Factor Scores PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
 Fritha Milne posted on Wednesday, September 07, 2011 - 3:04 am
I am wanting to calculate factor scores from an EFA to use in further analyses.
I understand that you cannot use the save=fscores command with an EFA. I have tried running my EFA in a CFA format using '@' to set the factor loadings to those calculated using Type=EFA.
It runs fine, except it says that the factor scores cannot be calculated as one of the residual variances for the categorical variables are negative.
How do I get around this problem? Can I set that residual variance to be positive?

 Linda K. Muthen posted on Wednesday, September 07, 2011 - 1:48 pm
This would point to the need for a different factor solution. This would have shown up in the original EFA if the EFA in a CFA model is set up correctly. Perhaps that is the problem.
 Bruno Figueiredo Damásio posted on Thursday, September 27, 2012 - 11:24 pm
Dear Linda and users,

I am a new user of MPLUS and I have two basic questions:

1) What is the estimation method that generate the factor scores when we use the command SAVE = FSCORES;

2) Althought I have read the Appendix 11 of the MPLUS user guide, I am not able to compute factor scores using WLSMV method. Is that possible with a simple sintax? That appendix is completely techinal and I could not implement it.

Best regards,

 Linda K. Muthen posted on Friday, September 28, 2012 - 12:21 am
1. It differs by whether the variables are continuous or categorical and the estimator that is used. Which are you asking about.

2. Factor scores for categorical and WLSMV use an iterative procedure. You cannot do this by hand.
 Elik Ruzimatov posted on Monday, June 24, 2013 - 6:40 pm
Dear Dr. Linda Muthen and users,

Running Principal axis factoring with promax rotation in SPSS I extracted 2 factors out of 25 variables (measures). I saved the factor scores using “regression” method and intending to use them as composite scores in further analysis (Mancova) as DVs, while I have two IVs each with two levels as well as 3 covariates.

What I need to know are: 1) about half of the saved reg scores have negative values, although variables have cleanly and positively loaded on two factors (I did it by defining fixed number of factors to extract). How should I interpret them?
2) will it be the correct method to use the saved scores as DVs in Mancova?

If this topic has been covered, can you please, direct me to it?

Thanks a lot for your help.

BR, Elik
 Linda K. Muthen posted on Monday, June 24, 2013 - 7:14 pm
For questions not related to Mplus, use a general discussion forum like SEMNET.
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