Categorical DV in mediation analysis
Message/Author
 YuChang posted on Monday, September 12, 2011 - 3:00 am
Hi,would you like help me for 2 question?
I have a mediation model: one continuous IV, one continuous nediator variable and one three-category DV. I used model indirect and STDYX command in Mplus 6.After run, I found two things I couldnt understand:
1. In StdYX Estimate output, there is only Estimate,NO S.E..Why? Is the standardized formula same as binary DV of mediation analysis£¿as MacKinnon's paper£¨2007£©mentioned?
2. In the indirect output, how can we interpret unstandardized indirent effect?
Yu Chang
 Linda K. Muthen posted on Tuesday, September 13, 2011 - 9:48 am
1. With WLSMV and covariates, we do not give standard errors for standardized estimates. Yes, the forumla is the same.
2. The interpretation is the same as for a regression coefficient, the change in y for a one unit change in x.
 Wen-Hsu Lin posted on Tuesday, November 24, 2015 - 7:19 pm
Hi, Mplus team, would you mind to help me on mediation. My model: 3 Xs, 1 mediating variable, which as three categories (a), and 1 distal outcome variable (y) (count variable).
My model:

X1 X2 X3---> a---->y
x1 x2 x3--->y

How do I get: (1) the effects of each X on a (2) how do I get each the effects of each category of a on y, and finally (3) the mediating from x through a to y?

Do I use known class command? Thank you.
 Bengt O. Muthen posted on Wednesday, November 25, 2015 - 10:16 am
For (1) and (3), see the paper on our website:

Muthén, B. & Asparouhov, T. (2015). Causal effects in mediation modeling: An introduction with applications to latent variables. Structural Equation Modeling: A Multidisciplinary Journal, 22(1), 12-23. DOI:10.1080/10705511.2014.935843

For (3), you could do a separate analysis of y regressed on 2 dummy variables representing a.
 Wen-Hsu Lin posted on Wednesday, November 25, 2015 - 4:47 pm
Thanks a lot.