Multiple Imputation and EFA PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
 Emily Rosenzweig posted on Monday, May 25, 2015 - 6:27 pm
Hi --

I need to conduct an exploratory factor analysis on a dataset that requires multiple imputation for MCAR values. Is it possible to include the multiple imputation command in my exploratory factor analysis? And if not, how should I go about this? If I do the multiple imputation on its own, it will generate multiple datasets. What then? Would I simply run the EFA on all of them? Or is there a way to combine the multiple datasets into one and then EFA on that?

 Bengt O. Muthen posted on Tuesday, May 26, 2015 - 2:36 pm
First you should ask if MI is necessary. EFA using ML handles missing under the MCAR and MAR assumption.
 Emily Rosenzweig posted on Wednesday, May 27, 2015 - 12:08 pm
Hi Bengt --

Thanks for your reply. Perhaps I can describe my dataset and you can give me a sense of what you think the best approach might be? I have a dataset with 400 variables, and each participant completed a random selection of 40 of those. Thus my data is truly MCAR. My hope was that multiple imputation prior to EFA would allow me to take into account both variable-specific and participant-specific sources of variance. But if EFA using ML already accounts for both of those, that is fantastic -- I'm happy to cut out that additional imputation step.

Many thanks for your input!
 Bengt O. Muthen posted on Wednesday, May 27, 2015 - 3:31 pm
Yes, you don't need MI.
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