Early Predictors of Nonmarital First Pregnancy and Abortion
Family planning perspectives
acteristics and their psychosocial traits during childhood and adolescence. Previous Studies Only two studies we found have used characteristics of respondents that predate the first pregnancy to predict the likelihood of abortion. In the first, Costa, Jessor and Donovan followed a sample of high school and college students from 1972 or 1973 to 1981. 3 The researchers gathered data on the women's psychosocial and family characteristics at the early date and asked at the last follow-up whether
... follow-up whether they had ever had an abortion; the data did not allow the investigators to classify pregnancies by women's marital status or to identify women who had never been pregnant. When Costa and colleagues compared the characteristics of women who had had an abortion with those of women who had not, they found that psychosocial unconventionality (as defined by such characteristics as having large numbers of sex partners and having liberal attitudes regarding sex roles) was the primary predictor of who would have an abortion. R esearch describing the demographic and socioeconomic characteristics of women at the time they have an abortion indicates that in the United States, unmarried women account for the bulk of these procedures. In 1987, 9% of pregnancies among married women ended in abortion, compared with 57% of those among unmarried women. 1 In 1990, 82% of abortions were performed on unmarried women. 2 Women who have a nonmarital pregnancy are at the highest risk of having an abortion. Thus, to predict what group of women are the most likely to have an abortion, we must first predict what group are the likeliest to have a nonmarital pregnancy. Ideally, we must start with a group of women for whom we have data from the period before any married and before any became pregnant. Such a study directs our interest to the women's childhood and adolescent characteristics and requires a life-course approach. In this article, we seek to answer two questions: First, what factors predict a nonmarital first pregnancy? Second, among women who have such a pregnancy, what factors predict abortion as the resolution? We focus on two sets of attributes of the women: their family char-among women who became pregnant outside marriage between 1973 and 1985. 4 She found that living in an intact, two-parent family during early adolescence and having parents who had at least a high school diploma decreased the probability of a nonmarital pregnancy and increased the probability of an abortion in the event of a nonmarital pregnancy. The analysis also revealed that the more siblings a woman had, the more likely she was to have a nonmarital pregnancy, and the less likely she was to end such a pregnancy by abortion. Beyond these two studies, existing theory provides little in the way of guidance concerning the effects of early attributes on the likelihood of abortion. In this article, we follow the lead of Costa and colleagues, who theorized that abortion was unconventional behavior, and of Cooksey, who found that family characteristics directly predicted who among women with a nonmarital pregnancy would have an abortion. On the other hand, in an earlier analysis based on the same data set used in this article, we showed that preadolescent psychosocial characteristics were stronger predictors of early initiation of intercourse than were family characteristics. 5 For the present analysis, we therefore hypothesized that psychosocial characteristics would better predict who had a nonmarital first pregnancy, and that family characteristics would better predict which women who had a nonmarital first pregnancy would have an abortion. Study Design The Sample Our sample consisted of women born between March 1960 and March 1963 to white women taking part in the Child Health and Development Study (CHDS). (The number of women of other races in the CHDS sample was too small for separate analysis.) The CHDS followed up mothers and daughters once when the daughters were aged five, once when they were aged 9-11 and once at age 15-17; the survey data, including medical records and follow-up interviews, are available as public-use data sets. 6 In 1990-1991, we conducted another follow-up of the daughters (who were aged 27-30), in Most research on abortion has focused on women's characteristics at the time of the procedure, but individuals' behavior may also be shaped by their experiences from younger ages. This study uses longitudinal data on 351 California white women aged 27-30 in 1990-1991 to identify characteristics in childhood and adolescence that predict who will have a nonmarital first pregnancy and, of those who do, which women will seek an abortion. Bivariate analyses reveal that psychosocial characteristics indicating a strong sense of autonomy, such as feeling it is important not to be tied down and engaging in socially undesirable behavior, are significantly associated with the likelihood of having a nonmarital first pregnancy (odds ratios of 1.7 and 1.5, respectively), but family characteristics are not. However, among women who have a first pregnancy out of wedlock, the odds of having an abortion are mostly influenced by family rather than psychological characteristics, particularly having been a good student and having a well-educated mother (2.0 and 1.7).