Hi Stephen,
> >Well I agree with the low frequency as I said in the previous e-mail..
> >but for A and B separately I really don't 'see' the prblm. Why, taking
> >your example, estimates would be -1 and 1? Given that A and B are
>
> (I meant that they wouldn't be -1 and +1, but definitely - and + and
> nonnegligible.)
>
> I didn't actually run a regression, but just drew a mental picture. If A,
> r, and B are each represented by a series of boxcars, then the closest fit
> to the line is given with beta_r = 0, beta_A = -M, beta_B = +M, where M is
> some (non-negligible) positive number.
well, the basic idea behind such design is to compare A and B. When looking at
the frequency of the contrast when one can see that A and B have similar power.
The design is ok (but possible to do better)
have a look at the script I send you as separate e-mail to demonstrate that...
best
cyril
-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/
-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/
|