Abstract
Temporal variations in conventional fertility measures reflect the operation of instrumental variables: quantitative and temporal intentions; success in achieving intentions; and reproductive conditions. A set of such variables is described, using data from the 1975 National Fertility Study. There was a large decline in the number of intended conceptions, a recent large rise in the extent of their delay, a very large decline in rates of failure to delay or terminate fertility, and a very large recent rise in sterilization. But one problem proved important and intractable: When the data source is a cross-sectional survey, the length of open interval is inherently different for real and for synthetic cohorts, it is strongly related to reproductive intention, and that affects the classification of exposure to risk in the open interval.
