RE: st: GLLAMM SYNTAX MEDIATION

["Smith, Stephanie Moller" <Stephanie.Moller@uncc.edu>]

The equation is:

y=b0 + b1x1 
b0=b2w1+u
logit link
where w1 is predicted to impact y via x1


I found the attached logistic mediation spreadsheet online.
Could I group mean center x1, and then put the slope, sd, etc of x1bar... into the attached spreadsheet
here y=b0+b1(xij-xbarj)
b0=b2xbarj + b3w1j + U
(logit link)

so that w1 is the causal variable and xbarj is the mediator variable?   If so, would I need to get descriptives for yij or for ybarj?

Thank you.

Stephanie Moller, Ph.D.
Associate Professor and Director of Graduate Studies, Department of Sociology
Interim Director, Masters of Arts in Liberal Studies Program
University of North Carolina at Charlotte

http://sociology.uncc.edu/people/smoller/
Specializations: Sociology of Education, Social Stratification and Political Sociology

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Stas Kolenikov
Sent: Thursday, February 17, 2011 11:53 AM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: GLLAMM SYNTAX MEDIATION

That's more complicated, then. While -reg3- does not support
-vce(cluster)-, you can work around this with the combination of
-bsweights-/-bs4rw-... unless you really need the variance
decompositions between levels (rather than just the estimates with
reliable standard errors.)

Can you write down the model explicitly, equation by equation, with
multilevel random effects?

On Thu, Feb 17, 2011 at 9:10 AM, Smith, Stephanie Moller
<Stephanie.Moller@uncc.edu> wrote:
> Thank you.  it is all observed, but it is multilevel.
>
> Stephanie Moller, Ph.D.
> Associate Professor and Director of Graduate Studies, Department of Sociology
> Interim Director, Masters of Arts in Liberal Studies Program
> University of North Carolina at Charlotte
>
> http://sociology.uncc.edu/people/smoller/
> Specializations: Sociology of Education, Social Stratification and Political Sociology
>
> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Stas Kolenikov
> Sent: Thursday, February 17, 2011 10:00 AM
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: GLLAMM SYNTAX MEDIATION
>
> Do you have observed or latent variables? With observed variables, you
> probably would only need -reg3-.
>
> On Thu, Feb 17, 2011 at 8:37 AM, Smith, Stephanie Moller
> <Stephanie.Moller@uncc.edu> wrote:
>> Is there a site where I can find a fully worked up example of mediation analysis using gllamm with logit link?
>>
>> I have looked at the GLLAMM manual, but I find it confusing.
>>
>> this is a simple 2-1-1 path model.  I want to make sure that I properly design and interpret the direct and indirect effects.
>>
>> Thank you in advance.
>> Stephanie Moller
>>
>> *
>> *   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/
>>
>
>
>
> --
> Stas Kolenikov, also found at http://stas.kolenikov.name
> Small print: I use this email account for mailing lists only.
>
> *
> *   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/
>
> *
> *   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/
>



-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.

*
*   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/

Attachment: Copy of Nate's Logistic Mediation Spreadsheet.xls
Description: Copy of Nate's Logistic Mediation Spreadsheet.xls