In response to recent questions about the relative merits of sampling from different groups of births when studying the birth interval and its components, we derive analytically the implications of sampling births according to several different schemes and of estimating the distribution of waiting times from both current status and retrospective history measures. Breastfeeding is used as an example to illustrate that the distributions of weaning times estimated from the open birth interval or from the open and last closed birth intervals are not the same as the distributions of weaning times in the population, either for mothers or for infants. Furthermore, the distribution of weaning times estimated from current status data differs from that estimated from retrospective history data. If, however, a sample of all births that occur in a fixed period or a sample of all births of a given parity is drawn, then the distribution of weaning times estimated from both current status and retrospective history data is the same as the distribution of weaning times for infants in the population.