Re: Measurement Invariance: on initial CFA or revised CFA?

Posted by Art Kendall on Apr 10, 2012; 11:47am
URL: http://spssx-discussion.165.s1.nabble.com/Measurement-Invariance-on-initial-CFA-or-revised-CFA-tp5628686p5629699.html

I concur with David that you would be better off going to an SEM group. However, you should give them a more detailed question to react to.

However, how did you come up with the number of factors to retain?  Was this a pre-existing instrument?
Was the first factor analysis in fact an exploratory analysis?

How was the set of items selected?  Was there already a scoring key?
What is the response scale for the items?

Would I be correct in guessing that  the Kaiser criterion was used for the number of factors to retain? 

On 4/9/2012 8:15 PM, nnhartma wrote:

Hello:

I am testing for measurement invariance across two groups. Do I need to run
measurement invariance on the initial CFA or revised CFA model? I understand
how to run the measurement invariance test as well as how to interpret it.

Looking at the literature across various fields as well as several SEM texts
leads me to believe that I should do a CFA for each group separately and
then test the final CFA models for each group for measurement invariance.
However, I am working with a 9 latent factor model which has approximately
40 items. Intuition tells me that in larger models we will always fail to
find invariance (there are always items that need to be deleted IMO).

Thank you for your help.
Best!


--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Measurement-Invariance-on-initial-CFA-or-revised-CFA-tp5628686p5628686.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD