A FORTRAN computer program was written to select the norms and coefficients of variation (CV) for DRIS. The program splits the population into a high and low yielding subpopulation and then calculates the mean, variance and standard deviation for each element and form of expression (ratio and product of each element to all others). Based on the significance of the variance ratio (low yield/high yield), one then selects the form of expression which has the highest variance ratio and uses the corresponding high yield means and CV's as the norms from which DRIS indices are calculated.