I am trying to model a set of data that I am working on using contrasts. Since the data set is too large, I am interested in scripting the entire process to arrive at the minimum adequate model. Has anyone in the group done this before? If you I would appreciate their comments. Cheers ../Murli