When evaluating the results of infertility therapy, it is often of interest to know whether differences exist between a new form of therapy and the established one, between medical and surgical therapy, or between patient groups that differ in severity of disease, age, parity, and so on. In this paper, a likelihood-ratio test is developed to evaluate differences between cumulative pregnancy curves of different patient groups. The test takes into account variation in patient follow-up and can be applied to any type of infertility therapy. It is illustrated by consideration of a group of patients who underwent artificial insemination by donor (AID) and determination of whether pregnancy outcome differed according to parity. It was found that the cure rate for both nulliparous and parous patients was virtually 100% and that the monthly probability of pregnancy was not significantly different between the two groups.