New evidence on the value of birth expectations.

Past national surveys regarding birth expectations have usually been restricted to currently married women, a fact which has led demographers to question the usefulness of these data. Because the June 1976 Current Population Survey includes the expectations of all women in a cohort regardless of marital status, it provides the data needed to evaluate biases due to restricted survey universes. At older ages, where there are substantial differences in lifetime expectations between currently married and single women, there are relatively few single women; at younger ages, however, where the proportion of single women in a cohort is relatively large, the differences in expectations are small. This counterbalancing effect makes the lifetime birth expectations of currently married women a close approximation of all women in a cohort. The analysis also indicates that the observed intracohort declines in lifetime birth expectations since 1967 were due largely to the addition at subsequent survey dates of previously unmarried women; nevertheless, some "true" cohort declines also seem to have occurred.