Dear Maria,
For each covariate you specify, you will have to enter a vector with 69
values, ie for all subjects. In your case, you would have three
"Covariate" entries under "Covariates".
In recent versions of SPM12, you have a "Multiple covariates" option
that allows you to specify a file containing a matrix of 69x3 values for
all covariates simultaneously.
To look for interactions, you would have to create the corresponding
covariates by hand, eg if you have variables 'anxiety' and 'drugs', you
would compute (anxiety-mean(anxiety)).*(drugs-mean(drugs)) and enter
this as an extra covariate.
Best regards,
Guillaume.
On 24/08/17 17:30, Maria Tippler wrote:
> Hello SPM community!
>
>
> My study has 2 groups: high risk (n=33) and low risk subjects (n=36). I
> have a contrast map looking at the effects of a stimulus versus rest.
>
>
> At the 2nd level of analyses, I want to do an independent samples
> t-test, with group 1 being high risk and group 2 being low risk.
>
>
> I have 3 covariates: gender (0=male, 1=female), anxiety levels (range
> of 0-25 based on a questionnaire), lifetime number of drugs used (range
> of 0-500 based on self-report).
>
>
> How do I add in a covariate?
>
>
> Do I create a separate covariate module with one text file per covariate?
>
> Example:
>
> module/textfile 1: gender noted as 0 or 1 (n=69)
>
> module/textfile 2: anxiety levels (n=69)
>
> module/textfile 3: drugs (n=69)
>
>
> Or do I have to create a module for each group per corvariate?
>
> Example:
>
> module/textfile 1: high risk gender (n=33)
>
> module/textfile 2: low risk gender (n=69)
>
>
> Also, how could I look at the complex interaction effect of
> gender*anxiety or gender*anxiety*drugs?
>
>
> Thank you in advance!
>
> Maria
>
>
>
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|