Jean Marie Linhart wrote:
Please note the definition for invbinomial() -- it is a bit unusual:
[Excerpted]
----------------------------------------------------------------------------
Thank you, Jean Marie, for clearing up my confusion. I actually now
remember the reference manual calling it a "twist," but its significance
didn't register at the moment.
There must be a story behind the different behaviors of -Binomial()-
and -invbinomial()- . . .
Joseph Coveney
P.S. I have an alpha version of -blakerci.ado- working. It's surprisingly
not much more conservative than the Blyth-Still-Casella interval (from
StatXact) in the limited evaluation that I've had time to do; quite similar,
actually. If it survives certification testing, then I'll forward a package
to Kit Baum. It yields a so-called exact binomial confidence interval
(guaranteed to have at least nominal coverage), one that is nevertheless
less conservative than the Clopper-Pearson interval. It should offer an
additional alternative to the several approximate binomial confidence
intervals already available in 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/
