method 1:
in your design, determine the specific comparisons (paired or otherwise) that you REALLY want to make ... pre plan these comparisons and go after them ... in this scenario ... the notion of doing your omnibus ANOVA first and then only "following up" IF you find significant F ratios ... is out the window ... a priori arrangements
method 2:
if you have no particular comparisons that are more important than others ... you might do your omnibus F tests first and THEN if you reject the null .... go looking for WHERE differences might be ... in this scenario ... you don't do follow up tests unless the first omnibus screening F tests are significant ... post hoc arrangements
both methods have their supporters and opponents ...
i would say that a general weakness of method 2 is that, sometimes if you fail to reject the original F tests ... and stop ... you miss some differences that could be "significant' or, "important" ... but, in method 2, you never give yourself the chance to see them
on the other hand, method 1 has opponents too in that, for many designs ... one "could" only look for differences that are no brainers (ie, are bound to be significant) and ... fail to look at other comparisons that MIGHT cast doubts on your theories or propositions ... so, you just avoid looking for them
At 06:34 AM 12/19/2002, Ivan Balducci wrote:
Hi members,
Please, may someone explain to me this doubt:
why is not right to perform a Tukey HSD test
after an ANOVA came out non-significant ?
Why Tukey HSD test should be done only following an ANOVA that was significant ?
Which is the reason ?
TIA,
Ivan Balducci
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.427 / Virus Database: 240 - Release Date: 12/6/02
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
_________________________________________________________ dennis roberts, educational psychology, penn state university 208 cedar, AC 8148632401, mailto:[EMAIL PROTECTED] http://roberts.ed.psu.edu/users/droberts/drober~1.htm
. . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
