On 2 Jul 2003 12:03:13 -0700, [EMAIL PROTECTED] (Dennis Roberts) wrote: > At 02:33 PM 7/2/03 -0400, Wuensch, Karl L wrote: > > For example, I just computed the Pearson r between statophobia (the > >first-day amount of self-reported fear of the course expressed by students > >who have taken stats with me the last several years) and the height of the > >students' ideal mate. Although "significant" at .003, the r was rather > >small, .13. The sample size was moderately large, 489. More fun is to see > >how it is that the students interpret this correlation -- why does > >statophobia covary with the height of ideal mate? > > > >Karl W. > > but, we know the null is not true ... so, why test it? what's the point and > utility of that step? >
In modern practice, perhaps we do look at the 'test' as a first step, because - occasionally - we can be surprised, despite having an N that is huge (relatively speaking). > the real question is: what might the rho be between X and Y (whatever that > is) ... NOT ... should we reject 0 ... > [ snip, rest] Right. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html "Taxes are the price we pay for civilization." Justice Holmes. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
