A wald statistic is offered to test equality of proportions in matched samples when the probability of success of an observation in a given sample is constant within the stratum but possibly different between strata. Cochran''s Q statistic might be used to test this hypothesis, since it was used to test equality of proportions in matched samples when the probability of success within a sample is allowed to vary among observations and when this probability is assumed constant from observation to observation. A comparison is made between the Wald statistic and the above Q statistic. From simulation studies the Wald statistic seems to be more powerful than the Q statistic, even after adjustment for different significance levels. Wald statistics are also obtained for testing equality of proportions between strata and for interaction between the matched samples and the strata. A final Wald statistic is given which tests equality of the samples adjusted for interaction.