as for a summary ... i think that it would include the following:
1. for hypothesis testing to occur, there needs to be some null value
2. for the null, there is a model we assume about it and this is
HYPOTHETICAL ... sample data say something about whether this model makes
sense or not or is viable or not
3. the hypothesis test therefore is about the null but, this might be
stated 1 of at least 2 ways ... A: p value has something to do with the
evidence against the null ... B: p value has something to do with the
evidence of the sample data being discordant with A null ...
4. some raise the notion (as you have) that the null is bound to be false
... but, unless we don't take that position ... then null hypothesis
testing is worthless
5. some say null hypothesis testing is worthless anyway ... since it really
does NOT get at the question of real interest ... that being ... what IS
the parameter (or some estimate of it)?
6. i think there is agreement that no matter what one's perspective is on
NHT and p ... that p doesn't relate to "practical importance" ... in and of
itself, this is an important thing to stress and dissuade folks from
interpreting p as
7. most will agree i think that if ALL you do is some hypothesis test ...
you have not learned much (some say nothing but, i will say only that you
have not learned much)
you might have some other take on what some of the posts boil down to and,
perhaps i have "omitted" a point that i should have included (feel free to
edit!) or misstated a view ...
but it seems to me that these form the crux of what people have been saying ...
At 02:37 PM 1/31/01 +0000, you wrote:
>In article <[EMAIL PROTECTED]> you write:
> <...>
> >
> >"P values, or significance levels, measure the strength of the evidence
> >against the null hypothesis; the smaller the P value, the stronger the
> >evidence against the null hypothesis"
> >
> >my main questions of this are:
> >
> >1. does the general statistical community accept this as being correct?
> >
> >2. if the answer to #1 is yes ...
> >
> >then what does this tell us (only this p value) about what the real
> >parameter value is? (are)
> >
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
