This bulletin reports a likelihood ratio test of the significance of the difference between correlated proportions developed by Dr. S.S. Wilks. A test of the independence of the variables yielding these proportions on the assumption of equal difficulty is also presented. It is also shown that the tests of the significance of differences between correlated proportions used in the literature are in fact ways of estimating a conditional probability. Computational formulae and a computational layout as well as an illustrative example are included.