Stata colleagues:
Is there a particular reason why the list of diagnostic options available
with -predict- after poisson and negative binomial regression is so
limited? I can get a much more complete set of diagnostic parameters
using -predict- after -glm-. This approach is fine for poisson models, but
(as the Stata help points out) replicating -nbreg- results using -glm- is
difficult.
Are there any differences in the way that regression diagnostics are
computed after -glm- to how they would be computed after -poisson-, if
they were available (eg. the difference in residuals between -glm-
(computed 1 per observation) and -logit- (computed 1 per covariate
pattern))?
Thanks in advance for any help with this.
Cheers
Ian Dohoo
--------------------------------------------------------------
Ian R. Dohoo,
Professor - Epidemiology,
Department of Health Management
Atlantic Veterinary College
University of P.E.I.,
Charlottetown, P.E.I. C1A 4P3 CANADA
e.mail <[email protected]> phone 902-566-0640 FAX 902-566-0823
"Life is full of obstacle illusions" ... Grant Frazier
--------------------------------------------------------------
*
* 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/
