over time and across varied social groups. When an unmarried high school sophomore reports that her pregnancy occurred too early, her assessment of ''too early" probably means something different from the report of a married engineer who wanted a baby after working for several years but got pregnant her first year on the job.
Information on the demography of unintended pregnancy—the subject of this chapter—is dominated by data from the National Survey of Family Growth (NSFG), a federally-sponsored survey that has developed quite specific terminology and definitions to measure "unintended pregnancy." Over the past four decades, a series of questions has been regularly asked of women by this survey and its predecessors, the Growth of American Families surveys and the National Fertility Studies, in an effort to learn more about women's plans and intentions at the time they became pregnant. In these surveys, women are asked about pregnancies during the previous 5 years, including whether contraception was being used at the time the woman became pregnant. The relevant NSFG questions appear in Appendix C, along with additional material on the history and future plans of this survey.
On the basis of these questions, pregnancies in the NSFG are defined as
Among unintended pregnancies, a distinction is made between mistimed and unwanted:
The labels and terms from the NSFG have many limitations and ambiguities that should be mentioned. First, it is important to emphasize that the NSFG questions are designed to probe feelings at the time of conception, not at the time of birth. This distinction is important because a woman's feelings can change in many ways over the course of pregnancy (Miller, 1974). For example, Poole and colleagues (1994) report that among a sample of low-income women who were queried both early and late in their pregnancies, 12.5 percent reported a positive shift in attitude toward the pregnancy, and 10 percent reported a negative shift. In particular, an unintended pregnancy can result in a much