The correspondence between intention to avoid childbearing and subsequent fertility: a prospective analysis

CONTEXT: Retrospective studies of pregnancy intendedness have revealed some characteristics that can help identify which women are more likely than others to experience an unintended birth. A comparison of these findings with those from a prospective analysis may shed greater light on the characteristics associated with unintended pregnancy. METHODS: Data were taken from the 1988 National Survey of Fertility Growth and a telephone reinterview of respondents conducted in 1990. Separate analyses were conducted of women intending to postpone childbearing for at least three years and of women intending to forgo all future childbearing. Logistic regression models were used to identify the effects of social and demographic characteristics, as well as change in marital status and certainty of intentions, on the odds of experiencing a birth in the interval between interviews. RESULTS: Only 10% of women intending to postpone pregnancy for more than three years and 8% of respondents seeking to forgo future childbearing had a birth in the interval between interviews. (These births, referred to as unpredicted births in this article, are roughly analogous to those labeled unintended in retrospective analyses.) Women with incomes below the poverty level were 2-3 times as likely as women with incomes between 100% and 199% of poverty to experience an unpredicted birth. Race was not a significant factor among women intending to avoid future childbearing, and became nonsignificant among those intending to postpone when change in marital status and contraceptive status were taken into account. Women aged 35 and older who wanted no more children were significantly less likely than women aged 20-29 to have an unpredicted birth. Women aged 30-34 who wanted to postpone childbearing were roughly 70% less likely than women aged 20-29 to experience an unpredicted birth. Overall, women who were at risk for a pregnancy but not practicing contraception were 2-3 times more likely than women using an effective method to have an unpredicted birth. CONCLUSIONS: There are at least two potential explanations for instances where the correlates of unintended births in the prospective analysis differ from those identified in retrospective studies. Certain subgroups of women may be more likely to classify births as wanted when they are asked retrospectively; alternatively, they may be more likely to experience changes in their living conditions that alter their fertility intentions.