What is the difference between omnibus and univariate W-S effects in repeated measures?
Is the only difference that thehh latter assumes sphericity? I ask because the the omnibus and W-S are identical when only one repeat is taken, ie., no sphericity needed. I also notice most software prints out an error term in the W-S but not in the omnibus. Most importantly Is there a way to trick the omnibus into assuming sphericity and producing the same result as the W-S. Or some conceptual explanation as to why univariate is called "univariate".. Thanks. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
