Re: st: Re: Poisson and marginal effects

[Maarten Buis <maartenlbuis@gmail.com>]

On Mon, May 14, 2012 at 12:57 AM, Jessie C wrote:
> Thank you for your help.  I am interested in comparing the Poisson
> estimates to OLS estimates.  Would the comparison be the OLS
> regression of log y versus Poisson of y and I multiply the coefficient
> coming from f the Poisson regression by the mean of y?
>
> The variable of interest is discrete.  If the coefficient is beta_1,
> how I interpret that?  Is it a going from 0 to 1 increases y by beta_1
> percent?  or beta_1 * mean y percent?

You can use the -eform()- in regress, but it will give the ratio
effects in terms of geometric means rather than arithmatic means, see:
Roger Newson (2003) Stata tip 1: The eform() option of regress. The
Stata Journal, 3(4):445.
<http://www.stata-journal.com/article.html?article=st0054>

This will get you a bit closer to comparabiltiy of the results with
-poisson- with the -irr- option, but not quite. Bottom line: just do
not use -regress- on a log transformed dependent variable.

Hope this helps,
Maarten

> On Sun, May 13, 2012 at 6:41 PM, Jeffrey Wooldridge
> <jmwooldridge60@gmail.com> wrote:
>> In Poisson regression the average partial effect of a continuous
>> variable is just the sample average of y times the coefficient.
>> Naturally, it is harder for a discrete change. There is makes sense to
>> obtain the predicted value at the two different settings of the
>> explanatory variable, with other variables evaluated at their observed
>> values, and average the difference.
>
>> The reason this is not regularly done in Poisson regression is that
>> the coefficients have interpretations as percentage changes.
> *
> *   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/



-- 
--------------------------
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/