One major source of confusion in the literature dealing with teen pregnancy and childbearing is precisely the distinction between pregnancy and its outcomes. People often say they're referring to teenage pregnancy when they only have information on births. Pregnancy can be resolved in a number of ways, only one of which is a live birth kept by the mother. However, in talking about the problems of teen pregnancy, the problems that have been well-documented to date are those associated with that one outcome—bearing and raising a child as a teenager. Another set of confusions revolves around the process which leads ultimately to childbearing and its implications for policy and programs. For example, an agency may be interested in developing a profile of young women at risk of teen childbearing to target them for intervention. As discussed in earlier chapters, in order to become a teen mother, a young woman must first become sexually active, next, not use contraception or fail in its use in some way (including experiencing method failure), and, finally, once pregnant, decide to bear and raise the child herself. There are several points at which alternatives present themselves. Some teens choose one way, others choose another. Thus the agency has several possible points at which to target its interventions: at initiation of sexual activity, at contraceptive use, or, at the resolution of a pregnancy.
In this chapter some basic demographic description of the number and rates of teen pregnancies, births and abortions are first presented for the United States. Comparisons are drawn with Denmark, a country with registers of health events. Statistics showing the actual way pregnancies to U.S. teens are resolved are presented, followed by a discussion of research that sheds light on the factors associated with resolving a pregnancy one way rather than another. A summary and conclusions section closes the chapter.
In 1984 there were 469,682 births to teenagers 15 to 19, 9,965 births to teens under 15. This represents a considerable decline in births to teens over the decade, from a high of 656,000 in 1970. The number of pregnancies rose slightly until 1980 and has declined slightly since then. There were over a million pregnancies to teens in 1984 (Table 3.1).
However, the change in the absolute numbers of births and pregnancies does not adequately indicate the incidence of teen pregnancy and childbearing because it does not take into account changes in the number of teen women. The number of teens rose during the 1970s, leveling off in the mid 1970s and declining since 1979. Nor does it take into consideration the number of women at risk, that is, the number of women who are sexually active (see Hofferth et al., 1986). This is especially important for teenagers, only a portion of whom are sexually active. Pregnancy rates1 per 1000 women 15 to 19 rose 9 per- cent between 1974 and 1984; however, because the proportion who were sexually active also rose over the period, the pregnancy rates per 1000 sexually active women 15 to 19 actually fell 8.7 percent between 1974 and 1984.
What does this mean for individual women? The pregnancy rate in 1984 was 231 per 1000 sexually active women. This means that in 1984, 23 percent of sexually active teenagers would have become pregnant. This figure, however, only indicates the proportion of teens who would become pregnant in any one year. A more interesting figure is the proportion of young women who would ever become pregnant before reaching age 20. That is, what is the chance that a young woman would become pregnant as a teenager? Although this probability has been estimated using survey data, since abortions are underestimated in such data, the estimates of pregnancy will be low. Better estimates are obtained from reporting data such as those collected by the Centers for Disease Control and the Alan Guttmacher Institute. Based on such data it was estimated that in 1981 about 44 percent of young women will become pregnant before reaching age 20, 40 percent of white and 63 percent of black women (Forrest, 1986; Table 3.3).
Of course, this estimate, too, is rather crude, since among those young women are some who became sexually active very early in their teens, others who became sexually active very late and others who were still virgins at age 20. The data that are most helpful in showing what the actual risk of pregnancy is among those who are sexually active, breaks the probability down by the length of time since first intercourse and uses a life table methodology to estimate the risk of conception within the first two years after first intercourse (Zabin, 1979; Koenig and Zelnik, 1982). Data collected in 1976 (Zabin, 1979) indicate that within the first three months 9 percent of white and 14 percent of black teenage women will have experienced a first premarital pregnancy (Table 3.7). By the end of the first year that figure has risen to 17 percent for whites and 27 percent for blacks, and by the end of two years, 30 percent of whites and 37 percent of black teenagers will have experienced a first premarital pregnancy. Data from 1979 (Koenig and Zelnik, 1982) suggest a slight increase in the probability of pregnancy during the first two years after first intercourse between 1976 and 1979, with 33 percent of white teenagers and 43 percent of black teenagers experiencing a first premarital pregnancy within two years after first intercourse (Table 3.5). The probability of a first pregnancy is strongly affected by two factors—the age at first intercourse and the use of contraception (Tables 4.6 and 4.8). Pregnancy rates are much for those older at first intercourse and for those who always used a contraceptive method. There was little difference in pregnancy between those who used a prescription and non-prescription methods, as long as they always used it (Koenig and Zelnik, 1982).
THE RESOLUTION OF TEEN PREGNANCIES
What happens to these one million teenage pregnancies? Many more young women under 20 become pregnant than bear a child, almost twice as many. In 1982 the total births to teenagers 15 to 19 represented 47 percent of the total number of pregnancies (abortions plus births plus miscarriages [Table 3.1]).
Table 3.2 shows how pregnancies in 1982 were divided: 40 percent of the pregnancies were aborted, and 13 percent miscarried; thus slightly under half, 47 percent, resulted in a live birth. The 47 percent which were live births are divided as follows: 13 percent were postmaritally conceived births, 11 percent were premaritally conceived but born postmaritally, and 23 percent were born out-of-wedlock (estimates from Table 3.1 and O'Connell and Rogers, 1984).
The resolutions to a premarital pregnancy considered here are abortion versus having a live birth, marriage versus non-marriage, and adoption versus keeping the child.
Live Birth versus Abortion
The proportion of teenage pregnancies that ended in a live birth decreased over the past decade (Table 3.1). The number of teen pregnancies has risen, but because the number of abortions has risen even faster, the number of births has been declining. Both the number of abortions and the abortion rate increased by 50 percent between 1974 and 1980. The percent of teenage pregnancies terminated by abortions climbed rapidly, increasing from 27 percent to 40 percent between 1974 and 1980. Since 1980, the abortion rate and ratio have remained level. Birth rates for all women have remained fairly level; rates for those sexually active have declined. (Table 3.1).
Of course, it is difficult to interpret these figures without some comparison. What is a high level of pregnancy, of births, of abortions for teenagers? Unfortunately, there are only limited international data on abortions, especially by age of the woman. The United States has a high abortion rate for young women compared to western European countries (Jones et al., 1985; Henshaw and O'Reilly, 1983). The United States also leads in the percent of abortions to teenagers (Tietze, 1983; Bachu, 1983). In spite of the large number of abortions, births to United States teens are also high, relative to other countries (AGI, 1981; Jones et al., 1985).
Denmark is a good country with which to compare the United States. Levels of sexual activity among teenagers are actually higher in Denmark than in the U.S. (Rasmussen and David, 1981). Abortion laws were liberalized there about the same time as in the United States—the early 1970s. Most important, Denmark has an excellent abortion reporting system. With a unique identifying number for each person and a centralized information gathering system, the data on abortion in Denmark are among the most complete in any nation.
Pregnancy rates in the United States have been about twice the level of Denmark for the past decade (David et al., 1982; Table 3.1). In both countries the pregnancy rates increased initially after liberalization of abortion, but levels in Denmark returned to those prior to liberalization, while those in the United States continued to rise. As a result, rates of abortions and births in the U.S. in 1980 and 1981 are considerably higher than in 1970. Abortion rates in both countries rose. However, while they have leveled off in Denmark, they have continued to rise in the United States.
The rapid increase in pregnancy and abortion rates in the U.S. during the 1970s was due to the rapid increase in sexual activity over the same period. Apparently, levels of sexual activity rose dramatically in Denmark during the 1960s (Rasmussen and David, 1981); thus by the time abortion was legalized in both countries, sexual activity had begun to level off in Denmark at a higher level. In contrast, the major increase in sexual activity in the U.S. occurred during the 1970s, with a leveling off during the early 1980s (see discussion, Chapter 1). As Table 3.1 showed, pregnancy rates among those sexually active actually showed a decline between 1974 and 1984.
Two valuable lessons from these data and from a recent study of five western European nations (Jones et al., 1986) are that 1) high levels of sexual activity do not necessarily result in high pregnancy rates, given adequate use of contraception, and 2) low birth rates do not necessarily imply high abortion rates; they may simply imply low pregnancy rates. Low abortion rates and low birth rates are compatible.
Among teens, the proportion of pregnancies terminated by abortion is higher in Denmark than in the United States, primarily due to the high abortion ratio among 15 to 17 year old Danes (David et al., 1982). 15 to 17 year old United States teens are much more likely to bear their babies than Danish 15 to 17 year olds. Jones et al. (1985) also found that in each of 5 developed nations they investigated, that 15–17 year olds were much more likely to abort a pregnancy than 18–19 year olds: the difference was smallest in the U.S. This suggests substantial differences between United States and other countries in choice of resolution for unplanned pregnancies, differences which will be pursued a little later.
