I think Mike meant to say that p is only meaningful if you remember that it
is conditionalized upon the null being true.  Even if the null is never or
almost never true, p as an abstract quantity can be meaningful, just like a
sampling distribution is a meaningful thing even if it is never obtained.
This is not to say that there are not even more meaningful things we could
do with the measurements from which we typically derive these p values.

I disagree, however, with the assertion that replication (if that means
repeating the research) is only useful if the null is rejected.  With
replication the base upon which a meta analysis can be conducted grows,
allowing greater precision (less wide confidence intervals) when estimating
effect sizes.  Please keep in mind that a meta analysis can detect a
non-zero effect size from a collection of studies each of which individually
resulted in retention of the null.
----- Original Message -----
From: "Jean-Marc Perreault" <[EMAIL PROTECTED]>
To: "Teaching in the Psychological Sciences" <[EMAIL PROTECTED]>
Sent: Tuesday, November 12, 2002 5:56 PM
Subject: RE: p is continuous, not dichotomous


Mike, could you expand on how it is that a p-value is
only meaningful if the Null is true? I understand the
second part of your statement (Replication only useful
is Null is rejected)... but would like more info on
the first...

cheers!

Jean-Marc



--- Mike Scoles <[EMAIL PROTECTED]> wrote: >

A p-value is only meaningful if the null hypothesis
is true.  Replication is only  meaningful if the null
is false.




>
> > -----Original Message-----
> > From: Martin J. Bourgeois [mailto:MartyB@;uwyo.edu]
> > Sent: Tuesday, November 12, 2002 1:30 PM
> > To: Teaching in the Psychological Sciences
> > Subject: RE: p is continuous, not dichotomous
> >
> >
> > Maybe I should quit before I get too far behind,
> but what I'm trying to
> > say (and apparently failing) is that an observed
> difference between
> > means is more likely to be replicated when the p
> is .001 than when the p
> > is .1. You can certainly calculate the probability
> of replicating a
> > result with a given p value, and results with
> smaller p's are more
> > likely to be replicated (yes, it has been
> supported by data). I'll dig
> > up a reference when a get a chance.
>
>
> ---
> You are currently subscribed to tips as:
> [EMAIL PROTECTED]
> To unsubscribe send a blank email to
> [EMAIL PROTECTED]
>
>

=====
Jean-Marc Perreault
Arts & Sciences
Yukon College
Whitehorse, Yukon
867-668-8867

______________________________________________________________________
Post your free ad now! http://personals.yahoo.ca

---
You are currently subscribed to tips as: [EMAIL PROTECTED]
To unsubscribe send a blank email to [EMAIL PROTECTED]


---
You are currently subscribed to tips as: [EMAIL PROTECTED]
To unsubscribe send a blank email to [EMAIL PROTECTED]

Reply via email to