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

Re: st: Wald test limit?

From   Keith Dear <>
Subject   Re: st: Wald test limit?
Date   Wed, 06 Feb 2008 14:54:08 +1100

I fitted the same three models using poisson instead of nbreg.  The results:
  Actual  nbreg  poisson
   261     255     253
   262     257     253
   263     257     256
So the answer would appear to be no.
("actual" is parameters fitted excluding _cons and, for nbreg, alpha).

At 01:44 PM 6/02/2008, someone wrote offlist:
Hi Keith,

Is it perculiar to nbreg?

Are the df correct if you replace nbreg with poisson?

At 01:14 PM 6/02/2008, you wrote:
Good idea. This is what happened:

xi: nbreg rrv_count (i.fin_year) $BASE, exposure(person_days) cl(pcode)
gen x1=uniform()
xi: nbreg rrv_count (i.fin_year) $BASE x1, exposure(person_days) cl(pcode)
gen x2=uniform()
xi: nbreg rrv_count (i.fin_year) $BASE x1 x2, exposure(person_days) cl(pcode)

The first model reports Wald chi2 on 255df, the second 257, the third also 257! Parameters for x1 and x2 (and the rest) are all estimated ok, so the models actually have 262, 263 and 264 linear parameters resp., including _cons.

So clearly it's not that 256 is a limit, but something odd is happening.

(The 256 mentioned before came from the stepwise run, I think I had those the wrong way round.)

At 11:57 AM 6/02/2008, you wrote:

How about tossing a few more variables in - and see if the d.f. stays at 256? That would indicate whether there actually is a limit you are running up against.

Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME: (574)289-5227
EMAIL: Richard.A.Williams.5@ND.Edu

* For searches and help try:
*   For searches and help try:

Dr Keith B.G. Dear
Senior Fellow in Biostatistics
National Centre for Epidemiology and Population Health
Australian National University
Canberra, ACT 0200, Australia
Tel: 02 612 54865, Fax: 02 612 50740
CRICOS provider #00120C

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index