Were you to define p as the probability of getting data exactly as discrepant with the null as those you obtained, given the null, then, assuming you are dealing with a continuous variable, that probability is always going to be quite small, eh? About as small as the probability of the null being true. I've never seen a true null, why should I make all this fuss about testing it? ;-) ----- Original Message ----- From: "Jerry Dallal" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Thursday, March 20, 2003 2:18 PM Subject: Re: What is a P value?
"Wuensch, Karl L" wrote: > > Simple explanation: The probability of obtaining data as or more discrepant > with the null hypothesis than are those in the present sample, assuming that > the null hypothesis is absolutely correct. > > OK, I waiting for you all to tear my simple explanation to shreds. Why should I care about data more discrepant than what I've observed? I haven't seen them. Why should they affect the way I judge what I did observe? :-) . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
