[Mediation-information] Fwd: Help

[Teppei Yamamoto]

Hi Peter,

The three-way interaction (T-M-X) allows the T-M interaction effect to 
be moderated by X. That is, the degree to which the mediation effects 
depend on the baseline value of the treatment may itself vary as a 
function of the covariate X.

(I'm forwarding this to the mediation list because it may be useful for 
others --- I hope you are fine with it.)

Teppei


Peter Belmi wrote:
>
>
> Hi Dustin and Teppei,
>
> In your response to one of the questions posed by another subscriber,
> you indicated that it might be helpful to include a three-way
> interaction in the outcome model for testing moderated mediation. I'm
> curious -- is there any particular advantage or reason to include a
> three-way interaction term? Most prescriptions I've seen (Edwards &
> lambert; Preacher & Hayes) do not seem to have it.
>
> Peter
>

-- 
====================================
Teppei Yamamoto
Assistant Professor
Department of Political Science
Massachusetts Institute of Technology
http://web.mit.edu/teppei/www/
====================================