Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Maarten buis <maartenbuis@yahoo.co.uk> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: Would _sgmediation_ work for dichotomous dependent variable? |
Date | Sun, 1 Aug 2010 22:54:50 +0000 (GMT) |
--- On Sun, 1/8/10, Bo MacInnis wrote: > The help file of binary_mediation indicates that mediator > variables shall be standardized. Is it a recommendation or a > requirement? In my current study, unstandardized mediator > seem to make more sense than standardized mediator. > > Another question on effect size: I am an economist and > trained in being interested in marginal effects rather than > coefficients in the nonlinear model like logit or probit but > I am working on project for a psychologist. Would the > mediational analysis be extended to marginal effect rather > than coefficients in the probit model? If I understand Phill's model correctly, he solves the problems involved with non-linear models by declaring that we care about the effects on the latent variable rather than the effects on the probability or the odds. In that case the model becomes linear again. This answers both your question: Yes, the standardization is a requirement, because the scale of the latent variable is unidentified, so you need to fix it by standardizing. No, you can't use this with marginal effects, because then the model becomes non-linear again. Hope this helps, Maarten -------------------------- Maarten L. Buis Institut fuer Soziologie Universitaet Tuebingen Wilhelmstrasse 36 72074 Tuebingen Germany http://www.maartenbuis.nl -------------------------- * * 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/