Prev: how to produce color output for "ods graphics"?
Next: Predictive Analytics for Business, Marketing and Web 2010
From: clyde013 on 15 Mar 2010 16:14 Hello, I have a data set similar to the following: dummy NJ NH 0 3 7 0 7 8 0 6 9 1 4 6 1 5 7 1 6 3 proc ttest data=sum; paired nj*nh; run; I would like to do a paired t-test because NJ and NH are dependent. But I need to compare them by the dummy variable. In a regular t-test I would use a class dummy statement but I cannot use a class variable in a paired t-test. Is there an alternative way to calculate a paired t-test?
From: Paige Miller on 16 Mar 2010 15:06
On Mar 15, 4:14 pm, clyde013 <clyde...(a)yahoo.com> wrote: > Hello, > > I have a data set similar to the following: > > dummy NJ NH > 0 3 7 > 0 7 8 > 0 6 9 > 1 4 6 > 1 5 7 > 1 6 3 > > proc ttest data=sum; > paired nj*nh; > run; > > I would like to do a paired t-test because NJ and NH are dependent. > But I need to compare them by the dummy variable. In a regular t-test > I would use a class dummy statement but I cannot use a class variable > in a paired t-test. Is there an alternative way to calculate a paired > t-test? proc ttest data=sum; paired nj*nh; by dummy; run; -- Paige Miller paige\dot\miller \at\ kodak\dot\com |