[Mediation-information] moderated mediation analysis using 'mediation' package

[Teppei Yamamoto]

Liviu,

One thing you could do is to set the moderator to different values along 
its range (running a separate mediate call for each of those values) and 
plot the estimated mediation effects against those moderator values. 
That way, you will see how the indirect effect (or direct effect if 
that's what you are more interested in) varies as a function of the 
moderator, which I think is exactly the goal of moderated mediation 
analysis.

Teppei


(3/22/12 1:17 PM), Liviu Andronic wrote:
> Dear Teppei
>
>
> On Thu, Mar 22, 2012 at 5:34 PM, Teppei Yamamoto<teppei at mit.edu>  wrote:
>> Moderated mediation is supported in the mediation package as of the most
>> recent update (4.0), which we were about to announce to the list
>>
> Very timely update, indeed. :) Thank you so much.
>
>
>> subscribers. Your code looks correct except the last step, where you need to
>> use the "covariates" argument and specify the value of the moderator on
>> which the estimate will be conditioned.
>>
> I tried this and I'm slightly surprised that one needs to use fixed
> values for the moderator. For example, the following ends in an error:
>> med_b<- mediate(mod.m2, mod.y2, dropobs=F, treat='EVPUBLIC', mediator='TRUST', covariates=list(ENGAGESP),  boot=T, sims=50)
> Error: object 'ENGAGESP' not found
>
> while this version succeeds:
>> med_b<- mediate(mod.m2, mod.y2, dropobs=F, treat='EVPUBLIC', mediator='TRUST', covariates=list(ENGAGESP=0.5),  boot=T, sims=50)
>
> I was given to understand that one may use a variable in its entirety
> (as a moderator) instead of using only 'specific values of the
> pre-treatment covariates'. Perhaps we are not talking of the same
> procedure?
>
> Regards
> Liviu

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