[Mediation-information] Question about mediation package in two level design

[Kosuke Imai]

Hi Sun Yoon,

 Unfortunately, the current version of our package does not incorporate multi-level models (if that's what you are using).  However, if you know some programming, then implementing our general algorithm as described in our Psychological Methods paper should be in principle straightforward: http://imai.princeton.edu/research/BaronKenny.html   I'm ccing Michael Donnelly who used mixed effects models for mediation in case he can share some of his codes. 

Good luck,
Kosuke

Department of Politics
Princeton University
http://imai.princeton.edu

On Oct 12, 2011, at 5:28 PM, Sun Yoon wrote:

> Dear professor Kosuke Imai,
> 
> Hi, my name is Sun, a doctoral student in the Educational Leadership and Policy Analysis department at University of Wisconsin-Madison. I 
> greatly appreciate your effort and contribution on your works. While I'm investigating the mediation analysis on my study, gladly I found "mediation" R package and your papers which I appreciate. But I have some questions about "mediation" package in multi level design. 
> 
> My current study is a randomized cluster experimental design and I am investigating the mediating effect between treatment and outcome. Mediation variable is a count variable in level 2, an independent variable is binary variable inlevel 2, and outcome variable is continuous in level 1. I found that you demonstrated the count mediation variable in the design, but I'm not sure whether I could still use that model in cluster design too. I'd sincerely appreciate it if you could suggest any ideas or reference. I look forward to hearing from you soon. Thank you.
> 
> Best wishes,
> 
> Sun Yoon
> 
> --
> Sun Young Yoon 
> Ph.D. Student 
> Department of Educational Leadership and Policy Analysis 
> University of Wisconsin-Madison