is better to use to compare alternative models, SRMR or
RMSEA or both, above and beyond CFI, TLI, and chi-squares?

Further, which estimation method is better to use, ULS or ML? FYI, the
sample size is quite large, more than 2000. Is ULS really OLS?

Any thoughts on this would be great appreciated.

Thank you so much,
In-Sue
--
In-Sue Oh, Doctoral Student, MA in I/O Psychology
Department of Management and Organizations
W321, 108 Pappajohn Business Building
The University of Iowa
Iowa City, IA 52242-1000

Office: 319-335-1809 ACT: 319-339-3043 (only Friday)

--------------------------------------------------------------
To unsubscribe from SEMNET, send email to ***@bama.ua.edu
with the body of the message as: SIGNOFF SEMNET
Search the archives at http://bama.ua.edu/archives/semnet.html

[Brent Coker]

In-Sue,

There are probably more complex answers to your questions, but I think you should:

(1) Report RMSEA and SRMR. But if you have to choose, report RMSEA over SRMR.
(2) Use ML.

Brent.


-----Original Message-----
From: Structural Equation Modeling Discussion Group on behalf of In-Sue Oh
Sent: Mon 18/09/2006 13:19
To: ***@BAMA.UA.EDU
Subject: RMSEA vs. SRMR / ULS vs. ML

Dear SEMnetters,

I have two questions. I am conducting meta-analytic path analyses
based on a meta-analytically derived correlation matrix and testing
several alternative models using LISREL.

In this case,
which index is better to use to compare alternative models, SRMR or
RMSEA or both, above and beyond CFI, TLI, and chi-squares?

Further, which estimation method is better to use, ULS or ML? FYI, the
sample size is quite large, more than 2000. Is ULS really OLS?

Any thoughts on this would be great appreciated.

Thank you so much,
In-Sue
--
In-Sue Oh, Doctoral Student, MA in I/O Psychology
Department of Management and Organizations
W321, 108 Pappajohn Business Building
The University of Iowa
Iowa City, IA 52242-1000

Office: 319-335-1809 ACT: 319-339-3043 (only Friday)

--------------------------------------------------------------
To unsubscribe from SEMNET, send email to ***@bama.ua.edu
with the body of the message as: SIGNOFF SEMNET
Search the archives at http://bama.ua.edu/archives/semnet.html

--------------------------------------------------------------
To unsubscribe from SEMNET, send email to ***@bama.ua.edu
with the body of the message as: SIGNOFF SEMNET
Search the archives at http://bama.ua.edu/archives/semnet.html