On 25 Nov 1999 04:56:04 -0800, [EMAIL PROTECTED] (Donald F.
Burrill) wrote:
< snip, much >
> I might be more impressed with the idea of "ordinal" and "disordinal" if
> either I could see a useful application of them in interpreting two-way
> data (I cannot see reporting to a client that in his data the interaction
> is ordinal (or dis-); he'd just ask what the %#& I meant by that, and
< snip, rest >
I don't have a formal definition, and it seems like the original
question was based on a different idea -- but here is one thing that I
thought was implied by disordinal interaction: The actual values of
the cell means can *not* be explained by *bad scaling* of the
variable.
That is the only way that I have used 'disordinal' - "Yep, there must
be something different going on, because it surely is not caused by
scaling." And I have seen plenty of interactions caused by scaling,
for instance, when analyzing 0-4 item scores, in samples that happen
to be 50 or 100 bigger than what is needed for a significant F.
This is consistent with what Donald says, about the appearance of cell
means being disordinal once you have subtracted out the main effects
(so, why would you want to do that?).
--
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
