Hi Pushpender Nath Saini
Friday, June 1, 2007, 7:16:39 AM, You wrote:
PN> Dear Lister,
PN> I am using one way ANOVA followed by Scheffe's post hoc test for testing
PN> some hypotheses. Sometimes ANOVA gives significant results but all paired
PN> comparison results of Scheffe's test are not significnat. Why does this
PN> happen?
First, when you say "paired", do you mean "pairwise"? (quite different
idea). Assuming you meant "pairwise", here's the reply:
Because, answering the question you ask below, Schefee's post hoc is
the MOST conservative test. Use Tukey's HSD instead (good control of
the experimentwise type I error, but still a bit sensitive)
PN> Suppost out of 10 paired comparisons,
pairwise?
PN> I get significant results at 4
PN> combinations, but not significnat at rest six combinations. Should I say
PN> that hypothesis is largely rejected?
NO. The ANOVA overall test is used to reject the null hypothesis
(equality of ALL means). Once you have rejected it (simply "rejected",
not "largely" or "quite" or "scarcely"...), use a post hoc test to
find out which groups are significantly different from others.
PN> I have read that Scheffe's post hoc test is the most conservative test among
PN> all post hoc test for ANOVA. To what extent I am right?
The MOST conservative, as a matter of fact. Why did you choose it? Are
you very worried about the chance of a type I error? This test will
protect you against it, but, as you have seen with your data, you will
probably be faced with type II errors, due to the lack of power of
this method.
--
Regards,
Dr. Marta García-Granero,PhD mailto:
[hidden email]
Statistician
---
"It is unwise to use a statistical procedure whose use one does
not understand. SPSS syntax guide cannot supply this knowledge, and it
is certainly no substitute for the basic understanding of statistics
and statistical thinking that is essential for the wise choice of
methods and the correct interpretation of their results".
(Adapted from WinPepi manual - I'm sure Joe Abrahmson will not mind)
Locked
