Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: Interpreting Poisson regression results with squared term


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: Interpreting Poisson regression results with squared term
Date   Wed, 13 Aug 2008 16:55:23 +0100

In addition, the position of the turning point on the quadratic of Y in
X1 given X2 and X3 can be derived easily. That may or may not have a
substantive interpretation. 

Nick 
n.j.cox@durham.ac.uk 

Maarten buis

--- Antonio Silva <asilva100@live.com> wrote:
> I ran a Poisson regression in which # of groups founded per year was
> the dependent variable. My independent variables of interest are X1
> and X1-squared. 

For these non-linear effects in non-linear models I prefer to show
graphs, in this case of the predicted rate against X1. For these kinds
of graphs -adjust- is your friend, see the example below. The key trick
is to specify in the -adjust- command -by(X1)-, in this case
-by(ment)-, and not mentioning the square term. This way for each value
of X1/ment the mean of the square term is used in the computation,
which is the value of X1/ment squared.

*---------------------- begin example --------------------
// get data 
net from http://www.stata-press.com/data/lf2/
net get rmcdvs

// start the analysis
use couart2, clear
gen ment2 = ment^2
poisson art fem mar kid5 phd ment ment2
adjust fem=0 mar=1 kid5=1 phd , by(ment) exp replace ci
twoway rarea lb ub ment || ///
       line exp ment,      ///
       lpattern(solid) legend(off) ///
       ytitle(rate)
*------------------------ end example ---------------------
(For more on how to use examples I sent to the Statalist, see
http://home.fsw.vu.nl/m.buis/stata/exampleFAQ.html )


*
*   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   |   What's new   |   Site index