Re: st: interpretation of squared term

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Thu, 26/8/10, David Bai  wrote:
> >   I have a question regarding how to interpret squared
> > term:

--- On Thu, 26/8/10, Maarten buis wrote:
> You can only interpret the time trend graphically. If you
> want to have a non-linear trend where you can interpret
> the coeficients, you can look at a linear spline, see
> -help mkspline-.

Here are some examples on how to implement these recomendations
in Stata. Here is how you can graph the quadratic effect of 
age:
*------------------ begin example -----------------
webuse lbw, clear
gen age2 = age*age
xi: logit low age age2 smoke i.race if age < 45
adjust smoke=0 _Irace_2=0 _Irace_3=0, pr by(age) gen(prhat)
twoway line prhat age if age < 45, sort
*-------------------- end example ------------------

Here is how you can add a linear spline of age to your model
(notice that a linear spline is a type of nonlinear curve)
*------------------- begin example -----------------
mkspline sp1 21 sp2 = age
xi: logit low sp1 sp2 smoke i.race if age < 45, or
*-------------------- end example ------------------
the effect of sp1 is the effect of age when one is younger
than 21, and the effect of sp2 is the effect of age when
one is older than 21. So the odds of getting an underweight
baby increases by 3.7% ((1.037-1)*100%) for every year one 
gets older when one is younger than 21, and it decreases by 
5.8% ((.942-1)*100%) for every year one gets older when 
one is older than 21.

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/