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: compare effect size between dummys and metrics variables in logistic


From   "Clyde Schechter" <clyde.schechter@einstein.yu.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: compare effect size between dummys and metrics variables in logistic
Date   Tue, 28 Sep 2010 08:22:32 -0700

Although I think these situations are very rare, I have seen at least one
instance where a sensible comparison between the coefficient of a dummy
and the coefficient of a metric variable was done.  The context was a
linear regression rather than a logistic one, but the same logic could
apply to a dichotomous outcome.

I saw this a long time ago (it was not my work), my memory of it is hazy,
and I cannot find the reference right now.  But the gist of it was that
the unit of analysis was the county or some other health-jurisdiction, the
dependent variable was the infant mortality rate (or, probably, the log
thereof), and the two predictors of interest were a dummy variable
indicating whether the county had implemented some particular policy, and
calendar year.  The inclusion of calendar year was to capture the effect
of a long-standing secular decline in infant mortality rates.

In their discussion of results, the authors remarked that the coefficient
of the dummy variable was approximately twice the coefficient of time--and
they even did a test of that hypothesis.  From that they concluded that
implementation of the policy in question had an impact roughly equivalent
to leaping forward two years in time with respect to progress in reducing
infant mortality.

I wish I could remember more details and find the reference.  Obviously
that conclusion is also contingent on the model's being properly
specified, and substantive assumptions supporting causality.  But I think
it illustrates a (possibly) sensible comparison between the coefficient of
a dummy and the coefficient of an interval-level variable.

Clyde Schechter, MA MD
Associate Professor of Family & Social Medicine
Albert Einstein College of Medicine
Bronx, NY, USA

Please note new e-mail address: clyde.schechter@einstein.yu.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/


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