Discussion:
spss post hoc test
(too old to reply)
madhushi dasanayaka
2021-03-25 07:28:21 UTC
Permalink
I want to know that what are differences of post hoc tests in spss and what is the relevant test when sample size equal or not equal
Rich Ulrich
2021-03-26 22:22:18 UTC
Permalink
On Thu, 25 Mar 2021 00:28:21 -0700 (PDT), madhushi dasanayaka
Post by madhushi dasanayaka
I want to know that what are differences of post hoc tests in spss and what is the relevant test when sample size equal or not equal
Check out textbooks for the differences. These tests were not
invented by SPSS.

Read up on SNK; Least Significant Differences; and Duncan's
(larger control group vs multiple smaller groups). The first two are
the most used ones, and Duncan's is for that special case.

Duncan's is an alternative to the main test, not "post hoc".

Most of the tests for special contrasts were developed with
the assumption of equal Ns, and the versions that followed
for unequal Ns are approximations which are pretty good
so long as the Ns are pretty equal. They are less-good for
greater deviations from equal, so take that into account in
your reporting. Ns within 10% are "pretty equal" (my standard).
--
Rich Ulrich
Loading...