Estimation of the level and timing of fertility control from reports on children ever born

This paper proposes a new procedure for estimating the level and timing of fertility control within marriage as well as the social and economic correlates of these reproductive parameters. The method precludes the need for data such as trends in birth intervals and age at last birth and other data that is often not available. The method does require data on children ever born, age at marriage, and the socioeconomic characteristics of couples, information often gathered in censuses and surveys. Some assumptions of the model are: in the absence of deliberate fertility control, a woman can reproduce at each age according to a natural fertility schedule. A couple can choose the date after marriage at which control begins, and the level at which it is controlled. Application of this procedure to samples of US women aged 45-54 from the 1965 National Fertility Survey and women aged 20-44 from the 1970 National Fertility Survey yielded results that were consistent with analyses of special surveys on contraceptive use. In particular, for the sample of older women, it was estimated that the control of fertility began relatively soon after marriage, with control beginning at later durations for younger women, and that the extent of control was quite high, with the average level of control of non-Catholics exceeding that of Catholics. The evidence of early control is consistent with other research; specifically, the 1960 Growth of American Families Study. For the sample of younger women, it was estimated that the duration of marriage after which control began was longer for older women and for Catholics, but that the extent of fertility control, once it began, did not differ by religion.