Testing mediation involving two media... PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Tony Reed posted on Tuesday, March 29, 2011 - 7:11 am
Dear Drs. Muthen,

I am using SEM to test a model with 3 IVs (X1-X3), 2 mediators (M1, M2, both at the same level), and 2 DVs(Y1-Y2). I want to see if M1 and M2 have full mediating effects.

For the Mplus code, I just wrote:

X1 BY ****;
***********
***********
Y2 BY ****;

Y1 Y2 ON M1 M2;
M1 M2 ON X1 X2 X3;
Y1 Y2 ON X1 X2 X3;

Model Indirect:
Y1 IND M1 X1;
************;
Y2 IND M2 X3;


But how could I tell if the high correlation, if any, between M1 and M2 would compromise the results I ended up with? Would journal reviewers have problems with this?

Thanks!!!!
 Linda K. Muthen posted on Tuesday, March 29, 2011 - 8:55 am
You may have a problem with multicollinearity in

Y1 Y2 ON M1 M2;
 Tony Reed posted on Wednesday, March 30, 2011 - 6:03 pm
Hi Dr. Muthen,

Thanks for the response. So, should I write instead as follows:

Y1 Y2 ON M1;
Y1 Y2 ON M2;

thanks!!

Tony
 Linda K. Muthen posted on Wednesday, March 30, 2011 - 6:28 pm
There is no difference between the two specifications.
 Tony Reed posted on Thursday, March 31, 2011 - 11:16 am
So, would it be meaningful if I controlled for the correlation of M1 and M2?
 Linda K. Muthen posted on Thursday, March 31, 2011 - 1:11 pm
I know of no way to deal with multicollinearity other than not to use both variables at the same time.
 Tony Reed posted on Friday, April 01, 2011 - 12:53 am
Dr. Muthen,

Thanks for your response!! I will give a second thought on my model....

Thanks!!

Tony
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: