Examined the validity coefficients of all possible WISC-R short forms of several subtests. Comparisons were made between coefficients given by McNemar's and Silverstein's formulas to determine "best" short forms for different uses. Results indicated only a slight difference between short forms selected by the two methods. (Author)