On 16 Oct 2003 09:01:36 -0700, [EMAIL PROTECTED] (Enda Kelly) wrote: > Hi. I have a query regarding whether it is logical to place a confidence > interval about a P-value. The computation involved uses a permutation > method to produce a P-value for a hypothesis test. In an effort to check > the reliability of this P-value I have been bootstrapping the raw data > to produce a confidence interval about this P-value. Curiously, for > significant P-values the CI is in general quite narrow, whereas for > non-significant values, I get enormous CIs.
You are observing 'proportions'. The simple SD (ignoring N) for a proportion is sqrt(p *(1-p) ) ; so the SD for 0.50 is 0.50 . By that formula, the SD for 0.0001 is 0.01. In bootstrapping, of course, you would get CI limits that would not be symmetrical at the extremes. > This leads me to the > suspicion that there is something flawed about the process. Am I correct > in my suspicions? Permutations are used to get *exact*, counted limits. Actual "bootstrapping" has been used to approximate those numbers. There are proper was to talk about doing those things. So, I am not entirely happy with your use of the terminology, and I *suspect* that you are misinterpreting the natural, expected sort of results. But I can't be sure of that. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html "Taxes are the price we pay for civilization." . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
