On Thu, 9 Apr 2009 15:38:42 -0700, Olga Prilipko <[log in to unmask]>
wrote:
>Dear all,
>
>I have a basic question concerning the two-sample t-test analysis where I
have 20 subjects in group1 and 10 in group 2: is it correct to specify the
contrasts 1 -1 and -1 1 or should the contrast be corrected for the uneven
group size (n1/(n1+n2) -n2/(n1+n2))?
>Thank you very much!
>
>Olga
You do _not_ correct for differences in group size. Use e.g.
1 -1
|