This would seem like fairly straight forward thing to do but I ma having problems. I have variable SS which has three levels high. med. low. My hypothesis is that those with Low and moderate SS scores will have higher mean speeds (dependent variable). I want to compare high to low using speed as the dependent variable. HOW? I tried independent T-test but to no avail. Another question could someone explain to me why it better to use Niles to divide your group into three equal groups that using 33/66/99 percentiles.

Take care, Talino