I found the comment by Corcoran and Mehta particularly compelling. The
inverted LR test looks better to me than all the alternatives. On a related
topic, did anyone come up with a way to estimate confidence intervals for
proportions in -svyset- data that would be robust to cells with no observed
successes?
-----Excerpted Message-----
From: Ron�n Conroy [mailto:[email protected]]
Sent: Wednesday, September 08, 2004 11:56 AM
To: [email protected]
Subject: Re: st: Binomial confidence intervals
Reading the paper and examining their graphs, it's really much of a
muchness between Wilson and Jeffreys, which seem to suffer less from the
lucky/unlucky p and N combinations than the Agresti. Agresti certainly
gets third place, but did anyone reading the paper spot a compelling
advantage of Wilson over Jeffreys, or the other way around? I didn't.
*
* 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/
