Dennis Roberts <[EMAIL PROTECTED]> wrote: : At 07:06 PM 11/21/02 +0000, Jerry Dallal wrote: :>Confidence intervals are not betworthy.
: i disagree i guess "Betworthy" means one thing and one thing only. If you consider a CI to be betworthy, that means you are willing to generate a CI, let me look at it, and let me choose whether I want to bet on it or against it at nominal odds (since it's betworthy, you should be indifferent to my choice because the expectation appears to be always 0). I can make money doing this, not only for the example I already gave, but also for the two sided CI for a normal mean based on the t distribution. Confidence is not probability. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