One way of resolving an out-of-wedlock teenage pregnancy is by marrying. So far all teenage pregnancies have been lumped together. In fact, some 13 percent of all teenage births are postmaritally conceived (Table 3.2), and such births are not generally considered to be problematic. In 1980 only 5 percent of abortions to teens 15 to 19 (about 2 percent of all pregnancies) were to married women (Henshaw et al., 1985). Assuming that abortions indicate that a pregnancy was unintended, it can be inferred that most pregnancies to married women are intended. Zelnik (1979) found that 53 percent of first births to women who were married were unintended. If to the proportion of postmarital births are added a proportion of the miscarriages and a small proportion of the abortions, it can be seen that that between 15 and 20 percent of all pregnancies to women under 20 occur to married women. The remainder, 80 to 85 percent, are premarital pregnancies.
Earlier, it was pointed out that about 24 percent of sexually active teenagers age 14 become pregnant each year. However, this does not tell us how many teenagers age 14 become pregnant before they reach 20 or marry. According to 1979 survey data (Zelnik and Kantner, 1980), 16 percent of all metro teenage women 15 to 19 had ever experienced a premarital pregnancy, double that of 1971. Of those sexually active, 33 percent had ever experienced a premarital pregnancy, a small increase since 1971. Thus, when control is introduced for the increase in sexual activity over the decade of the 1970's, the incidence of premarital pregnancy has not changed very much. The major reason for the large apparent increase in premarital pregnancy is the increase in sexual activity. There was an increase in premarital pregnancy among sexually active white teens, but not among black teens. The lack of increase among blacks is probably due to underreporting of abortion. Thus premarital pregnancy has increased, but not as much among those sexually active as it appears from the increase in the population of teenagers. Data from the 1982 National Survey of Family Growth show a slight decline in premarital pregnancy among teenagers between 1979 and 1982, although the difference is probably not statistically significant. In 1982 14 percent of all teen women 15 to 19 had ever experienced a premarital pregnancy, compared with 16 percent in 1979. Of those premaritally sexually active, 30 percent experienced a premarital pregnancy.
These figures substantially underestimate the true proportion of teenagers who become pregnant before they reach age 20 or marry because abortions are substantially underreported in surveys—by as much as 50 percent. Some subgroups report more accurately than other subgroups (Mosher, 1985). Unmarried black teenage females are the least likely to accurately report their abortions, with unmarried white teenage females only slightly more accurate. Older married white females are the most accurate reporters of their own abortions. Since accurate pregnancy estimates depend on accurate abortion reports, the reports of pregnancy obtained from surveys will be lower than those estimated on the basis of nationally collected data from organizations such as the Centers for Disease Control and the Alan Guttmacher Institute. Recent calculations from the latter (Forrest, 1986; Table 3.3) suggest that based on 1981 data about 40 percent of white teenagers 15–19 and 63 percent of black teenagers would experience a first pregnancy before reaching age 20.
The increase in premarital pregnancy over the decade of the 1970s was not due to an increased wantedness of pregnancy. Table 3.6 shows that the proportion of premaritally pregnant teens who were unmarried at resolution who wanted the pregnancy actually declined between 1971 and 1979 for whites and blacks alike, and the proportion using contraception increased (Zelnik and Kantner, 1980; Table 4.4). Of course, premaritally conceived but marital births, which constitute about 11 percent of teen pregnancies, are excluded here. However, since the proportion who marry to resolve a premarital pregnancy also declined, the proportion who wanted a pregnancy probably also declined for all premaritally pregnant teen women.
Contraceptive use generally improved between 1971 and 1982. A smaller proportion reported never using contraception, a higher proportion reported always using it. A larger proportion used contraception at first intercourse and at last intercourse in 1982 than in 1971. Unfortunately, Table 3.4 shows that the percentage of premaritally sexually active teen women who ever experienced a premarital first pregnancy rose in all contraceptive use statuses 1976–79, except for those who used contraception at first intercourse but not always (Zelnik and Kantner, 1980). The largest increase was among never users, but increases also occurred among those who always used contraception. The authors attribute this increase in pregnancy, particularly among the youngest teens, to sharply increased frequency of intercourse and to decreased reliance on the most effective methods of contraception (Koenig and Zelnik, 1982). Data are not yet available from the 1982 NSFG to see whether pregnancy rates continued to increase among contraceptive users as well as non-users. We suspect they have not, since pregnancy rates have been declining.
Data from three surveys of young women (Bachrach, 1985) show that the proportion of teenage women whose first pregnancy ended in a first premarital birth and who gave their baby up for adoption declined in the 1970s between 1971 and 1976 and leveled off at a low level between 1976 and 1982 (Table 8.1). Eighteen percent of white teenagers reported having terminated parental rights in 1971, 2 percent of blacks. By 1976 only 7.0 of whites and no blacks reported having given up a baby for adoption. By 1982 7.4 percent of whites and fewer than 1 percent of blacks reported having given up a child for adoption. Based on data from the National Survey of Family Growth, the estimated annual number of unrelated adoptions declined to a low in 1976 and has been gradually increasing since then.
