Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Stephen Sullivan <stephen_sullivan@brown.edu> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: Re: pprob |
Date | Sat, 2 Oct 2010 09:33:13 -0400 |
Hello. Could someone please provide some Stata advice with respect to "pprob"? I have a dataset, which does not appear to fit the Poisson distribution. I then ran glm and nbreg commands as suggested (http://www.ats.ucla.edu/stat/stata/library/count.htm). To see how the my variable fit the Poisson and negative binomial distributions, I downloaded and used the nbvargr command. Nevertheless, something seems strange as all pprob values for the Poisson are 0. The output is: . nbvargr total Obtaining Parameter Estimates (9 observations deleted) here Negative Binomial Probabilities with mean = 47.92857 & overdispersion = .0867781 +------------------------------+ | k nbprob nbcum | |------------------------------| 1. | 0 0.00000001 0.00000001 | 2. | 1 0.00000006 0.00000006 | 3. | 2 0.00000029 0.00000035 | 4. | 3 0.00000105 0.00000140 | 5. | 4 0.00000307 0.00000446 | |------------------------------| 6. | 5 0.00000767 0.00001214 | 7. | 6 0.00001703 0.00002917 | 8. | 7 0.00003438 0.00006355 | 9. | 8 0.00006417 0.00012772 | 10. | 9 0.00011222 0.00023994 | |------------------------------| 11. | 10 0.00018568 0.00042562 | +------------------------------+ k was byte now float Poisson Probabilities for lambda = 47.92857 +------------------------------+ | k pprob pcum | |------------------------------| 1. | 0 0.00000000 0.00000000 | 2. | 1 0.00000000 0.00000000 | 3. | 2 0.00000000 0.00000000 | 4. | 3 0.00000000 0.00000000 | 5. | 4 0.00000000 0.00000000 | |------------------------------| 6. | 5 0.00000000 0.00000000 | 7. | 6 0.00000000 0.00000000 | 8. | 7 0.00000000 0.00000000 | 9. | 8 0.00000000 0.00000000 | 10. | 9 0.00000000 0.00000000 | |------------------------------| 11. | 10 0.00000000 0.00000000 | +------------------------------+ (11 observations deleted) Any thoughts on what I may have done wrong? Why all the zeros for the poisson prob? Do I need to somehow install pprob? Thank you so much for any help you can provide. Stephen Stephen R. Sullivan, MD, MPH Assistant Professor, Surgery and Pediatrics Brown University Plastic Surgery Rhode Island and Hasbro Children's Hospitals * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/