This brief explores the complications involved in computing the divorce rate in the United States. Reasons for the complications include the federal government’s decision to stop collecting divorce data from States in 1996 and the reliance on the National Center for Health Statistics to publish an annual divorce rate, although several States (including California) do not participate in data collection or dissemination. The difference between the divorce rate and the cohort divorce rate is then explained. Finally, results from a study of cohort rates for various birth years are cited and indicate that the probability of marriages ending in divorce increased more or less continuously until 1990 and then stabilized. If separations that do not end in divorce are included, then the current rate of marital disruption is about 50%, a figure that has not declined during the last quarter century.