<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7638.1">
<TITLE>RE: Causally dependent mediators</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<P><FONT SIZE=2>Thank you Teppei and Mashhood for your very helpful comments.<BR>
<BR>
I would be using then:<BR>
<BR>
1) deprivation --> education level --> aging well (covariates: age, gender, region)<BR>
(with the R mediation package, to study the direct and indirect effects of deprivation on aging well)<BR>
<BR>
2) deprivation --> education level --> economic position --> aging well (covariates: age, gender, region)<BR>
(with the R MULTIMED package, to study the direct and indirect effecs mediated through economic position of deprivation on aging well)<BR>
<BR>
Does it seem OK?<BR>
<BR>
Best regards,<BR>
<BR>
Angel<BR>
<BR>
<BR>
-----Mensaje original-----<BR>
De: mediation-information-bounces@lists.r-forge.r-project.org en nombre de Teppei Yamamoto<BR>
Enviado el: mié 27/08/2014 18:02<BR>
Para: mediation-information@r-forge.wu-wien.ac.at<BR>
Asunto: Re: [Mediation-information] Causally dependent mediators<BR>
<BR>
Hi Angel,<BR>
<BR>
The second model does not give you estimates of those effects "net of<BR>
deprivation" -- you will get the direct and indirect effects of<BR>
manipulating education levels from one level to another on aging.<BR>
<BR>
Including deprivation as a covariate in the models allows you to correct<BR>
the confounding between education and economic position and/or education<BR>
and aging due to deprivation, but that does not allow you to interpret<BR>
the estimated mediation/direct effects as those effects "net of"<BR>
deprivation.<BR>
<BR>
If you are interested in how the effect of deprivation on aging is<BR>
transmitted by those two interrelated mediators, you want to use<BR>
multimed (as in your original message) instead of model 2.<BR>
<BR>
Best,<BR>
Teppei<BR>
<BR>
(8/27/14, 10:35 AM), Angel Rodriguez wrote:<BR>
><BR>
> Thank you for your answers, Teppei.<BR>
><BR>
> So, would these two models make sense?<BR>
><BR>
> 1) deprivation --> education level --> aging well (covariates: age,<BR>
> gender, region)<BR>
> (with R mediation package, to study the direct and indirect effects of<BR>
> deprivation on aging well)<BR>
><BR>
> 2) education level --> economic position --> aging well (covariates:<BR>
> DEPRIVATION, age, gender, region)<BR>
> (with R mediation package, to study the direct and indirect effecs on<BR>
> education on aging well, net of deprivation)<BR>
><BR>
> Best,<BR>
><BR>
> Angel<BR>
><BR>
> -----Mensaje original-----<BR>
> De: mediation-information-bounces@lists.r-forge.r-project.org en nombre<BR>
> de Teppei Yamamoto<BR>
> Enviado el: mié 27/08/2014 15:27<BR>
> Para: Mediation-information@r-forge.wu-wien.ac.at<BR>
> Asunto: Re: [Mediation-information] Mediation-information Digest, Vol<BR>
> 25, Issue 3<BR>
><BR>
> Hi Angel,<BR>
><BR>
> The difference between the two estimates (multimed vs. mediation<BR>
> analysis on education level as the "treatment") is that the former is<BR>
> about the effect of manipulating infant deprivation on aging whereas the<BR>
> latter is about the effect of manipulating education directly. That is,<BR>
> they are estimating effects of different kinds of hypothetical<BR>
> interventions and how they are transmitted through the variables that<BR>
> are assumed to be causally descendant.<BR>
><BR>
> So I think your choice boils down to what causal quantity you are trying<BR>
> to estimate.<BR>
><BR>
> Best,<BR>
> Teppei<BR>
><BR>
><BR>
> (8/27/14, 5:38 AM), Angel Rodriguez wrote:<BR>
> > Thank you again, Mashhood.<BR>
> ><BR>
> > What I do not understand yet is why we need a package like multimed if<BR>
> > we can decompose a model with two causally dependent mediators (infant<BR>
> > deprivation --> education level --> economic position --> aging well)<BR>
> > into two models (infant deprivation --> education level --> aging well &<BR>
> > education level --> economic position --> aging well).<BR>
> ><BR>
> > Following your advise, I should include infant deprivation as a<BR>
> > confounder in the model education level --> economic position --> aging<BR>
> > well. But if deprivation conditions education strongly, I could end not<BR>
> > finding any direct nor indirect effect of education on aging well.<BR>
> ><BR>
> > Best regards,<BR>
> ><BR>
> > Angel Rodriguez-Laso<BR>
> ><BR>
> ><BR>
> > -----Mensaje original-----<BR>
> > De: mediation-information-bounces@lists.r-forge.r-project.org en nombre<BR>
> > de Mashhood Sheikh<BR>
> > Enviado el: mar 26/08/2014 20:56<BR>
> > Para: mediation-information@lists.r-forge.r-project.org<BR>
> > Asunto: Re: [Mediation-information] Mediation-information Digest, Vol<BR>
> > 25,Issue 3<BR>
> ><BR>
> ><BR>
> > Hi,<BR>
> ><BR>
> > In the model (education level --> economic position --> aging well) you<BR>
> > should include all the variables that potentially confound the<BR>
> > association between education level and economic position, education<BR>
> > level and ageing well, and economic position and ageing well.<BR>
> ><BR>
> > Best wishes,<BR>
> > Mashhood<BR>
> > ---<BR>
> > Mashhood Ahmed Sheikh<BR>
> > Research Fellow,<BR>
> > Department of Community Medicine<BR>
> > Faculty of Health,<BR>
> > University of Tromsø,<BR>
> > Norway<BR>
> ><BR>
> ><BR>
> ><BR>
> ><BR>
> > _______________________________________________<BR>
> > Mediation-information mailing list<BR>
> > Mediation-information@lists.r-forge.r-project.org<BR>
> ><BR>
> <A HREF="https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/mediation-information">https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/mediation-information</A><BR>
><BR>
> --<BR>
> ====================================<BR>
> Teppei Yamamoto<BR>
> Assistant Professor<BR>
> Department of Political Science<BR>
> Massachusetts Institute of Technology<BR>
> <A HREF="http://web.mit.edu/teppei/www/">http://web.mit.edu/teppei/www/</A><BR>
> ====================================<BR>
> _______________________________________________<BR>
> Mediation-information mailing list<BR>
> Mediation-information@lists.r-forge.r-project.org<BR>
> <A HREF="https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/mediation-information">https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/mediation-information</A><BR>
><BR>
><BR>
><BR>
><BR>
><BR>
> _______________________________________________<BR>
> Mediation-information mailing list<BR>
> Mediation-information@lists.r-forge.r-project.org<BR>
> <A HREF="https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/mediation-information">https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/mediation-information</A><BR>
<BR>
--<BR>
====================================<BR>
Teppei Yamamoto<BR>
Assistant Professor<BR>
Department of Political Science<BR>
Massachusetts Institute of Technology<BR>
<A HREF="http://web.mit.edu/teppei/www/">http://web.mit.edu/teppei/www/</A><BR>
====================================<BR>
_______________________________________________<BR>
Mediation-information mailing list<BR>
Mediation-information@lists.r-forge.r-project.org<BR>
<A HREF="https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/mediation-information">https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/mediation-information</A><BR>
<BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>