Factor loadings differ (SEM vs CFA)... PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Hannah Wittmann  posted on Thursday, July 19, 2012 - 8:07 am
Hi,

Ive run three confirmatory factor analysis (one for each of my three questionnaires; for example "A1 by a b c" and so on). After that, i ran a SEM, containing all three questionnaires:

questionnaire 1:
A1 by a b c
A2 by d e f

questionnaire 2:
B1 by g h
B2 by i j k l

questionnaire 3:
C1 by m n o
C2 by p q r s

A1 on B1 B2 C1 C2
A2 on B1 B2 C1 C2

When I look at the CFA-outputs the factor loadings differ from the factor loadings presented in the SEM output.

Is that ok, might that be due to different number of degrees of fredom?
Or does this indicate a problem?

Thank you very much for your help!
 Linda K. Muthen posted on Thursday, July 19, 2012 - 10:55 am
This comes about because the observed variables do not correlate only through the factors. You can use modification indices to examine this. Use

MODINDICES (ALL);

in the OUTPUT command.
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: