A procedure for applying the results of an analysis from one sample of subjects to a new sample of subjects to evaluate how well they generalize. It is frequently used in regression.