Hi Omar,
The design
> matrix I came up with, for randomize, is the following (for
> the sake of
> simplicity I will just represent the data with 1 control, 2
> replased
> subjects, and 3 abstinent subjects).
>
> Group/EV1(control time1)/EV2(control time2)/EV3(relapse
> time 1)/EV4 (relapse
> time2)/EV5 (abstinent time1)/EV6 (abstinent time2)/
> EV7(subject1)/EV8(sub2)/EV9(sub3)/EV10(sub4)/EV11(sub5)/EV12(sub6)
>
> 1/1/0/0/0/0/0/1/0/0/0/0/0
> 1/0/1/0/0/0/0/1/0/0/0/0/0
> 2/0/0/1/0/0/0/0/1/0/0/0/0
> 2/0/0/0/1/0/0/0/1/0/0/0/0
> 3/0/0/1/0/0/0/0/0/1/0/0/0
> 3/0/0/0/1/0/0/0/0/1/0/0/0
> 4/0/0/0/0/1/0/0/0/0/1/0/0
> 4/0/0/0/0/0/1/0/0/0/1/0/0
> 5/0/0/0/0/1/0/0/0/0/0/1/0
> 5/0/0/0/0/0/1/0/0/0/0/1/0
> 6/0/0/0/0/1/0/0/0/0/0/0/1
> 6/0/0/0/0/0/1/0/0/0/0/0/1
>
> (the first two lines represent the 1 control subject, lines
> 3-6 represent
> the 2 relapsed subjects, and lines 7-12 represent the 3
> abstinent subjects)
Seems ok to me
> Examples of contrasts:
>
> To compare t1 and t2 of the relapsers (simple t-tests):
>
> 0/0/1/-1/0/0/0/0/0/0/0/0
> 0/0/-1/1/0/0/0/0/0/0/0/0
That's right
>
> To compare t1 and t2 of abstinent subjects:
>
> 0/0/0/0/1/-1/0/0/0/0/0/0
> 0/0/0/0/-1/1/0/0/0/0/0/0
Still ok
> Here is where I am very confused, how do you model
> interactions? is the
> following correct:
>
> To model a 2x2 interaction with controls and relapsers:
>
> 1/-1/1/-1/0/0/0/0/0/0/0/0
> -1/1/-1/1/0/0/0/0/0/0/0/0
Not ok. Your question is where are the significant differences (-) between the changes in the controls (1 -1) and the relapsers (1 -1) i.e.:
(1 -1)-(1 -1)=1/-1/-1/1
>
> To model a 2x3 interaction with all groups:
> 1/-1/1/-1/1/-1/0/0/0/0/0/0
> -1/1/-1/1/-1/1/0/0/0/0/0/0
Not quite. Not sure here what you're asking for here... but if you want something like "overall is there any significant difference between the changes across groups", then you'd need an F-test...
> Finally, When I input these matrices into Glm, I get the
> following error
> messages:
>
> Problem with processing the model. Warning: at least one EV
> is (close
> to) a linear combination of the others. You probably need
> to alter
> your design. (Design matrix is rank deficient - ratio of
> min:max
> eigenvalues in SVD of matrix is 1.21792e-17).
>
> Warning-design matrix uses different groups (for different
> variances),
> but these do not contain "separable" Evs for the different
> groups (it
> is necessary that, for each EV, only one of the groups has
> non-zero
> values).
It is probably because you've run exactly the model above that is indeed rank deficient. You just need to add some subjects (and presumably you've got more than one control) as some additional EVs and it should not be a problem anymore...
Hope this helps,
Gwenaelle
|