Competing mediators PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Yinxuan Huang posted on Sunday, September 27, 2015 - 6:11 pm
Hi all, I have one question regarding competing mediators in path analysis. In my model, I want to test the differences between indirect paths via different competing mediators. So this is for example normally the way how I do it.
MODEL:
Y on M1 (b1)
M2 (b2)
X ;
M1 on X (a1);
M2 on X (a2);
MODEL TEST:
0 = a1*b1 a2*b2
Now I am trying to construct a path model using one exogenous variable which is nominal (5-cats). Therefore I created four dummy variables to represent this variable. However, a potential problem that 'a standardized version of this type of relative arbitrary scale is still arbitrary; standardized effect size measures are not necessarily any more meaningful theoretically or practically than unstandardized measures' (Hayes and Preacher, 2014:461).
My question is: how does Mplus test for competing mediators in this particular case when k-1 dummies are used to represent nominal exogenous variable with k categories? I assume the outcomes are calculated based on unstandardized path coefficients. In this case, could the results in the model test for competing mediators fall into the 'arbitrary trap' as well?
Thank you very much.
 Bengt O. Muthen posted on Monday, September 28, 2015 - 12:26 am
To test such differences in indirect effects you want M1 and M2 to be on the same scale. It doesn't matter what the X variable scale is since you use the same X for each indirect effect. That's also true for dummy X's.

You may want to ask these general analysis questions on SEMNET.
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: