Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

# Re: Re: st: binary mediation command

 From Pina Valle To statalist@hsphsun2.harvard.edu Subject Re: Re: st: binary mediation command Date Tue, 05 Jul 2011 12:37:21 -0400

Philip,
Yes that was helpful, as that is what my co-author and I believed as well from a website on binary mediation. One last question..when using the bootstrap CI, and getting a CI for the indirect effect that does not include zero, we then can report proportion of total effect mediated. However, proportion of total effect mediation is -.084. Would we still report 8.5% or does the negative mean something else?

Thank you.
Pina

----- Original Message -----
From: Philip Ender <ender97@gmail.com>
Date: Tuesday, July 5, 2011 11:12 am
Subject: Re: Re: st: binary mediation command
To: statalist@hsphsun2.harvard.edu

> Dear Pina,
>
> You didn't include the bootstrap results for the direct and indirect
> effects so I can only give general advice.  Check the bootstrap
> confidence intervals for your direct and indirect effects.  In
> particular, look at your indirect effects because it seems rather
> small.  If the confidence interval does not include zero then you can
> report the proportion of total effect mediated.  If, on the other
> hand, the confidence interval includes zero then the proportion
> mediated is essentially meaningless since the indirect effect is not
> reliably different from zero.
>
> I hope this response helps.
>
>
> Pina Valle wrote:
> >Even when using percentile confidence intervals, is this output
> relevant since I am trying to determine the >proportion mediated
> by the indirect effect?
>
> >Indirect effects with binary response variable evercoh
> >        indir_1 = .00601173        (adrel, binary)
> >total indirect  = .00601173
> >direct effect  = .10223182
> >total effect  = .10824356
> >      c_path   = .10680445
> >proportion of total effect mediated = .05553895
> >ratio of indirect to direct effect  = .05880491
> >Binary models use logit regression
>
> >Thanks
> >Pina
>
> --
> Phil Ender
> UCLA Statistical Consulting Group
> *
> *   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/
>

Giuseppina Valle, M.S.
Department of Sociology
Doctoral Candidate
Florida State University
526 Bellamy Building
Tallahassee, FL 32306-2240

Email:   gv08d@fsu.edu
*
*   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/