Yulia Marchenko wrote:
[Abridged]
Sample size = 613 Exact power:.9131
Sample size = 610 Exact power:.9159
Sample size = 580 Exact power:.9001
The above results support the guess about other packages reporting sample
size for one-sided test.
-------------------------------------------------------------------------------
That was my guess, too. But the decrease in power with the increase in sample
size from 610 to 613 reminds of an alternative guess: have you considered that
the other packages might be using a different exact confidence interval to
invert, such as the Blyth-Still-Cassela (StatXact) or the Blaker (S-Plus)?
I implemented the latter in a do-file for a one-off use last year. If I can
get time to dig through my CD-R/W collection and find it to polish it into an
ado-file, I'll try to post it. The Blaker confidence interval begs the
question of sidedness.
Joseph Coveney
*
* 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/
