Differences between programs PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
Message/Author
 Quint Oga-Baldwin posted on Friday, September 25, 2015 - 7:24 pm
I have run an EFA in both MPlus and Stata and found considerable differences in the representation of the factor coefficients. Both programs indicated the same number (7) factor solution as the best fit for the data - MPlus through fit indices, Stata through Parallel Analysis.

Procedures for both:

ML extraction
Oblimin rotation

While I obtained a theoretically interpretable solution in MPlus (replicated in independent CFA), Stata provides very different coefficients for the factors. Is there any reason for this and how should I approach the interpretation?
 Bengt O. Muthen posted on Saturday, September 26, 2015 - 6:14 pm
First make sure that the same sample size is used and that you are comparing the factor patterns, not the factor structure (so looking at loadings, not correlations between factors and items). If all that squares, send the outputs from the 2 programs to support.

Note also that Mplus provides parallel analysis.
 Quint Oga-Baldwin posted on Wednesday, September 30, 2015 - 6:11 pm
I have solved the problem. For reference, Stata requires specification of Oblique rotation when using Oblimin.
 Jo Madison posted on Sunday, November 29, 2015 - 10:05 pm
Hi there,

I did EFA using Mplus 7.4 and SPSS and got quite different factor loadings

For Mplus I used the following script:

title: uses file data1.csv wich is based on 21.sav which is correct data.
data: file is ...
variable:
names are ....

usevariables are ...

ANALYSIS:
Type = EFA 5 6;
ROTATION= VARIMAX;


and for SPSS I did EFA using the following syntax:

DATASET ACTIVATE DataSet1.
FACTOR
/VARIABLES aca1 aca2 aca3 aca4 aca5 itc1 itc2 itc3 itc4 itc5 opp1 opp2 opp3 opp4 opp5 glp1 glp2
glp3 glp4 glp5 grp1 grp2 grp3 grp4 grp5 prs1 prs2 prs3 prs4
/MISSING LISTWISE
/ANALYSIS aca1 aca2 aca3 aca4 aca5 itc1 itc2 itc3 itc4 itc5 opp1 opp2 opp3 opp4 opp5 glp1 glp2
glp3 glp4 glp5 grp1 grp2 grp3 grp4 grp5 prs1 prs2 prs3 prs4
/PRINT INITIAL EXTRACTION ROTATION
/FORMAT SORT BLANK(.33)
/CRITERIA MINEIGEN(1) ITERATE(25)
/EXTRACTION ML
/CRITERIA ITERATE(25)
/ROTATION VARIMAX.


Please help
 Bengt O. Muthen posted on Monday, November 30, 2015 - 5:32 pm
The programs should agree if you have the same:

sample size
estimator
rotation

and you are looking at the same loading pattern part of the output.

If not, send the 2 outputs to Support along with your license number.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: