Stephen
Thank you for the quick, clear and simple explanation. Given your last
statement, my two groups are carefully aged matched, and adults, so it
may not be necessary to use age as a covariate. The results change
minimally when the covariate is applied so it appears that the
majority of variance is accounted for by group membership.
Cheers,
Deryk
Deryk Beal, M.HSc., S-LP(C), Reg. CASLPO
Speech-Language Pathologist
Doctoral Candidate
[log in to unmask]
*************************************************************************
This email may contain confidential and/or privileged information for the
sole use of the intended recipient. Any review or distribution by others is
strictly prohibited. If you have received this email in error, please
contact the sender and delete all copies. Opinions, conclusions or other
information expressed or contained in this email are not given or endorsed
by the sender unless otherwise affirmed independently by the sender.
On Thu, Jun 25, 2009 at 10:42 AM, Stephen J. Fromm<[log in to unmask]> wrote:
> On Wed, 24 Jun 2009 18:22:51 -0400, Deryk Beal <[log in to unmask]>
> wrote:
>
>>Dear SPM Users
>>
>>I have some very simple questions regarding the factorial design
>>specification and contrast design for an independent 2-sample t-test with a
>>single covariate. Although many new users have posted on this topic before
>>the replies to their questions do not seem to have made it to the list and
>>therefore I can not find an answer to this question already posted.
>>
>>1. While building the statistical model: I have 2 groups of 12 participants
>>(control vs. patient) and a covariate of age. Do I need to enter 1 or 2
>>covariate vectors? What I mean is: Do I enter a vector of control_age (e.g.,
>>32 30 40 25.... 0 0 0 0....) and then a patient_age vector (0 0 0 0..... 24
>>29 36 39....) or just one vector (32 30 40 25.... 24 29 36 39....)?
>
> You could do either; it depends on how you want to model the situation.
>
> A single age covariate means you think the effect of age doesn't depend on
> the group.
>
> Two age covariates: the effect of age depends on the group. (You might
> say that a "group-by-age interaction" is being modeled.)
>
> You have to be careful with covariates, however; they might make it harder
> to create meaningful contrasts of the group effects.
>
>>2. While defining the contrast do I use simple 1 -1 and -1 1 or do I need to
>>change the definition somehow to include the covariate.
>>
>>My apologies for such a simple question but it does not seem to have been
>>spelled out on the list.
>>
>>Cheers,
>>Deryk
>>
>>Deryk Beal, M.HSc., S-LP(C), Reg. CASLPO
>>Speech-Language Pathologist
>>Doctoral Candidate
>>[log in to unmask]
>>*********************************************************
> ****************
>>This email may contain confidential and/or privileged information for the
>>sole use of the intended recipient. Any review or distribution by others is
>>strictly prohibited. If you have received this email in error, please
>>contact the sender and delete all copies. Opinions, conclusions or other
>>information expressed or contained in this email are not given or endorsed
>>by the sender unless otherwise affirmed independently by the sender.
>>
>
>
>
|