Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Re: Poisson model: how to impose restrictions on coefficients


From   "Patrick Fitzgerald" <pfitzgerald@m-tag.net>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Re: Poisson model: how to impose restrictions on coefficients
Date   Thu, 13 Jun 2002 10:05:30 +1000

Since a2 + a3 = 1, a3 = 1 - a2, and you can re-write your model as:

Y =  a1 X1  +  a2 X2   + (1 - a2)  X3  +  a4 X4   +  error term,

which can be rearranged to

Y = a1 X1 + a2 (X2 - X3) + X3 + a4 X4 + error term,

so that you can fit the model you want using the STATA "poisson" command by

(1) creating a new variable, Z = X2 - X3 and fitting the model

Y = a1 X1 + a2 Z + a4 X4

(2) setting the option offset(X3).

The Poisson regression command for this would then be:

poisson outcome x1 z x4, offset(x3)

Notes

(a) I assume you wish to fit Y = log(mean of outcome).

(b) You don't estimate 4 coeffficients, but 3, as a3 is now completely
specified in terms of a2.

(c) Only FYI. If (a) holds, then the error term is added to the outcome
itself, not Y.

Best wishes,

Patrick F.


----- Original Message -----
From: <Sailangita@aol.com>
To: <statalist@hsphsun2.harvard.edu>
Sent: Saturday, June 08, 2002 11:15 PM
Subject: st: Poisson model: how to impose restrictions on coefficients


> Dear All:
>
> I need help on how to impose restrictions on coefficients.
> I am running a poisson model like this:
>
> Y =  a1 X1  +  a2 X2   + a3  X3  +  a4 X4   +  error term
>
> I want to impose   a2 +  a3  = 1
>
>
> Thanks in advance.
>
>
> Diafuka
> *
> *   For searches and help try:
> *   http://www.stata.com/support/faqs/res/findit.html
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   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