On 2 Sep 2003 09:10:54 -0700, [EMAIL PROTECTED] (Donald Burrill) wrote: > As Joe has somewhat more gently pointed out, "interactions between the > DV and covariates" do not make a whole lot of sense. < snip, rest >
Those interactions don't make sense for ANOVA. On the other hand, I was confused the first time that I ran into the conventions of 'log linear models.' - The *only* things that are not called interactions in the loglinear models are the silly little Main Effects. So, I have some sympathy for this difficulty with the words, since we statisticians seem to have created some contradictory descriptions. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html "Taxes are the price we pay for civilization." . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
