In article <[EMAIL PROTECTED]>,
Robert Dodier <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] (Warren Sarle) wrote:

>> While exclusive reliance on p-values is often foolish, so is
>> the idea of banning them. [...]

>Well then, when should we use p-values, and when is it foolish to do so? 

>If exclusive reliance is foolish, what form of partial reliance is wise?

If one can believe the point null hypothesis, or know that 
the alternatives for which one should accept are close 
enough that it makes little difference, then one can
compare the p-value with the evaluated consequence of
falsely taking the "acceptance" action when the state of
nature is such that one should not.

It is also the case that this can be used to obtain a
procedure which is robust against some misspecifications
of the problem, and the risk for improper acceptance can
only be approximated.

The expected value of the p-value can also be used in
deciding posterior expected risk, and thus in deciding
whether experiments should be informed.


-- 
This address is for information only.  I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
[EMAIL PROTECTED]         Phone: (765)494-6054   FAX: (765)494-0558
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to