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

st: re. Poisson Regression Goodness of Fit Tests


From   Michel Camus <[email protected]>
To   [email protected]
Subject   st: re. Poisson Regression Goodness of Fit Tests
Date   Fri, 03 Oct 2003 09:09:50 -0400

First, you should have typed "search overdispersion , all". |:-()
-nbreg- is your solution. Simply replace "poissson" by "nbreg" in your model, then check the "Likelihood-ratio test of alpha=0".
This is one real test for overdispersion. [ Is this not easy enough relative to SAS? /:-) ]
Still, your extreme -poisgof- GOF chi2 indicates that the Poisson regrssion model is inappropriate.
Read -nbreg- section in Stata Reference Manual N-R.
You may want to explore reasons for overdispersion by using -gnbreg-.
Read FAQ for both overdispersion and zero-inflated data if this applies: http://www.stata.com/support/faqs/stat/nbreg.html

Michel Camus
===========
roger webb wrote:


Dear Statalist

I'm generating Poisson regression models with an aggregated data set (i.e. each record in the data set represents a stratum of aggregated numbers of deaths and person-years of observation).

I wish to check that the models are not over-dispersed. The manual tells me that I can use either the 'poisgof' or the 'poisgof, pearson' command. These produce the following contradictory results:

poisgof

Goodness-of-fit chi2 = 1191.579
Prob > chi2 (5304) = 1.0000

poisgof, pearson

Goodness-of-fit chi2 = 29207.21
Prob > chi2 (5304) = 0.0000

A colleague has told me that these results have no meaning for my data set, because the degrees of freedon are incorrect (I think). He says that I should instead apply the Breslow adjusted score test (Breslow NE. Generalized linear models: checking assumptions and strengthening conclusion. Statistica Applicata 1996; 8: 23-41).
My colleague says this test is unavailable in standard stats packages, but he has programmed SAS to perform the test. He thinks it would take me 2 to 3 weeks to write a similar programme in Stata. Can anyone please advise me further on this? Has anyone attempted to programme Stata to run the adjusted score test? Are there any alternatives?

Thanks in advance

Roger Webb
University of Manchester (UK)
+44 161 275 0728
*
* 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/


--

~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~

Michel Camus, Ph.D.

�pid�miologue, Div. Biostatistique et �pid�miologie, DGSESC, Sant� Canada

Epidemiologist, Biostatistics and Epidemiology Div., HECSB, Health Canada

Courriel / e-mail : [email protected] <mailto:[email protected]>

T�l�phone / phone : (514) 850-0157

T�l�copieur / fax : (514) 850-0836

~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~




*
* 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 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index