> From [log in to unmask] Thu Nov 26 19:17:00 1998
> Date: Fri, 27 Nov 1998 03:15:30 +0800
> Subject: Question of time x response interaction
> From: lordowen <[log in to unmask]>
Dear Chung-Haow,
> First, in the "time x response interaction" option in SPM96 statstic,
> there are two options: "as effect of interest", and "as only effect of
> interest" in "treat as". What's the difference between these two
> option?
In the second the responses per se (i.e. those that you are expecting
to show adaptation) are treated as confounds. This means they will be
removed from the adjusted data and that the ensuing SPM{F} pertains to,
and only to, the time x response interaction (as opposed to the
interaction and response itself)
> Second, after "treat as", I must be give a number in "apply of
> effect". What's these number working for, and what's their meaning?
They are simply the columns of the design matrix that specify the
effect that shows an interaction by time that you want to test for.
This may be a single regressor or a number of columns ,if you have used
basis functions.
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|