Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

re: st: mediation analysis with multi-level logit models


From   "Ariel Linden, DrPH" <ariel.linden@gmail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   re: st: mediation analysis with multi-level logit models
Date   Thu, 26 Apr 2012 10:49:26 -0400

Hi Edward,

Take a look at the following UCLA website:

http://www.ats.ucla.edu/stat/stata/faq/ml_mediation.htm

I also strongly suggest you read the recent paper by Zhang et al., which
highlights some of the problems when performing multilevel mediation
analysis. There are additional concerns when the either the mediator or
outcome variables are binary, and some commentators suggest rescaling (see
for example, Krull and MacKinnon). Others have argued that this rescaling
does not solve the problem and that the potential outcomes approach is more
appropriate (see Imai et al.). I have not seen any more recent literature on
the multilevel models with binary mediators/outcomes, or potential outcomes
approaches with multilevel data.

I hope this helps

Ariel

References:

Imai K, Keele L, Yamamoto T. Identification, inference, and sensitivity
analysis for causal mediation effects. Statistical Science 2010;25(1):51-71.

Krull JL, MacKinnon DP. Multilevel modeling of individual and group level
mediated effects. Multivariate Behavioral Research 2001;36(2):249-277. 

Zhang Z, Zyphur MJ, Preacher KJ. Testing multilevel mediation using
hierarchical linear models: Problems and solutions. Organizational Research
Methods 2009;12:695-719.

Date: Wed, 25 Apr 2012 10:25:04 +0200
From: Edward Lorenz <Edward.Lorenz@gredeg.cnrs.fr>
Subject: st: mediation analysis with multi-level logit models

Hello,

I am interested in carrying out a mediation analysis with a multi-level
logit 
model. I am using 'xtmelogit' on Stata 11.0 for Mac. My level-2 IV is
continuous 
and the DV and level-1 mediators are binary. I am aware of the Stata program

'binary-mediation' for single level logit models. The bootstrap command is
used 
to compute the standard errors for the direct and indirect effects. Is there
a 
program for multi-level logit models?

Thanks for any help on this.


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index