ULSMV and EFA PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
 cathy labrish posted on Monday, December 12, 2011 - 8:36 pm

If this is a repeat post, please forgive me but I do not see a message I posted last Friday.

In addition to checking scree plots, and performing parallel analyses to determine number of factors to extract, I also check various fit indices. When using WLSMV, I use the RMSEA, CFI, TLI, chi-square and SRMR fit stats.

There are times when I might try a ULS estimator. To my knowledge, no other package except CEFA produces RMSEA or chi-square when using ULS. I note that ULSMV produces all the fit stats produced when using WLSMV.

I have checked the technical appendices. Except for a description of how ULSMV computes df, I can find no other reference to how ULSMV works. Aside from the calculation of df, are there any differences between ULSMV and ULS that I should take note of, or should the two provide identical results.

Thank you!
 Linda K. Muthen posted on Tuesday, December 13, 2011 - 9:50 am
ULS and ULSMV will give the same parameter estimates. ULS does not give standard errors or fit statistics. Only ULSMV does that.
 claudio barbaranelli posted on Wednesday, March 18, 2020 - 7:36 am
I am trying to locate where is the formula for USLMV estimator in EFA. I did not find it in the technical material dowloaded from the website.
I am curious to know if it is similar to the one for the so called "ALPHA" FACTOR EXTRACTION (see e.g. Mac Callum et al., 2007, equation 9.4 p. 170).
I run a factor analysis problem both under SPSS (EXTRACTION = ALPHA/ROTATION = PROMAX), and under MPLUS (ESTIMATOR = ULSMV, ROTATION = GEOMIN, CATEGORICAL), and obtained almost the same results.
So, I am wondering whether these two estimator are sort of equivalent, and if ULSMV in MPLUS allows you to perform the Kaiser's ALPHA factor analysis
 Bengt O. Muthen posted on Wednesday, March 18, 2020 - 12:08 pm
Appendix 4 of the Version 2 Tech appendices that we post on our web site at


shows the formulas (105) -(108) which are relevant to ULSMV with W=I. I am not familiar with Kaiser's alpha method - you may want to ask on SEMNET.
 claudio barbaranelli posted on Wednesday, March 18, 2020 - 12:14 pm
thank you so much Benght
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message