Path Analysis and Covariates PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Stephen Salbod posted on Monday, August 12, 2013 - 1:55 pm
Let me start off by saying I’m a newbie to the program and this site.

I have a problem with some nuisance demographics that I want to control: age (continuous) and gender (two groups, dummy coded).

To deal with the problem I regressed my five variables: x1, x2, m1, m2, and Y on age and gender and saved the residuals which I used in this model:

m1 ON x1, x2;
m2 ON x1, x2;
y ON m1, m2, x1, x2;

I thought this was a neat way to approach the problem (at the time), but now I’m having second thoughts. Both age and gender were related to y, but only age was related to x1 and x2, while gender was related to m1 and m2.

Would I just modify the last line: y ON age gender m1 m2 x1 x2?

Any suggesting will be greatly appreciated. Thank you.
 Linda K. Muthen posted on Tuesday, August 13, 2013 - 9:04 am
I would include age and gender in all of the equations:

m1 ON x1, x2 age gender;
m2 ON x1, x2 age gender;
y ON m1, m2, x1, x2 age gender;
 Stephen Salbod posted on Monday, September 02, 2013 - 11:44 am
Dear Linda, Thank you for your suggestion. When I want to control for a variable I just need to include in the equations: just like I would do in regression analysis. (My sample is only 150. I plan to use CINTERVAL(bcbootstrap) to develop confidence intervals for my indirect effects.)


What I want to do next is to examine the model within males and females. I'm planning on reviewing the discussion on moderation in this Mplus Discussion topic. Is there any article I should read?
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: