- if you don't figure out that there are various ways to use p-values, you might end up as negative and silly as Dennis, here -
On 19 Mar 2003 07:31:30 -0800, [EMAIL PROTECTED] (dennis roberts) wrote: > actually, p values are rather useless (i am almost prepared to say > "useless") since, it would be the RARE case when the null is REALLY exactly > true - "Exactly" depends on how you state the null; one-sided can be true quite easily - Even if we know that an effect is not absolutely zero, p-values are one report on the strength of the evidence. > > thus, in 99.9999% of the cases ... we KNOW the null is not true so, setting > some cutoff for rejection and then actually rejecting the null ... what has > this added to our knowledge? - the strength of the evidence? Whether *chance* might be sufficient to account for what has been observed? (We do want to keep track of 'nominal p-values' -- the tabled value is not the whole story when multiple tests are performed, or assumptions are violated.) > and, p values don't speak to the notion of the null being "approximately" true > - If you philosophize that everything is "approximately" true/ untrue, p-values are a way to measure it; not the whole story, but salient. I suggest for study, Robert P. Abelson's 1995 book, "Statistics as Principled Argument". -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