Agency data support survey evidence which showed declining adoption placements from the early to the mid-1970s (Bachrach, 1985). Legal abortion became an alternative to adoption for many young women who had an unintended pregnancy and who would have adopted if abortion were not available. It has been argued that the reduced social stigma attached to unwed pregnancy caused a shift away from adoption as an alternative to childbirth. The subsequent apparent increase in adoption may be a response to the substantial demand for babies to adopt as well as a response to the many concerns about the ethics of abortion. This is just speculation, since there is no research that would allow us to shed light on these changes. Just documenting the changes that have occurred is a difficult task.
Factors Associated with Resolution of Premarital Teen Pregnancies: Delivering the Baby
Once a teenager is pregnant, what factors are associated with whether she has an abortion or carries the pregnancy to term and delivers the baby? One study found that the younger the teen at conception, the more likely she was to carry the pregnancy to term (Zelnik et al., 1981). In this study 13 to 16 year olds were more likely to have a live birth compared with 17 to 19 year olds comparable on other factors. This is supported by data from another study, which found that of those 13 to 19, the 16 to 17 year olds were most likely to have a live birth. However, national statistics on abortion ratios do not support these findings. The true explanation may be the underreporting of abortions in sample surveys of teenagers, which is likely to be most serious for the younger teens. An underreporting of abortions would increase the apparent proportion who carry pregnancies to term. Thus, due to underreporting of abortion, it is not clear whether factors are related to choice of abortion or birth or to whether abortion is reported. This is a serious problem for analytic study of abortion using sample surveys.
The birth year of the teenager is important. At a given age, earlier birth cohorts are more likely than more recent cohorts to have a live birth (Zelnik et al., 1981).
Young women are more likely than in the past to resolve a premarital pregnancy by abortion (Table 4.5). White teenagers were 1.3 times and black teenagers 2.5 times more likely to have an induced abortion in 1978 than in 1972. Although in the early 1970s black teenagers had a lower likelihood of using abortion to resolve pregnancy, according to these abortion ratios, after 1974 the abortion ratios are similar or slightly higher for blacks than whites. Since abortion data appear to be underreported more for blacks (Zelnik and Kantner, 1980), the difference in levels between blacks and whites may be underestimated. The abortion ratio appears to have levelled off after 1980, according to national figures (Table 3.1).
The black-white difference in likelihood of abortion varies by age. Among young teenagers the ratio of abortions to births is lower for blacks than whites (Table 4.6). However, this difference declines such that ratios are similar for 19 year olds. Among older women, ratios are higher for blacks than for whites.
One source of difference is the age at which abortions and pregnancies are measured. The Ezzard et al. (1982) study (Table 4.5) adjusted age to age at conception. This is particularly important at younger ages. Only a third of women who became pregnant before age 15 were still under 15 at delivery, while three-fourths of those obtaining abortion were still under 15 at the time of abortion (Henshaw et al., 1985). Thus differences between the figures will be sharpest at youngest ages.
Zelnik et al. (1981) found that the more religious a young women, the more likely she is, once pregnant, to bear the child. Another study using data from a small study of health providers in Ventura County California found white Catholics to be less likely to have a live birth, once pregnant than either white non-Catholics or Hispanic Catholics (Eisen et al., 1983). Thus the particular religious affiliation appears less important in the decision than the strength of religious conviction.
Teens living in the East or North central United States or in an urban area are more likely to have a live birth, once pregnant, than those in other regions or in non-urban areas (Zelnik et al., 1980).
The most important family factor associated with delivering a baby versus aborting a pregnancy is parental education. The higher the education of parents, the lower the likelihood that a teenager, once pregnant, will have a live birth (Zelnik et al., 1980). The mother's opinion of abortion is important, with girls whose mothers are more favorably disposed toward abortion less likely to have a live birth (Eisen et al., 1983).
Peer environment is important. The more positive a likely a young pregnant girl is to have a live birth (Eisen et al., 1983). In addition, girls who know a single teen mother are more likely to have a live birth (Eisen et al., 1983).
Among the most important factors affecting the outcome of the pregnancy was whether the pregnancy was wanted. Girls who said they wanted the pregnancy were much more likely to have a live birth than those who didn't (Zelnik et al., 1980). Of course, this measure of wantedness was obtained after the resolution of the pregnancy; ex-post facto rationalization may be measured here.
Beliefs about abortion and birth are important. Having favorable attitudes toward and beliefs about abortion prior to the event were associated with a lower probability of having a live birth (Eisen et al., 1983) and with a positive abortion intention (Smetana and Adler, 1979). Intention to have an abortion was associated with a lower probability of having a live birth (Smetana and Adler, 1979). Positive beliefs about having a child were associated with a low intention to have an abortion. Finally, women choosing either abortion or birth believed others wanted them to follow this alternative, with women intending abortion most motivated to comply with friends' expectations (Smetana and Adler, 1979).
