Vina, > Thanks. Someone (not on through this list) also suggested that I might be > able to code it [1 -0.05 -0.95] and have it be close enough. Could that also > work? Thanks much! > I don't know what that would buy you. I realize it seems weird, but if you're trying to tell if 3 numbers (i.e. beta1, beta2, beta3) follow a straight line, and you specifically don't care about the mean (i.e. you don't want your answer to be influenced by the magnitude of (beta1+beta2+beta3)/3), then the only information you need is on the difference of the extremes (beta1-beta3). -Tom ____________________________________________ Thomas Nichols, PhD Director, Modelling & Genetics GlaxoSmithKline Clinical Imaging Centre Senior Research Fellow Oxford University FMRIB Centre