tests of within subjects contrasts vs pairwise comparisons

#1
What is the difference between tests of within subjects contrasts vs pairwise comparisons given in the SPSS output when running ANOVA?

It seems one is simply comparing levels of the within subjects factors and the other the between subjects factors, but what analysis is being run in the within subjects contrasts - some kind of paired t test?

SPSS does not give so many options in the contrasts tests (e.g. repeated, polynomial) in comparison to the post hoc comparison tests. Another niggle I have is that SPSS does not give ALL the possible contrasts. It would be useful to know how to produce a table of all possible contrasts...

Thanks for your help!
 

Karabiner

TS Contributor
#3
What is the difference between tests of within subjects contrasts vs pairwise comparisons given in the SPSS output when running ANOVA?
Contrasts are linear contrasts (see the SPSS help for some
introductory information), while pairwise comparisons are
post-hoc tests, which account for multiple testing.

Wth kind regards

K.
 
#4
Thanks for your reply. I understand that post hoc tests adjust for multiple testing, e.g. bonferonni correction, but don't the contrast tests do the same? I would have thought so because you can run multiple contrasts, and therefore you should adjust p values accordingly right?

I would have likened the post hoc tests to independent t tests and the contrasts to paired t tests, but perhaps this thinking is incorrect.