Among the most important factors associated with choice of pregnancy resolution are expectations and academic achievement. High school dropouts and those not enrolled in school, those with a low grade point average, and those with low educational expectations have been found more likely, once pregnant, to have a live birth (Eisen et al., 1983; Leibowitz et al., 1980; Devaney and Hubley, 1981).
Two studies have looked at the relationship between receipt of AFDC and pregnancy resolution decision. Moore and Caldwell looked at the probability of abortion, marriage and out-of-wedlock birth among premaritally pregnant U.S. women aged 15 to 19 in 1971, data collected by Kantner and Zelnik in the National Survey of Young Women. Controlling for a number of individual characteristics, such as education of the father, wantedness of pregnancy, importance of religion and race, they found the probability of abortion to be significantly lower in states having relatively generous AFDC benefit levels (Moore and Caldwell, 1977).
Eisen et al. (1983) and Leibowitz et al. (1980) examined a group of 299 pregnant teenagers who went to health providers in Ventura County, California between 1972 and 1974 for assistance in terminating a pregnancy or for prenatal care. The teens were interviewed twice, once prior to abortion or delivery and a second time six months after the resolution of the pregnancy. The authors hypothesized that young women who received state support would be more likely to choose delivery than girls who were self-supporting. They found that both receiving financial aid from the family and receiving financial aid from the state (AFDC) were associated with choosing delivery (Eisen et al., 1983; Leibowitz et al., 1980). However, more young women than those currently living in welfare families would be eligible for welfare if they did give birth; thus the study really measures the effect of actual receipt of welfare benefits, rather than their availability.
Factors Associated with Marriage Before Birth (Legitimation)
Young women are less likely now than in the past to resolve a premarital pregnancy by marrying. The proportion of women pregnant before marriage who resolved a premarital pregnancy by marrying dropped by 50 percent between 1971 and 1979 for both whites and blacks (Zelnik and Kantner, 1980). The data show very little additional change between 1979 and 1982, although the data are not completely comparable, and the total number of pregnancies is underreported (Horn, 1985).
If we look only at pregnancies that end in a live birth, we see that of the total first births to white and black teenagers, the proportion conceived outside of marriage has risen, and the proportion premaritally conceived but legitimated before birth rose then declined to about the same initial level (O'Connell and Rogers, 1985). As a result, the proportion born out of wedlock rose sharply.
Two studies have examined factors associated with whether a premaritally pregnant teenager who subsequently had a birth married prior to that birth: Zelnik et al. (1981) used data from the National Survey of Young Women in 1971 and 1976. They found that (among those who were premaritally pregnant and gave birth) white teenagers, those from a higher socioeconomic status background and those who wanted the baby were more likely to marry before bearing the child. The second study used the data from Ventura County, California (Eisen et al., 1983). They found that (among those who carried to term) the only factor that discriminated between those who married before the birth and those who didn't was whether the family had been receiving financial aid from the state. Those girls whose families had been receiving financial aid from the state during pregnancy were less likely to marry than those who had not been receiving such assistance (Eisen et al., 1983).
Factors Associated with Bearing an Out-of-Wedlock Child
The resolution many people are interested in is that of bearing a child out-of-wedlock compared with all other options. The previous analyses have explored the decisions in temporal sequence: that is, they have looked at, first, the decision to abort or carry a premarital pregnancy to term, and, second, the decision to marry or not marry before birth among those who carry to term. Several analyses have studied this decision as a joint one with three choices: 1) abortion, 2) marriage and birth, and 3) bearing an out-of-wedlock child. The results of studies viewing the decision this way do not differ from the results of studies using paired comparisons only, but this approach allows simultaneous comparison among all alternative resolutions. Young women who are black, who live in a metropolitan area, whose parents are of low educational levels, who are young at first conception, and who live in a large family are more likely to bear a child out-of-wedlock than to either abort or marry (Eisen et al., 1983; Leibowitz et al., 1980; Devaney and Hubley, 1981; Zelnik et al., 1980). In addition, Leibowitz et al., 1980 and Eisen et al., 1983 found teens living in families receiving financial aid from the state to be more likely than their peers to bear an out-of-wedlock child. In contrast, using 1971 data from the National Survey of Young Women, Moore and Caldwell (1977) found no relationship between level of AFDC benefits and having an out-of-wedlock birth. The latter found a negative relationship between AFDC acceptance rates and the probability of having an out-of-wedlock birth. That is, young women in states with high acceptance rates were less likely to have an out-of-wedlock birth (Moore and Caldwell, 1977). There was no significant association between AFDC benefit levels and acceptance rates and the probability of marrying before the birth (Moore and Caldwell, 1977).
Only a few studies have compared teens who have made adoption plans with teens who have kept and parented their children. These are summarized in Resnick (1984). The results suggest that teenagers who make adoption plans are similar to those who have abortions but different from those who take on parenting responsibilities. The former tend to be older, to have more parental influence and less male partner influence, and to be of higher socioeconomic status. Parenting teens tend to be younger, to have less schooling, to not be attending school and to come from non-intact homes. Thus those who make adoption plans tend to have more prospects for the future. In addition, they were reared in smaller towns and cities and have more traditional attitudes about abortion and family life (Resnick, 1984).
Recent data from the 1982 National Survey of Family Growth (Bachrach, 1985) show that teenagers under 18, whose parents have had some college, whose baby was born before 1973, and who were living with both parents at age 14 were more likely than other teenagers to place the child for adoption if they had a premarital birth.
Two recent studies (Kallen, 1984; Resnick, 1984) are funded by the Office of Adolescent Pregnancy Programs to look more closely at the factors affecting the decision of unmarried pregnant teens to make an adoption plan. At this writing no results are yet available.
Factors Associated with Decision Satisfaction
It is obvious that no one decision is the “right” decision for all adolescents, since the circumstances differ among individuals. However, researchers have found some regularities in the extent to which individuals express satisfaction or dissatisfaction about the decisions they have made in resolving their pregnancies. A study of a Danish sample found that the degree of satisfaction with the decision depended on the firmness of the decision in the first place. Of those who had made a firm decision to abort soon after learning about pregnancy, 94 percent said that the decision was correct 6 months later. Of those who were not so certain, 72 percent said that the decision was correct 6 months later (David et al., 1982). Of those whose decision was firm, 59 percent experienced relief afterward, compared with 28 percent of the less firm. None of the Danish women expressed feeling of guilt over the decision.
A study of United States teen women (Rosen, 1983) found that the more alternatives considered, the greater the dissatisfaction with the decision. This probably reflects greater uncertainty as to what to do, and is consistent results from the Danish study (David et al., 1982).
The Eisen and Zellman (1984) study of pregnant teens in Ventura County, California found no significant difference in decision satisfaction 6 months after pregnancy resolution by type of decision made, age or ethnic group. Nearly all—80 percent—expressed satisfaction in their decision. There were some differences in degree of satisfaction depending on the decision made. Among teenagers who chose abortion, those with better educated mothers, who had advocated abortion for themselves, who were more approving of abortion in general and who used contraception more consistently following abortion were more satisfied (Eisen and Zellman, 1984). Among teens who chose single motherhood, those not enrolled in formal schooling during the six months after birth were more likely to be satisfied with the decision, as were those with maternal support for single parent status. Among teens who married, none of the variables utilized significantly differentiated those who were satisfied with their decision from those who weren't.
Interesting and Controversial Issues
Three issues are worth looking at further. The first issue is the relationship between age and pregnancy resolution. Young teenagers in the United States have a very high probability of bearing the child, once pregnant, compared to older teenagers or teenagers in other countries (Jones et al., 1985). Data from the Danish study (David et al., 1982) show abortion ratios (abortions divided by births plus abortions) to 15 to 17 year olds that are twice those of U.S. 15–17 year olds. Three-quarter of the pregnancies to young Danish teens are terminated by abortion, compared with 40 percent of those to young U.S. teens. Abortion ratios for 18 to 19 year olds are very similar in the U.S. and in Denmark. Results from the National Survey of Young women suggested that, net of other factors, girls younger at conception are more likely than older teens to carry a pregnancy to term. Although the differences are exaggerated because of the underreporting of abortion at younger ages, it could be expected that abortion would be higher at younger ages than at older ages, as shown by the Danish sample, since few young women wanted these pregnancies.
Thus the lack of difference by age in the United States is of interest. Why are 15 to 17 year old pregnant teens in the United States so much more likely to bear a child than comparable teens in a country such as Denmark and other countries? Why are they as likely to bear a child as their 18 to 19 year old peers in the U.S.?
The second important issue is that of race differences in pregnancy resolution. The chapter has emphasized differences between blacks and whites, but conclusions about race differences in pregnancy resolution based on analyses of survey data are of necessity weak because of differential reporting of abortion by race in those data sets. The best information on subgroup characteristics come from the Centers for Disease Control, AGI, and from the National Center for Health Statistics and they are good. However, such data do not provide the depth of information needed to explore causal factors in decision-making. Another problem is whether to use abortion rates or ratios. The abortion ratio is higher among blacks than whites for all ages except the teen years (Table 4.6). During the teen years, the ratio of induced terminations of pregnancy to live births is higher for whites than for blacks. However, if you look at the abortion rate (Table 4.4) the rate is higher for non-whites than for whites at all ages. This is because the pregnancy rate for non-whites is also higher. Thus, in this case, using the abortion rate would lead to a completely different and erroneous conclusion about black-white differences. Analysts need to choose the appropriate measure for their purposes.
One reason for the differences between blacks and whites in abortion is that blacks appear to use abortion for spacing or to end childbearing more than to postpone a first birth. Sixty-five percent of abortions to whites occurred to childless women, compared to 39 percent of abortions to blacks (Table 4.7).
However, there is another problem with the data. Figures are often based on age of the woman at pregnancy outcome. Since birth occurs nine months and abortion approximately 3 months after a conception, a proportion of the young women who conceived (and who eventually bore a child) at the same time as those who conceived and who eventually terminated the pregnancy through abortion would be one year older at outcome. Thus the event (pregnancy) occurred at the same age, but this would not be reflected in the statistics. Adjusting the data to age at conception would take care of this problem, but would also alter the number of births and abortions, especially at younger ages. Thus the Ezzard et al. (1982) study (Table 4.5) shows almost no black-white difference in abortion ratios when abortions and births are adjusted to age at conception. This raises an important issue of comparability of measures across studies. The Alan Guttmacher Institute has moved toward reporting ratios adjusted to age at conception. The other organizations that report abortion statistics do not yet do so (the Centers for Disease Control and the National Center for Health Statistics).
A third interesting issue is that of repeat abortion. In 1980 one-third of U.S. aborters had previously had an abortion (Tietze, 1978; Henshaw and O'Reilly, 1983: Table 7). The figure is smaller for teenagers, as could be expected, since they have not had as much time to have one, let alone two abortions. NCHS data suggest that 12 percent of abortions to 15 to 17 year olds, and 22 percent of abortions to 18 to 19 year olds are repeat abortions (Table 4.7). There are two potential reasons for concern. First, there may be negative effects of abortion on later childbearing and subsequent pregnancies. Second, there may be (over)utilization of abortion as substitute for contraception.
Are there negative effects of abortion on later childbearing and subsequent pregnancies? This literature has been reviewed in Strobino (in this volume) and Hogue (1982); the reader is referred to those sources. After adjusting for the fact that abortions performed on teenagers are performed later in pregnancy, which is somewhat more risky, rates of mortality and morbidity from abortion are somewhat lower for teenagers than for adult women. There is only one instance in which teenagers appeared to be at higher risk of injury than adults. Teenagers appeared to be at higher risk of cervical damage than older women (Cates et al., 1983; Cates, 1981).
Although there is little evidence that having had one prior abortion increases a woman's risk of miscarriage, premature birth or bearing a low birth weight baby, there is some evidence that having had multiple abortions may increase this risk, although, again, the results of several different studies do not agree (Levin et al., 1980; Chung et al., 1982).
Is abortion over-utilized as a substitute for contraception? The concern that abortion is becoming a substitute for contraception does not seem founded. Although in 1971 the percentage of teen women who had a premarital second pregnancy was higher 2 years after the outcome of the first premarital pregnancy for those who had an abortion than for those who had a birth, by 1979 the figures were reversed. In 1979 teen women who had terminated their premarital first pregnancy by abortion were less likely to have a second pregnancy within two years than those who had carried the first pregnancy to term (Koenig and Zelnik, 1982). Tietze (1978) argued that the increasing number of repeat abortions reflects the increasing number of women who have had a first abortion and are, therefore, at risk of having a second abortion. This appears to be born out by a recent study that shows few differences between women obtaining a first and those obtaining a repeat abortion (Berger et al., 1984). Those obtaining a repeat abortion were older, less likely to be married and more tolerant of legal abortion than were women having a first abortion. They had intercourse more frequently and they were more likely to have been contracepting when they became pregnant. They did not differ on type of method used or on any other demographic, psychological or attitudinal measures. Finally, results from a 1982 national survey show that fewer than one half of 1 percent of women exposed to the risk of unintended pregnancy, who did not use contraception, mentioned the availlability of abortion as a reason for nonuse (Forrest and Henshaw, 1983).
SUMMARY AND CONCLUSIONS
How women choose to resolve their pregnancies has become one of the major factors determining the number and rate of births to teens. Only about half of all pregnancies to teens end in a live birth. Yet only a very small amount of research has been conducted on this important issue. One important issue that researchers have just begun to address is whether miscarriage and abortion have psychological, social, health, familial, educational, economic or other consequences for adolescents and for their families. A few studies have focused on short term psychological effects, but there are no long term studies. The many studies of health effects that have been conducted have found little negative impact on health (Hogue et al., 1982).
One major question that several researchers have addressed is why individual women choose one form of resolution to a pregnancy over another. The major studies in this area use two data sets: the National Surveys of Young Women (1971, 76, 79) and a study of 299 women in Ventura County, California in 1972–74. These are the only studies to provide multivariate evidence on the issue, and they are the only studies to have focused on the resolution of premarital teen pregnancies (as distinguished from postmarital teen pregnancies). It is important to make this distinction. Few people consider maritally conceived pregnancies problematic, although, among young teenagers, they may be. Research suggests that a premaritally pregnant teen is more likely to give birth rather than obtain an abortion if she wanted the pregnancy, is of lower socioeconomic status, is unfavorably disposed to abortion, has lower aspirations and educational expectations, receives parental financial assistance, currently lives in a family that receives public assistance, and lives in a state with higher AFDC benefit levels. These results are based on a very limited set of studies, however, and all these studies suffer from underreporting of abortion.
Among those who give birth, those who are of lower socioeconomic status, who are younger, and who are black are less likely to marry than their peers.
Two types of data are needed: 1) Vital statistics data that can provide national estimates of abortion (and, as a result, pregnancies) by age and, simultaneously, by race/ethnicity, and 2) Survey data that not only provide reasonable estimates of abortion but also contain variables that could be used to test hypotheses about relationships among variables both at one point and over time. At the present time there are no national reporting requirements for abortions. Abortion data are presently estimated from three sources: a national survey of providers by the Alan Guttmacher Institute, counts of characteristics of abortion patients obtained by the Centers for Disease Control and counts of abortions obtained in 12–13 reporting states by the National Center for Health Statistics. National estimates of abortions in survey data can be obtained from the National Survey of Young Women (1971, 1976) and the National Survey of Young Women and Young Men (1979), the National Survey of Family Growth, Cycle III (1982), and the National Longitudinal Survey of Youth, Ohio State University (1979–1985). Unfortunately, all these surveys have documented substantial underreporting of abortions, so they should be used cautiously until we have a better understanding of the bias this introduces into our analyses.
Pregnancies=Births and abortions plus miscarriages. Accurate abortion data are needed to calculate the number of pregnancies. Abortion was legalized in the U.S. in 1973. Prior to this year, the annual number of abortions in the U.S. could only be estimated. Therefore, 1974 was selected as a comparison year since it is probably the first full year with good abortion statistics.
individuals clearly need increased attention and services, reducing unintended pregnancy will require that influential organizations and their leaders—corporate officers, legislators, media owners, and others of similar stature—address this problem as well.
As noted above, the committee calls for a campaign that is both multifaceted and long-term, emphases that derive from the data presented in Chapters 4 through 7 showing that no single factor accounts for unintended pregnancy and that the underlying issues are very complex. In truth, there are many antecedents to the problem: socioeconomic, cultural, educational, organizational, and individual. Therefore, only a comprehensive effort will succeed in reducing unintended pregnancy, as has been the case for other national campaigns, such as those to reduce smoking, limit drunk driving, and increase the use of seat belts. Unintended pregnancy will not be reduced appreciably, the committee believes, unless more individuals and institutions make a major commitment to resolving this problem. Similarly, the campaign must be long-term. Past experience teaches that brief, intermittent efforts to address important social and public health challenges have very limited success.
The U.S. Department of Health and Human Services, through its National Health Promotion and Disease Prevention Objectives, has urged that the proportion of all pregnancies that are unintended be reduced to 30 percent by the year 2000 (U.S. Department of Health and Human Services, 1990). The committee endorses this goal, and stresses that it is a realistic one, already reached by other industrialized democracies. Achieving this goal would mean, in absolute numbers, that each year there would be more than 200,000 fewer births that were unwanted at the time of conception and about 800,000 fewer abortions annually as well.
The Campaign to Reduce Unintended Pregnancy
What should the campaign emphasize? Should it stress contraceptive services? School-based information? Abstinence? parent-child or male-female communication about contraception? Community norms regarding reproductive behavior? The specific skills required to use reversible methods? Or, to put these questions in a slightly different way, which factors best predict unintended pregnancy and should therefore be the main targets of action?
The information presented throughout this report, past experience in the public health sector with complex health and social issues, and common sense itself are all helpful in sorting through various options. The committee proposes a portfolio of activities to prevent unintended pregnancy that, like many public health campaigns, emphasizes basic information and preventive services accompanied by comprehensive program evaluation and research. It also addresses the important domain of personal feelings and relationships