The Wilcoxon test tests the null hypothesis two treatments applied to N individuals do not differ significantly. It assumes that the differences between the treatment results are meaningful. See S. Siegel and N. J. Castellan, Nonparametric Statistics, 2nd ed., 1988, pp. 87-95.
Be sure to CALL DFLT before calling COCHRANQ and to include the COMMON/BPRINT/ common block. The call is
CALL WILCOXON(N,A1,A2,ZVAL,IER) where N = the number of individuals (input) A1 = single precision array dimensioned A1(N) containing one set of measurements A2 = single precision array dimensioned A1(N) containing another set of measurements ZVAL = single precision real variable containing the calculated asymptotic normal statistic (output) IER = error return (=-3 input error,=0 only asymptotic test valid, =101 small sample test not significant, =100 small sample test significant)