Outliers and EFA PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
Message/Author
 Jiran Hou posted on Saturday, April 05, 2014 - 5:20 pm
Hi, I am using mplus to conduct EFA to examine 36 motive items so as to get a motive structure. My sample size is 306, and I have identified 39 multivarite outliers by using the Mahalanobis distance. Should I directly delete these 39 cases? Or keep them but transform them? When I analyze the 306 cases, I have identified 6 factors. When I analyze the 267 cases after deleting these 39 outliers, I have also identified 6 factors, 5 of them are the same factors identified by using 306 cases, and 1 new factor that is different from that in the exam of 306 cases. Do you have any suggestion regarding my situation?

Thank you!

Gina
 Linda K. Muthen posted on Sunday, April 06, 2014 - 5:45 pm
If they are truly outliers, they should be deleted. You may want to check some of the other outliers like loglikelihood to see if you get agreement.
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: