On Sat, 22 Jul 2000 20:54:24 -0400, Mike Hewitt
<[EMAIL PROTECTED]> wrote:
< snip, details >
> What then does this mean as far as the results for the test x modeling
> result? I guess my question is does a higher-level interaction
> "overrule" a lower-level interaction?
>
I have seen one good answer.
But it did not ask about the relative size of the interactions. If we
are talking about the same DF, then if one effect is barely .05 and
the other is well under p of .001, then the interpretation is mostly
going to follow the smaller p. So, a huge main effect can basically
swallow a tiny interaction, or vice-versa.
Also, the answer did refer to graphing, which is a pretty good way to
discover a lot of problems and also explanations; but it did not
explicitly mention the live possibility of "bad scaling" which is
quite often responsible for apparent interactions. That is, if you
have 'ceiling' or 'basement' effects on scores, that can give you
significant tests for interactions, that are (in the sense of research
hypotheses) apt to be meaningless.... What is most of interest or of
concern are the interactions that *cannot* be explained as a function
of bad scoring/scaling.
--
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
