The Cochran Q test tests whether three or more sets of matched frequencies differ significantly. The data are assumed to be 0-1 responses by N individuals under K conditions. See S. Siegel and N. J. Castellan, Nonparametric Statistics, 2nd ed., 1988, pp. 170-174.
Be sure to CALL DFLT before calling COCHRANQ and to include the COMMON/BPRINT/ common block. The call is
CALL COCHRANQ(JRESP,N,K,Q) where JRESP = INTEGER array dimensioned JRESP(N,K) containing the responses (input) N = the number of individuals (input) K = the number of conditions (input) Q = single precision real variable containing the calculated Chi Squared value (output)