Dynamic models of contraceptive use fill a theoretical gap. They bring behavioural content to the mathematically elegant Markov models of family building. They meet the criterion that Bulatao (1989) advances that method use, discontinuation and switching be considered not in isolation, but with reference to individual reproductive goals. Dynamic models help to link the literature on method choice with the literature on method effectiveness. These theoretical advantages are, at present, counter-balanced by empirical limitations. This article explores the theoretical and empirical underpinnings of dynamic contraceptive choice models.