Prev: GJR model in SAS - Help!
Next: Multiple windows to multiple homes
From: greenwillow on 12 May 2010 15:39
I am comparing pre and post dependent matched patients samples.

However, I need to compare more than 1000 peaks. If I use the
following codes to do the paired t test, that means I need to create
more than 1000*2 variables first. Is that possible not to create so
many variables? Can we do similar thing in independent comparison,
where we can use CLASS then VAR.

PROC TTEST DATA=masil.drug;
PAIRED propeak1: postpeak1;
RUN;

Thank you very much for your help!
From: data _null_; on 12 May 2010 16:16
On May 12, 2:39 pm, greenwillow <yangliuy...(a)gmail.com> wrote:
> I am comparing pre and post dependent matched patients samples.
>
> However, I need to compare more than 1000 peaks. If I use the
> following codes to do the paired t test, that means I need to create
> more than 1000*2 variables first. Is that possible not to create so
> many variables? Can we do similar thing in independent comparison,
> where we can use CLASS then VAR.
>
> PROC TTEST DATA=masil.drug;
>    PAIRED propeak1: postpeak1;
> RUN;
>
> Thank you very much for your help!

Make the data have PEEK 1:1000 and variable PRE and POST an then do
PROC TTEST by PEEK.
From: greenwillow on 12 May 2010 16:41
However, there are around 16 matched patients, which means pre and
post should be subset to 16 variables each.

On May 12, 3:16 pm, "data _null_;" <datan...(a)gmail.com> wrote:
> On May 12, 2:39 pm, greenwillow <yangliuy...(a)gmail.com> wrote:
>
> > I am comparing pre and post dependent matched patients samples.
>
> > However, I need to compare more than 1000 peaks. If I use the
> > following codes to do the paired t test, that means I need to create
> > more than 1000*2 variables first. Is that possible not to create so
> > many variables? Can we do similar thing in independent comparison,
> > where we can use CLASS then VAR.
>
> > PROC TTEST DATA=masil.drug;
> >    PAIRED propeak1: postpeak1;
> > RUN;
>
> > Thank you very much for your help!
>
> Make the data have PEEK 1:1000 and variable PRE and POST an then do
> PROC TTEST by PEEK.

 | 
Pages: 1
Prev: GJR model in SAS - Help!
Next: Multiple windows to multiple homes