Hi On Sun, 9 May 2004, DG wrote:
> I have a 3x3x3 within subjects design. I am doing a test of linear contrast > to compare performance between cells. I want to compare 3 cells with the > rest of the 24 cells. I am using SAS to perform the linear contrast however > I end up with no values for F and p in the output. > > Can someone please check my SAS code and tell me where I am going wrong. I > have pasted both my input and output. Rich has pointed out the clear problem with your dataset given the desired analysis. But ignoring that, a contrast for all within-subject factors can be done simply by computing for each of your cases a score that represents the contrast for each case. I don't use SAS that often, but in spss, it would be compute contr = 8*y1+8*y2+8*y3-1*y4 .... -1*y27 Then use either t or F to test whether the mean contrast score differs significantly from 0, the null hypothesis value. If you actually want to reproduce the ANOVA values (i.e., SSs, MSs, as well as the final F, which would equal t^2), then you normalize the coefficents before computing the contrast score (i.e., divide original coefficients by the square root of the sum of coefficients squared). Best wishes Jim ============================================================================ James M. Clark (204) 786-9757 Department of Psychology (204) 774-4134 Fax University of Winnipeg 4L05D Winnipeg, Manitoba R3B 2E9 [EMAIL PROTECTED] CANADA http://www.uwinnipeg.ca/~clark ============================================================================ . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
