Monte carlo for EFA PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
Message/Author
 Danielle Clark posted on Saturday, August 10, 2013 - 7:52 am
I'm running a simulation for an EFA with 50 continuous indicators and 5 factors, a residual variance of .4 for all indicators, and a correlation of .4 for factor 1 with factor 3. I looked at sample sizes of 100, 300, 500, and 1000 and noticed that for factor 3, the parater estimates became more biased as the sample size increased. Any suggestions?
model population:
f1 by y1-y10*.65 y11-y20*.05 y21-y30*.25 y31-y40*.3 y41-y50*.1;
f2 by y1-y10*.05 y11-y20*.85 y21-y30*.1 y31-y40*.05 y41-y50*.05;
f3 by y1-y10*.1 y11-y20*.05 y21-y30*.6 y31-y40*.1 y41-y50*.12;
f4 by y1-y10*.05 y11-y20*.12 y21-y30*.1 y31-y40*.65 y41-y50*.1;
f5 by y1-y10*.05 y11-y20*.12 y21-y30*.07 y31-y40*.15 y41-y50*.7;
f1@1 f2@1 f3@1 f4@1 f5@1;
f1 with f3@.4;
y1-y50*.40;
model:
f1 by y1-y10*.65 y11-y20*.05 y21-y30*.25 y31-y40*.3 y41-y50*.1 (*1);
[repeat as above for f2-f5]
y1-y50*.40;
 Bengt O. Muthen posted on Saturday, August 10, 2013 - 9:20 am
Please send the output for n=1000 to support.
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: