Social Dynamics of Adolescent Fertility in Sub-Saharan Africa (1993)

Any contemporary study of adolescent fertility inevitably confronts the topic of education. Throughout the world, formal schooling for women is the single most consistent correlate of their low fertility. ¹ Yet within this general agreement is a question of perpetual contention: How, precisely, does education work this reproductive magic? Does it teach a woman Western scientific facts about reproduction and health, instruct her in the national language in which radio messages about contraception are broadcast, expose her to ideals of low fertility, extricate her from the authority of kin who demand high fertility, imbue her with career aspirations outside the home, embolden her to ask for contraceptives from intimidating family planning personnel or in the face of an irate husband? (For important reviews and statements, see Schultz, 1973; Leibenstein, 1974; Easterlin, 1975; Cochrane, 1979; Graff, 1979; Caldwell, 1982; Easterlin and Crimmins, 1985; Oppong and Abu, 1987; Cleland and Rodriguez, 1988.)

Though each of these mechanisms undoubtedly plays some role, con-

vincing evidence to evaluate their various effects is virtually nonexistent. For Africa, these questions have undergone hardly any systematic scrutiny. In any event, as Graff (1979) cautions, it is easy to overestimate the virtues of Western education:

Instead, the contributions of education and literacy

Following Graff's lead, we attempt to address the question of education as concretely as possible in relation to fertility among adolescents in the African context.

This chapter and the next cover issues related to education (or, more broadly, training) and fertility. The first issue, covered in this chapter, concerns changes in formal education and why they are considered so important in the fertility equation.

Education is usually assumed to affect fertility, rather than vice versa, because women usually acquire education in their youth and bear children as adults, after their education has ended. By the logic of temporal order, education is assumed to have imprinted women in ways that make them substantially alter their lifetime reproductive behavior. Though this assumption itself is problematic, our more limited focus is not mature women, but adolescent women whose reproductive careers have hardly begun. The question of causation is further complicated by the fact that among adolescents, several critical events may be happening concurrently or in rapid succession. A schoolgirl may have struck up a sexual relationship, her elders may have begun negotiations for her marriage, and her work plans may still be vague. Sorting out a sequence of events—the end of education, the beginning of reproduction—in such a complicated situation may bring us no closer to establishing cause: Individuals often make decisions less in response to past events than in anticipation of future ones. Still, by singling out adolescents for consideration, we raise the possibility that the opposite causal direction may apply: Those young women who manage to avoid childbearing may gain the opportunity to obtain advanced schooling.

This chapter also examines whether education should be used to explain delays in adolescent fertility. We examine a variety of evidence that suggests that other factors may also be important in delaying entry into childbearing. The next chapter pursues this point as we consider other forms of training for adult life that may be significant in postponing childbearing.

Although many African countries have low rates of enrollment in formal education, most people, both literate and illiterate, place enormous weight on the importance of schooling. We present a brief case study, that of Sierra Leone (Bledsoe, 1992), that gives a flavor of what can best be described as the culture of education in many African countries. To be sure, primary school enrollment rates in Sierra Leone are not unusually high, compared with those in other countries: The gross primary enrollment rate was 53 percent in 1988 (World Bank, 1988). Yet the country's modest enrollments only make it all the more compelling as a focus of study.

Sierra Leone has been called the "Athens of West Africa," because high-quality formal schooling was established there quite early. Reacting to the horrors of slavery, an Englishman named Granville Sharp laid out a plan in 1786 for helping freed slaves and their descendants in England by creating a blueprint for a perfect society in West Africa, in which no one's labor could be coerced. Inspired by this Utopian vision, an initial group of settlers seeking to establish such a society set sail in 1787 for Sierra Leone. The cornerstone for their experiment was education, and setting up schools became an immediate priority.

These historical events laid a powerful template for what can best be seen as the contemporary culture of education. Many young people today yearn to become educated in order to leave rural villages where farm work demands exhausting labor. They want to be literate; to wear well-tailored Western-style clothes; to speak English and Krio, the national linguae francae; and to acquire civil service or business jobs from which they can earn regular monthly wages, instead of relying on irregular seasonal harvests. The obstacles to realizing these ambitions, however, are myriad.

Students must compete aggressively for admission, especially to secondary schools. Even for those who obtain places, tuition and school expenses require diverting family income from medical treatment, farm loans, and food purchases. Among those who are less fortunate, boys may be sent to learn skilled occupations with trade masters or to learn Arabic with an Islamic scholar; girls may be kept at home or sent to other families to be trained in domestic or trading skills. Most rural girls marry by age 17.

In much of the country, a distinctive feature of a town of any size is the presence of local chapters of the region-wide secret Poro and Sande societies. For at least three or four centuries these societies have initiated local boys and girls, respectively, into their membership, and have claimed to control powerful knowledge from which low-ranked members and the uninitiated are excluded. The initiation period is considered a time of singular danger. If a female initiate must leave the enclosed "bush" temporarily, she

dons special ritual clothing and paints white chalk on her body to indicate that she is not at liberty to interact freely, especially with men, during her perilous liminal state.

Though the Poro and the Sande continue to initiate practically all local children, a town of any size also boasts a contemporary sign of the times: a primary school. Local people comment that the Western outsiders whom they encounter lead an easy life largely because of their knowledge of the outside world and its wondrous technology. Those who desire a higher standard of living strive to send their children to school to acquire this special knowledge. Such schools are quite humble physically, often merely an open-walled, mud-brick structure with a corrugated iron roof. Sitting tightly packed, four or five to a bench, children take up the day's lessons. Because of the cost of paper, pens, and chalk (several times the actual cost of what it is in the United States), students in rural schools may learn the alphabet by tracing letters in the air or outside on the sand. But the eagerness to learn is so palpable that most children persist in their studies even in the face of extreme hardship.

Larger towns may have several primary schools and perhaps a secondary school. But because the quality of education is considered crucial to a child's chances of success, many children living near schools with poor facilities or mediocre academic reputations leave to attend better schools. They live with relatives, friends, or patrons, and spend their free hours in the service of the household: caring for young children, cooking, cleaning, or hawking trade items.

The schools most children would like to attend, whether private or public or run by Christian or Islamic missions, are in cities, where the education industry thrives. City newspapers advertise "institutes" that tutor students for college entrance exams and companies that claim access to overseas scholarships. Islamic ritual specialists, many of whom have never attended Western schools, do a thriving business among secondary school students, selling potions that are proclaimed to seal into memory the welter of facts presented during class.

In better years in the past, most urban schools had ample supplies of books, paper, and chalk. Science labs were well equipped, and highly trained teachers were paid regularly. But even the memories of such halcyon days are dim. Since the early 1970s, education has been hard hit by economic problems, and opportunities for succeeding through education have been sharply curtailed. Even students who manage to finish secondary school, college, or professional school cannot be assured of finding good jobs. The number of white collar jobs in the national business and civil service bureaucracy has declined, and the qualifications required to compete for them have risen. Recession undermines the quality of even the wealthiest urban schools.

In the face of these economic woes, some evidence suggests that families' inclinations to send their children to school are ebbing (International Labour Organisation, 1989). Families complain about the cost of schooling and express growing doubts about their ability to afford it. Rarely does one encounter parents who vow to keep all their children in school. Instead, there is more talk of which child to invest in. Despite these mounting hardships, children's own aspirations to go to school have not abated. Indeed, what is perceived as a growing scarcity of white collar jobs, and an inflation of job requirements has only intensified the national culture of education.

Desires for schooling have produced a new seasonality; now, the annual cycle is as much affected by exam schedules and school holidays as by agricultural cycles. In September, the ferryboat crossing the harbor from Freetown to Lungi International Airport teems with older students from wealthy families heading off to boarding schools in Europe or the United States. Up-country, stores stock up on pencils, exercise books, and textbooks, and small tailor shops keep their apprentices working late to fill orders for uniforms. Throughout the day, children traveling from small villages with no schools to towns with schools climb out of dusty trucks. Clutching small valises or plastic bags filled with a few clothes, they make their way to the houses of the families who have agreed to take them in.

As the school term comes to an end and exam time draws near, students all over town hover over study papers on verandas or under trees after school. Those from wealthy families write on paper with ballpoint pens to practice for exams; others write on the ground or abscond with school chalk to write on large metal water bins behind houses. Studying goes on into the night. Dim kerosene lamps shine through the cracks of window shutters as tired students worry into the night.

After exams are over and the results announced, lorries and ferries fill again with children returning home with their all-important school report cards. During the holiday, many boys in secondary schools travel around the country, seeking money from relatives for school fees and expenses, making sure, if they did well, to bring their report cards with them.

Desires for schooling produce an extraordinary degree of child mobility even on a weekly basis. Walking through even the smallest village with no school can be a quite different experience, from one day to the next. Because many children are fostered out to attend schools in large towns nearby, a visitor during the week is likely to get the impression of a quiet village of adults and a few young children. On the weekend, the missing segments of the population pyramid suddenly appear: Chattering schoolchildren come home to help their families with farm work. Chairs and porches are draped with drying uniforms in the official colors of schools in distant towns. Whereas only local languages may be spoken in a village during the week,

on the weekend young voices show off new language skills by calling out teasing greetings to Western visitors in English.

For girls, education significantly changes their life prospects. Before education was available, a girl's training was directed toward her future roles as wife, farmer, and caretaker of children, and marriage confined her geographically and economically. Today, rural girls increasingly seek urban careers. Some go simply to be ''trained'' in domestic skills in urban households; others are trained in marketing. But many girls go to primary school, and a few go on to secondary school.

Almost without exception, girls entering secondary school have ambitions to go on to college or professional schools. But, although the number of girls and boys beginning primary school is now roughly equal, the proportion of girls declines rapidly as puberty approaches. Rural families support a girl's goals for advanced education with mild enthusiasm at best. Her labor is needed in the household, and her early marriage to a wealthy man may underwrite the education of her younger brothers. But because a daughter who remains in school becomes more desirable as a wife for an urban man, her family elders may risk leaving her there against the chance that she will become pregnant and be forced to drop out.

Schools are widely charged with keeping their students uncontaminated both by knowledge of, and experience in, reproduction while they prepare for adult life. Injunctions to avoid pregnancy permeate all aspects of school life, even down to the symbolism of a school uniform. Children who wear school uniforms are regarded with a mixture of respect and fear because of their potentials for achievement in the "civilized" world, with its esoteric mysteries and fearful powers. A girl's school uniform suggests that she is being prepared for marriage to a man of importance and as such should be treated with respect. More important, the symbolism attached to school uniforms, like attire worn by Sande society initiates, marks her as sexually unavailable. A teacher drew out these comparisons, implying that being pregnant and wearing a uniform are symbolically incompatible. For a schoolgirl who becomes pregnant,

Public resentment at a schoolgirl becoming pregnant is conveyed in the negative image of an unkempt school uniform. An older girl with a dirty,

torn uniform is suspected of having no one to watch over her properly, rendering her vulnerable to men who might tempt her with food and money.

The notion that reproducing is incompatible with training (Agounké et al., 1990) helps to explain the intense disapproval of adolescent mothers continuing in school. Affirmed a teacher in rural Sierra Leone, "[t]he schools in Sierra Leone do not generally admit girls who have given birth: mothers. She is not considered a schoolgirl again [any more]" (quoted in Bledsoe, 1990:293–294). In The Gambia, the headmaster of a rural secondary school regularly admonished his female students about their responsibilities to avoid sexual activity. His policy was "no visiting 'friends' [lovers], no marriage, no pills while learning" (Bledsoe, unpublished field notes, 1991). Any girl suspected of being pregnant was sent to the local prenatal health center to be checked. If pregnancy was discovered, she was expelled from school. According to a young female graduate of this school, the headmaster's philosophy was, ''If you just follow the boys, it means you don't want to learn." Schools have much to lose, she explained, by being too lenient. If the students begin to get pregnant it will become known as a ''prostitute school." Families, particularly of good students, will withdraw their children, and the school's scores on national exams will plummet, along with its academic reputation.

How do these detailed observations about schooling in one country play out against the wider backdrop of facts about education of women across the continent? Obvious parallels exist in countries where secondary education for women has increased. The situation is likely to be quite different, however, in countries where few girls attend school.

School enrollments grew rapidly in the boom period from the 1960s to the 1980s (World Bank, 1988). Most African countries reported substantial increases in enrollment rates for primary school (the percentage of the relevant age group enrolled in primary school). As a result of this growth, those rates now exceed 90 percent in 13 of the 35 countries listed in Table 5-1. But despite these accomplishments, the average enrollment rate in sub-Saharan Africa remains lower than that on any other continent. In Burkina Faso, Mali, Niger, and Somalia, fewer than 30 percent of those in the relevant age group were enrolled in primary school in the early 1980s.

Table 5-1 also shows that secondary school enrollment rates are substantial in some countries but remain much lower than those for primary schools (as the differing definitions of low, medium, and so on, suggest). By the early 1980s, secondary school enrollment rates exceeded 50 percent in the Congo and Zaire. In 10 countries, however, fewer than 10 percent of the relevant age group were enrolled in secondary school in the early 1980s.

Secondary school enrollment grew faster from 1960 to 1980 than did primary school enrollment: 13.8 versus 6.6 percent (World Bank, 1988).

Enrollment rates of girls are lower, in general, than those of boys. In 1970, 32 percent of girls and 63 percent of boys aged 6–12 were in primary school in Africa (World Bank, 1988); by 1980 the percentages had increased to 58 percent and 87 percent, respectively. Although girls account for 40 to 50 percent of primary enrollees in most countries, in Benin and Chad no more than a third of the primary school students are girls (see Table 5-2).

Sex imbalances in enrollment rates intensify at the secondary and university levels. In 1970, 2 percent of women and 6 percent of men aged 18–23 were enrolled in secondary schools or universities. The corresponding percentages for 1980 were 5 percent and 11 percent (Gyepi-Garbrah, 1985a). Table 5-2 shows that by 1983, the proportion of girls among secondary school students exceeded 40 percent in only 8 of the 35 countries listed. The southern region stands out as having relatively high levels of girls in secondary schools. High levels of male labor migration may be part of the explanation.

Despite overall increases in secondary education, the World Bank (1988) reports, increases in primary school enrollments slowed between 1980 and 1983, and in several countries (Togo, Somalia, Tanzania, Liberia, and Nigeria) total enrollment declined. Between 1960 and 1980, increases in secondary school enrollment rates also tapered off in several countries; and in a few countries, such as Angola, Liberia, and Togo, the numbers enrolled in secondary schools actually declined. Economic crisis and the implementation of structural adjustment programs are likely central factors. In several countries the percentage of gross national product (GNP) expended for education declined markedly in the early 1980s. In Botswana, Cameroon, Congo, Ghana, Somalia, Tanzania, Uganda, and Zambia, such declines began as early as 1975 (World Bank, 1988).

It is too early to tell how these cuts in expenditures have affected either total levels of educational attainment or sex imbalances in education in these countries. So far, according to Demographic and Health Survey (DHS) data from the late 1980s, there has been little obvious decline in the levels of female education (see Table 5-3). Still, the fact that in Liberia and Togo, women aged 15–19 have lower literacy rates than women aged 20–24 may indicate that increases in female education are tapering off. (Liberia's case must be assessed carefully, however, because political strife since 1980 has likely disrupted schooling.)

The quality of schooling is allegedly decreasing as well (see, for example, World Bank, 1988), though it is difficult to document this change reliably. Still, public perception is that the quality as well as the availability of education is suffering. The following statements are representative:

This section discusses the influence of education on fertility. However, the causation can also run in the other direction, that is, adolescent fertility can affect education. We consider both possibilities in turn.

Some studies report a curvilinear relationship between fertility and education: high fertility among women with no education, even higher among those with some primary education, and lowest among those with secondary education (Cochrane, 1979; Gyepi-Garbrah, 1985a; Agounké et al., 1990; République de Côte d'Ivoire, 1990; République du Cameroun, no date). But most studies conclude that fertility declines significantly as education increases. How adolescent fertility fits the contours of this generalization is less clear.

The Demographic and Health Surveys include data on the highest level of school attended: none, primary, secondary, or higher. Questions on other forms of education such as trade apprenticeships and ritual initiations were not included. Table 5-4 shows the percentage of women aged 20–24 at the time of the surveys in DHS countries who gave birth before age 20, by

level of education. Almost all the countries show a predictable negative relationship between the highest level of formal education received and giving birth as teenagers: In general, the percentage of teenage mothers is slightly lower among those with primary education than among those with no education. Women with secondary or higher education are much less likely to have given birth as teenagers. In all countries except Liberia, most women with secondary education began childbearing after reaching age 20. The sharp discontinuity in percentages of teenage mothers between women with primary school and secondary school is not surprising because many girls drop out at precisely this point. One of the reasons is that families commonly find husbands for their daughters at this time. Another is that whereas primary education is free or low cost in many countries, the expenses required of a family to support a student rise sharply in secondary school.

DHS results for women aged 15–24 reveal a strong relationship not only between education and fertility but between education and fertility goals. These results (not shown) indicate that women of secondary or higher education would be least enthusiastic about becoming pregnant in the near future, and they report smaller ideal family sizes than do less educated women.

By what mechanisms, then, does education of women produce such apparently clear fertility results among adolescents? One commonly cited

way is simply by raising the age at first marriage for women, which in turn may lead to a later start in childbearing, and hence to lower levels of fertility among adolescents (see, for example, Gyepi-Garbrah, 1985b; Division of Family Health/GTZ Support Unit, 1988). As shown in Chapter 3 (Table 3-6), in all DHS countries women with secondary or higher education are indeed less likely to marry before 20 than are less educated ones. Families appear to be increasingly willing to delay giving their daughters to husbands in order to boost the girls' educational cachets. Hence, it may be more accurate to say that desires for education—among both families and daughters—produce lower fertility, rather than education itself.

Another often-discussed way in which education may act upon adolescent fertility is through influencing values. Formal education has been charged with eroding deeply ingrained cultural or familial constraints on premarital sexual relations (Cherlin and Riley, 1986; Worthman and Whiting, 1987; Division of Family Health/GTZ Support Unit, 1988; Adeokun, 1990). Education may also expose women to Western ideals of romantic love with one partner (what Caldwell, 1982, called "emotional nucleation") and to greater sexual autonomy for women, both of which may lead to earlier ages at first sexual experience and more frequent sexual encounters. As Table 3-7 has shown, more educated women aged 20–24 are more likely to report premarital sexual activity. The principal reason, of course, is that more educated women tend to delay marriage and have more exposure to risk. However, education may change young women's values from those of high fertility to low fertility or allay their misgivings about birth control. And education undoubtedly plays an indirect role in generating social disapproval for early pregnancies.

Another possible route of fertility influence for education is that of imparting knowledge. The most obvious kind of knowledge that girls might obtain in school that would induce them to avoid pregnancy is "family life" instruction. Certainly in the past, but also today, there was considerable resistance to teaching girls about reproduction in school because such instruction might encourage licentiousness. Nowadays, many teachers and families agree in principle, as do adolescents themselves, that population and family life education should be taught in schools. Yet there is considerable debate about who should teach adolescents, what they should be taught, and when they should be taught which information. Whereas education authorities often fear parental opposition to family life education courses, families are divided: Some want to retain the responsibility; others argue that it is the school's duty (National Institute of Development Research and Documentation, University of Botswana, 1988).

Official policies in a number of countries still restrict family life education. Until recently, for example, several francophone countries rationalized these restrictions by citing laws based on the French anticontraception

law of 1920, which prohibited "antinatalist propaganda" (Senderowitz and Paxman, 1985).

But despite the often bitter disagreement over such sensitive issues, most African countries, as Table 5-5 shows, have introduced courses in the primary and particularly in the secondary school curriculum (Gyepi-Garbrah, 1985b; Senderowitz and Paxman, 1985; Gachuhi, 1986; Centre for Development and Population Activities, 1988; Kiragu, 1988; Liskin et al., 1989; Mashalaba, 1989; Haider, 1990; Nigerian Educational Research and Development Council, 1990; Ukaegbu, 1990). These courses cover reproduction, family life, and (occasionally) contraception. Reproduction and family life education courses are seldom taught at the university level, and only a few countries (Angola, Liberia, Mozambique, Nigeria, Sierra Leone, Somalia, Uganda, and Zambia) offer them in teacher training colleges.

No assessments have yet been made about how effectively these programs prevent pregnancies. Because few girls go on to secondary schools, where most family life education is introduced, it would be difficult to claim that such programs significantly affect adolescent fertility. Indeed, the percentage of women in Botswana who said they dropped out of secondary school because of pregnancy was almost identical to the percentage of pregnancy-related primary school dropouts (see Table 5-6). Even within secondary schools, information specifically on family planning may come quite late in the curriculum series. In The Gambia, family planning information is introduced only in the third year of secondary school (Mary Hafna, personal communication).

Furthermore, the extent to which family life programs are actually implemented and the amount of information they cover vary considerably. In Nigeria, few secondary school teachers actually know much about the subject outside their own informal sources (Nigerian Educational Research and Development Council, 1990); similar results were found in Kenya (Division of Family Health/GTZ Support Unit, 1988; Ferguson et al., 1988). Nigerian secondary school teachers also complain of inadequate resource materials and of insufficient time for family life lessons. Family life subjects are seldom included in national exams, a fact that decisively reduces both students' and teachers' interest in them (Dynowski-Smith, 1989; Ukaegbu, 1990).

Besides family life education per se, schools offer other kinds of information that might help girls to delay fertility. How effective is information about physiology in this effort—for example, about when during the menstrual cycle a pregnancy is most likely to occur? Data from DHS respondents aged 15–24 give at best an indirect answer to this question. Table 5-7 indicates clearly that women with more education have greater knowledge on this subject. However, in the late 1980s, fewer than one in four women aged 15–24 still had accurate information. Knowledge is somewhat better

among women who attended primary school than among those with no formal education. But in most countries significant increases in reproductive knowledge do not appear until secondary school.

To be sure, education does not always improve knowledge about reproduction. In Botswana, Liberia, and Zimbabwe, only about one in ten women with secondary or higher education could identify the fecund period. And in teacher training colleges in Kenya fewer than 60 percent of the women in the sample of trainees could identify that period (Ferguson et al., 1988), though they knew more about reproduction then did uneducated women.

School, of course, is not the only source from which women can learn how to prevent unwanted births. Friends, health clinics, women's associa-

tion, husbands, can provide such information as well. A potentially rich source of information is the media (Senderowitz and Paxman, 1985; Gachuhi, 1986; Barker and Rich, 1990; Piotrow et al., 1990; Barker et al., 1991). In most African countries, radio broadcasts reach much of the young adult population aged 15–24. In countries such as Ghana, Kenya, Mali, and Nigeria, family planning and health messages are broadcast regularly on the radio and/or television, often in local languages or through drama or musical performances (République du Mali, 1988; Rimon et al., 1988; Church and Geller, 1989; Piotrow et al., 1990). Table 5-8 displays the percentages of women aged 15–19 and 20–24 who report that they listen to the radio weekly and, for two countries only, remember hearing a family planning message. The crucial question, of course, is how listening to the radio may affect reproductive knowledge.

Taking first the issue of knowledge, the DHS results displayed in Table 5-7 confirm that women aged 15–24 who listen to the radio weekly have more accurate knowledge of the fecund period than those who do not. The results also show that there is much more knowledge among secondary school attendees than among radio listeners. However, in almost all countries, women who listen to the radio weekly know the fecund period as

accurately as do women who attended primary school, if not more so. In fact, in some instances, the radio messages are more effective than secondary or higher education. In separate calculations for Ondo State, Nigeria, women who listen to the radio weekly are slightly more knowledgeable about the reproductive cycle than women with secondary or higher education—about 22 and 20 percent, respectively.

Associations in the DHS data between radio listening and use of contraceptives are even more striking than those for knowledge about reproduction. In most countries, women from 15 to 24 who listened to the radio were far more likely than primary school attendees to have used contraceptives. The Kenya DHS pressed further, asking women specifically whether they had heard family planning messages on the radio. Our calculations (not shown) revealed that at every educational level, women who reported hearing family planning messages on the radio were more likely than those who did not to know the fecund period and to use contraception. In fact, women with secondary or higher education who reported they had not heard radio family planning messages were only slightly more likely to use contraceptives than women with no education at all who heard these messages.

It is by no means clear, of course, that radio messages actually teach women about reproduction or induce them to use contraception. These results simply suggest that the content of formal education, at least through primary school and quite frequently through secondary school, may be less effective than conventional wisdom suggests in changing reproductive behavior or in imparting knowledge of the reproductive cycle.

Clearly, women's behavior concerning fertility, as well as their reproductive knowledge and contraceptive use, vary with their education. However, many of the differentials presented above underscore the presence of threshold effects between primary and secondary school education in Africa. Because women with secondary schooling or more still account for only a small proportion of African women, the influences of formal education on adolescent fertility should not be overemphasized.

The earlier description of Sierra Leone suggested that schoolgirls should be off limits sexually. Yet schoolgirls do become pregnant, with what some regard as shocking frequency. Regardless of her age, many African societies view a girl as an adult, or at least a non-child, after she gives birth: They expect her to assume adult responsibilities and to terminate her education (Mwateba et al., 1988). Most students have been sent to school less to better themselves as individuals, than to serve as representatives of families that seek to better themselves. A family may have sacrificed a great deal to educate one child out of a set of siblings, and may have spent considerable

time selecting and preparing that child. For a girl who carries such hopes, pregnancy is a serious family tragedy. Why, then, do such pregnancies occur?

Some researchers posit poor school performance as a prime cause of pregnancy-related school dropouts. In a Kenyan study more than one in three of the girls in the bottom 25 percent in school achievement dropped out because they were pregnant, compared with only one in seven of the best students (Division of Family Health/GTZ Support Unit, 1988). The difference may arise because poor students must repeat classes, thus exposing them to the risk of pregnancy for a longer period. But it is mainly those girls who do not perform well in school, whether because of poor ability or the press of chores at home, who have the least to gain from staying in school (Division of Family Health/GTZ Support Unit, 1988). Some low achievers may deliberately try to become pregnant in order to drop out of school (National Institute of Development Research and Documentation, University of Botswana, 1988; Dynowski-Smith, 1989). Again, however, separating poor achievement from economic handicaps is difficult. In Kenya community-funded (''Harambee'') secondary schools, which typically enroll students from poorer families, had pregnancy-related dropout rates more than twice as high as government-maintained and private schools (Division of Family Health/GTZ Support Unit, 1988). Another piece of evidence that supports the argument for economic hardship is that boarders and students in girls-only schools in Kenya had lower dropout rates than did day students and those in coeducational boarding schools, even though boys and girls are housed separately (Division of Family Health/GTZ Support Unit, 1988).

Related to economic hardship and academic difficulties is a final source of schoolgirl pregnancies: "sugar daddies." Secondary education, especially, requires substantial commitments of cash for school fees, uniforms, books, school supplies, and examination fees. Because families usually place priority on boys' educations, girls are less able to buy essential materials; so to help finance their education, some girls strike up sexual relationships with wealthy older men, their sugar daddies, who may provide places to live and money for food, as well as school fees and expenses (Busia, 1950; Schapera, 1971; Dinan, 1983; Brandon and Bledsoe, 1988; Brokensha, 1988). The payoff is judged worth the risk of becoming pregnant because the financial support that girls receive allows them to remain in school, as long as they can avoid pregnancy (see, for example, Acquah, 1958, for Accra). Akuffo (1987), studying a small Ghanaian town, found that about 93 percent of girls aged 15–19 had boyfriends who were employed, including some teachers (see also Bleek, 1976). So common has the sugar-daddy pattern become, that in rural Ghana over the last 40 years, the basis for imposing fines on men for pregnancy has been shifted from adultery to relations with schoolgirls (Vellenga, 1974).

There is some danger in overestimating the prevalence of the sugar-daddy syndrome. Schoolgirls are a highly visible and countable subsection of the female population, and families are understandably vocal about losing their return on years of school fee payments in cases of pregnancy. Yet, we cannot assume that the syndrome is everywhere a major cause of school pregnancy. The Kenyan Ministry of Health reports that schoolboys are responsible for most schoolgirl pregnancies in that country (Division of Family Health/GTZ Support Unit, 1988). Nor can we assume that sugar daddies always provide much support. Because parents usually pay most of girls' costs during their training, it may be ancillary expenses—clothes, uniforms, and pocket money—that are earned through sexual companionship (Akuffo, 1987). Even in Botswana, where most schoolgirls' lovers were adult men and not their schoolmates, only 11 percent of all mothers of out-of-wedlock children reported receiving support (National Institute of Development Research and Documentation, University of Botswana, 1988). But, the tremendous intercountry and intracountry variation in the frequency of sugar-daddy-schoolgirl situations makes it difficult to draw any general conclusions from the limited information cited for one or two countries.

Trying to untangle the causal relation between fertility and education is not an easy task. Earlier we presented evidence that educational attainment affects fertility desires and behavior, as well as contraceptive use. Yet the preceding section described some strong evidence that adolescent fertility also affects educational attainment.

In the past, when girls had few expectations of formal education, pregnancy posed little disruption in a young woman's life course. Pregnancy was, and still is, encouraged, especially in areas where women marry young. Nowadays, pregnancy decisively mars the prospects of girls who hope to pursue their education. In countries where secondary education is a real option, a schoolgirl may walk a fine line between pregnancy and motherhood, on the one hand, and, on the other, finishing an education, obtaining a prestigious urban job, and marrying a monogamous man. The widespread perception is that a girl who becomes pregnant in school and drops out may have to accept a low-paid job, enter a premature marriage, or become the head of an impoverished household, relying on meager assistance from her family and the children's fathers. Gyepi-Garbrah (1985a:22–23) writes:

Stories in the popular press frequently confirm that having a child effectively terminates a girl's education (Akuffo, 1987; Bledsoe, 1990). Interview data suggest that many local people hold this perception as well. One young Nigerian woman stressed that parents discourage sex for their adolescent daughters "[b]ecause the parents want them to finish school" (quoted in Hollos and Leis, 1989:121). Kenya is another country where advancement in education may be contingent on low fertility. Most teacher training colleges reportedly use pregnancy screening as a prerequisite to admission, although this practice is extremely controversial even among the staff (Ferguson et al., 1988). The more common way in which fertility affects education is that girls who become pregnant drop out of school, often through pressure from explicit policies that expel them and discourage them from returning after giving birth (République du Mali, 1988; Adje, 1992).

Numerical evidence, though spotty, supports these observations that early fertility precludes further education. Botswana was the only DHS country that included a direct question about whether a woman had dropped out of school because of a pregnancy. Reproduced in Table 5-6, these Botswana results show that 14 percent of primary school attendees and 15 percent of secondary school attendees had dropped out because of pregnancy—estimates that are probably low. Among the primary school dropouts, only 20 percent subsequently reenrolled, and only 17 percent of secondary school dropouts did so.

A number of studies concur that many young women drop out of school as a result of pregnancy (Gyepi-Garbrah, 1985a). In Kenya, a study conducted in 1985 estimated that about 10 percent of female students drop out of secondary school because they are pregnant (Division of Family Health/GTZ Support Unit, 1988; Barker and Rich, 1990). In 1986, 11,000 Kenyan girls dropped out as a result of pregnancy (Kiragu, 1988). In the Machakos area, 27 percent of families reported having a teenage daughter who dropped out of school because of pregnancy (Maggwa, 1987, cited in Division of Family Health/GTZ Support Unit, 1988). In Mali, in the years 1983–1987, 10 percent of female secondary school students were dismissed because of pregnancy (République du Mali, 1988); and in Zambia it is estimated that 2 percent of all primary and secondary schoolgirls were expelled from school because of pregnancy (Senderowitz and Paxman, 1985). (For Ghana, see Akuffo, 1987; and Oppong and Abu, 1987.) Pregnancy-related dropouts are widespread even at advanced levels. A survey of women in Kenya's teacher training colleges indicated that annual pregnancy-related dropout rates among enrolled women students were 7.1 percent in 1986 and 6.4 percent in 1987 (Ferguson et al., 1988).

Pregnancy rates, of course, exceed dropout rates due to pregnancy. Some girls who become pregnant near the end of their courses are allowed to complete their programs (Ferguson et al., 1988; Hollos and Leis, 1989). But many cases of schoolgirl pregnancies go unrecorded. Some girls withdraw quickly from school on other pretexts when they discover their pregnancies and go to live with distant relatives. They return after giving birth, leaving the baby with the relatives.

The most controversial fate of many schoolgirl pregnancies is abortion. A number of girls reportedly resort to this strategy to avoid detection and expulsion (see, for example, Bleek, 1978; Gyepi-Garbrah, 1985a; Senderowitz and Paxman, 1985; Nichols et al., 1986; Akuffo, 1987; Nichols et al., 1987; République de Côte d'Ivoire, 1990). Because a substantial proportion of young women conceive while at school or training institutions, the option of abortion is an important factor in their educational prospects (Oppong and Abu, 1987).

Although most girls must drop out when their pregnancies are discovered, what are their prospects for resuming education? Some countries have relatively accommodating policies concerning the readmission of young women who have been pregnant or given birth. In Kenya, schoolgirls who become pregnant are required to discontinue their studies for at least a year (Ferguson et al., 1988). Policies in many countries are much stricter; they decree that pregnant schoolgirls be expelled permanently. Table 5-9 summarizes the available evidence on policies related to schoolgirl pregnancy in eight sub-Saharan countries. All eight favor some form of expulsion, in many cases without provision for reentry after giving birth.

Even in the most liberal countries, a combination of policies, economic circumstances, and informal social pressures discourages all but the most advanced or determined students from reenrolling after a pregnancy. A few countries allow girls who have given birth to apply for readmission or for admission to a school other than the one they left (Gyepi-Garbrah, 1985a). But few girls do so. In Botswana, the DHS revealed, fewer than 20 percent return to school after a pregnancy. Part of the reason is that in rural areas with widely scattered schools, the policy of requiring adolescent mothers to apply to other schools would force most girls to move to other towns, leaving their families to care for the child (Dynowski-Smith, 1989). Gyepi-Garbrah (1985a:22–23) elaborates other reasons:

Child care itself imposes constraints on many would-be returnees. In theory, young mothers might return to school after giving birth by leaving

their babies with their families, but there is little evidence that they do so. Are female students who are mothers, few though they be, more likely to have children living away than are their peers who do not go to school? Unfortunately, though many of the young women aged 15–19 in some countries must have been students, neither the World Fertility Survey nor the Demographic and Health Survey asked them (or other women) whether they were still in school. A few individual surveys, however, provide some evidence. Surprisingly, according to a study in Côte d'Ivoire, 30 percent of schoolgirls who dropped out because of pregnancy returned after the delivery (République de Côte d'Ivoire, 1990); but most of those who returned reported that they could do so only because their parents or their partner's parents had been able to take care of the child. Some young women may not have such fostering opportunities. Although pregnant schoolgirls in Monrovia, Liberia, were allowed to transfer to night school, few could do so because of childcare difficulties (Gyepi-Garbrah, 1985a; see also Oppong and Abu, 1987).

Whereas strict policies limit educational achievements, informal social pressures may be equally effective. As a young woman in Sierra Leone related, people often taunt an adolescent mother who tries to return to school, and use her past to shame her:

Clearly, when childbearing begins, education usually stops. Women who begin to bear children young will have little chance to continue their education. Conversely, women who delay childbearing—whether they use abstinence, abortion, or contraception—or soften its consequences by fostering out their infants, can continue their education.

It is important to reiterate that there is tremendous heterogeneity among African countries in the number of teenage girls who attend school. Although cases of pregnancy-related school dropouts draw considerable publicity, in some countries the total number of cases may be quite small. On the other hand, the number of pregnancy-related dropouts is larger than we might suspect if we confined the analysis to secondary school students: Because of liberal age guidelines for enrolling in school in many countries, many teenage schoolgirls are still in primary school. These girls run similar risks to those of their age peers in secondary school.

As shown in Chapter 4, preparation for childbearing is a key component of sanctioned reproduction in Africa. This chapter has examined the most prestigious form of contemporary preparation for adult life: formal education. We have seen that primary and secondary enrollment rates, especially those for women, rose rapidly in Africa during the 1970s and 1980s. Even widespread declines in public expenditures in the early 1980s have so far had little effect on the percentage of women aged 15–24 who attended primary or secondary school. Enrollment rates for women also rose more rapidly over the period than did those for men, so that the ratio of boys to girls enrolled in school came into better balance. How have these dramatic changes in education of women over the last 20 years affected adolescent fertility?

Before education became widely available to them, girls were trained for roles as wives and mothers. Today, many have ambitions outside the home in the formal labor market, and they often delay childbearing in order to acquire more education. Not surprisingly, the probability of giving birth as a teenager is slightly lower for women with primary education than for women with no education, and much lower for women with secondary or higher education.

In trying to explain these results, we have avoided the conventional assumption that educational achievement necessarily lowers fertility because we are dealing with adolescents, among whom this kind of causality may be reversed: Reproductive events may curtail opportunities for formal education. Certainly the fact that schoolgirl pregnancies are usually followed by dropping out indicates that the usual logic of temporal order may not apply. Without becoming embroiled in the intractable debate about the direction of the education-fertility relationship, we stress that social changes are shaping both education and adolescent fertility.

What we can say with some confidence is that education holds out rewards of social and economic advancement to girls who can delay childbearing to pursue their schooling. Equally powerful are the negative sanctions that discourage pregnancy. Although some observers have argued that formal education increases premarital fertility by eroding traditional checks on sexuality and by increasing promiscuity among unsupervised adolescents, such arguments seldom take into account the very real threats of expulsion from school that girls face if they become pregnant. Obviously, such strictures fail in many instances, as the proportion of schoolgirl pregnancies attests. Still, the overall effects of school policies that discourage reproduction are undoubtedly quite powerful.

One of the most important social changes that shapes the education-fertility relationship among adolescents is marriage. If overall fertility among

adolescents in Africa is changing less dramatically than premarital fertility (or, perhaps more accurately, the proportion of children without full paternal recognition), what role does education play in this trend? Both DHS data and ethnographic descriptions leave little doubt that marriage, like fertility, almost always terminates a woman's education. Indeed, although a number of girls withdraw early from school because of pregnancy, others withdraw because their families pull them out of school to give them to husbands. Conversely, a girl who delays marriage—or a sanctioned procreative relationship with a man—has a better chance than others do of going on in school. Yet for educated women, delaying marriage may also precipitate what we call premarital pregnancies simply because such women may prefer to label themselves as unmarried. Despite the stigma they suffer, this strategy may leave the door open wider to the possibility, however remote, of resuming schooling at a later date.

In the realms of both marriage and fertility, then, education appears to have the effect of elongating the period separating menarche from fertility and marriage—or what women are willing to call marriage—and of expanding the probability that a woman will have a premarital conception. Schooling may also make marital status itself even more ambiguous than it was before: Young women at certain junctures of the conjugal process who previously would have declared themselves to be married may now be quite ambivalent about proceeding with a certain conjugal relationship. Trying to keep their options open, they may be more likely to declare themselves as unmarried.

The assertion that many young women nowadays would like to delay fertility in order to pursue training that prepares them for adult responsibilities should arouse little argument. Yet Chapter 4 stresses that education is only one of many other avenues of preparation for adult life—and thus childbearing—in Africa. The notion that training, more broadly defined, may affect the onset of childbearing is relatively unexplored in the literature on fertility. To flesh out its dimensions, the next chapter turns from schooling to other forms of training.