# Re: st: How to add restriction condition in Stata

 From Raymond <[email protected]> To [email protected] Subject Re: st: How to add restriction condition in Stata Date Fri, 29 Jul 2005 14:33:57 -0400

```Well, if you only put age in the equation, there is no difference.
Once you put age, period and cohort into the regression, the
difference is substantial.
I'm not sure that I can explain this well, but there is an article you
may want to take
a look at:
Fienberg, S.E. and W.M. Mason. 1978. "indentification and estimation
of age-period-cohort models in the analysis of discrete archival
data." in Sociological Methodology 1979 edited by K.F. Schuessler.

Thanks!

On 7/29/05, FEIVESON, ALAN H. (AL) (JSC-SK) (NASA)
<[email protected]> wrote:
> I still don't see why you can't sum the ages. What's the difference between:
>
> y ~ Poisson(la); log la = b0 + b1*age1 + b2*age2  with b1=b2
>
> and
>
> y ~ Poisson(la); log la = b0 + b1*(age1+age2) ?
>
>
> Al Feiveson
>
>
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Raymond
> Sent: Friday, July 29, 2005 11:05 AM
> To: [email protected]
> Subject: Re: st: How to add restriction condition in Stata
>
> Thanks!
> Actually, I'm using an age-period-cohort model, and I need to put all the
> three variables in the model. Without restrictions, there will be an
> identification problem since cohort=period-age.
> Thanks for your suggestions. I'll try it.
>
> On 7/29/05, Richard Williams <[email protected]> wrote:
> > At 11:51 AM 7/29/2005 -0400, Raymond wrote:
> > >Thanks!
> > >But the reason I make coeffients of age1 and age2 be equal is to
> > >solve an identification problem. So, I cann't just sum them up.
> >
> > I'm not sure how you are getting an identification problem in a
> > poisson regression or how setting the effects for 2 dummies equal
> > solves it.  Maybe you should write out your model.  But in any event,
> > I think the test command I gave before will still work; and the
> > poisson command does support constraints, so you could do
> >
> > constraint 1 age1=age2
> > poisson y age1 age2, constraints(1)
> >
> > *
> > *   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/
> *
> *   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/
```