4.1 Economic constraints
4.2 Girls' labour contributions
4.3 Family, community and social class
4.4. School-based factors
Equality of access is often seen as being the crucial factor affecting girl's education. But, from the experience of many SSA countries; including Zimbabwe, Zambia and Malawi, it is clear that persistence and attainment in the upper primary grades and at the secondary level is also a major problem for girls. Evidence from these three countries shows that boys and girls are enrolled almost at parity for primary school as a whole and yet educational outcomes for girls are still relatively poor. Even in Zimbabwe, where educational provision has expanded dramatically since independence, girls have been unable to fulfil their potential. The really crucial question is, therefore, the degree to which there is equality of process in the education system.:
Only a few studies have tried to address the factors affecting the poor educational performance of girls in Malawi, Zimbabwe and Zambia. Until the current World Bank initiative (involving the University of Zambia) reports its findings, Zambia appears to have the least amount of published research on the process of gender inequality in education particularly at a more detailed, school based-level. The various factors which have been found to adversely affect the successful participation of girls in relation to boys in the education system apply at every level. The constraints operating against girls are often divided into school-based and non-school based factors and they are multicausal, spanning a wide range of social, economic and political factors.
While the division between school-based and non-school based influences on girls' participation is often used for analytical convenience, it should not obscure the fact that inequalities stem from gender relations in society at large and these are reflected in and played out in the school system itself. While there has been considerable debate about the relative importance of school and non-school factors in the educational achievement of girls, no firm conclusions have been reached. In a large survey of Kenyan primary schools in 1994, Appleton concludes that school level variables appear less important for girls than boys in relation to student level variables. Using econometric modelling techniques, he found that girls were less responsive than boys to school level variables such as the quality of teaching staff (Appleton, 1994:78). The most plausible explanation is that the school is replicating the values of the home and community in terms of undervaluing girl's potential and abilities compared to boys. In his view, such attitudes might have the effect of inhibiting girls, making them psychologically distanced from schooling and therefore less sensitive to school conditions.
In marked contrast, Hyde stresses the importance of focusing on the institutional and pedagogical effects of school based changes in Malawi. Her research shows that a student's perceptions of encouragement and support from teachers can contribute significantly to the academic performance of females (Hyde, 1994). The positive results of the Cambridge Federation for Educational Development (CAMFED) project in Zimbabwe to support girls' through secondary school would also seem to support this supposition. A study of girls' achievement in Botswana's junior secondary schools suggests another possibility, namely that where family and community context strongly influence children's length of school attendance and their level of achievement, little effect would be expected from the school itself (Fuller et al, 1994). However, what seems to be happening is that the school simply reinforces gender differentials already set up in the household. What needs to be explored, therefore, is the juxtaposition of household, community and school. Gender relations are formed and expressed in all these areas. Limiting the discussion to 'supply and demand' factors is too simplistic to capture the process of gender inequality.
The following sections will outline some of the factors in the economy, school, community and household which affect the educational opportunities and achievements of females.
Research on Malawi, Zambia and Zimbabwe shows that one of the main reasons for low female participation and high drop out rates are economic constraints which have increased both the direct and indirect costs of education for girls and their parents. Education cost recovery and cost reduction measures have been in place in Zimbabwe since the 1990s and were imposed earlier in Zambia. In Zimbabwe, urban primary school fees were introduced in January 1992. Although some safety nets have been created (based mainly on the Social Development Fund), the real challenge is whether these can be implemented and targeted effectively (Chisvo, 1993). The introduction of school fees and the real cost of sitting international exams (due to devaluation) have raised the overall cost of education whilst real incomes are concomitantly being erode by recession. In Zambia, which has suffered massive declines in spending on social sectors in response to persistent economic crisis, a recent poverty assessment (1994) showed a strong correlation between primary school attendance and household poverty levels (Republic of Zambia, 1995).
The public sector employs a disproportionate number of women in the formal sector. Public expenditure cutbacks, therefore, can be seen as a particular threat for women on two fronts: school places are jeopardised and employment opportunities for those who attend school are reduced, a situation that will further lower the economic returns to education for women (Hyde, 1989). In Africa, devaluations which have been introduced as part of SAPs have been associated with sharp, often extreme falls in the value of salaries in the public sector, including those of school teachers (Colclough with Lewin, 1993). In particular, the real value of civil servants salaries has fallen precipitously in Zimbabwe and Zambia. During the 1990s, the employment of teachers in both primary and secondary schools in Zimbabwe has declined, with a growing number of male teachers seeking better paid jobs elsewhere. They are rarely replaced because of tight budgetary conditions (Chisvo, 1993). Consequently, higher pupil to teacher ratios have contributed to the declining quality of schooling. Although it is difficult to relate conclusively changes in social sector expenditure under adjustment to shifts in health and educational status of the population, evidence of increased poverty in these countries would suggest that to be the case.
A major constraint against the participation of girls in education is the real and perceived lack of labour market opportunities. It is generally recognised that boys have a wider range of choice with regard to jobs due to labour market segregation even in a context of shrinking formal sector markets. Females comprise between 15-20% of total wage employment in Malawi, Zambia and Zimbabwe (see Table 6). Time series data available for Malawi and Zimbabwe suggest that there was a slight upward trend in this share during the 1980s. In Zambia, on the other hand, the fragmentary data for the period 1980-86 indicate that the proportion of total formal sector employment accounted for by women fell from 24.5% to 16.1% in 1986.
With very sluggish and, in recent years, probably negative growth in wage employment in all three countries, the employment prospects for female school leavers remain as limited as ever. In Malawi, annual job openings for females, (i.e. replacements to make up for attrition and new jobs) have averaged no more than 10-15,000 during the last five years, which is no more than 1.5-2.0% of total female primary school enrolment. The corresponding guesstimates for Zambia and Zimbabwe are 0.5-1.0% and 1.0-1.5% respectively.
Given intense job competition, only graduates from higher education institutions and upper secondary school leavers stand any chance of eventually obtaining some type of skilled job in all three countries. So long as job prospects for all youth remain so grim, the vast majority of primary and lower secondary school leavers will continue to be absorbed in farm and non-farm household production in the rural areas and the urban informal sector. The policy implications are equally clear. Given that monetary benefits from education investments will continue to remain very low for the foreseeable future and the overall incidence of poverty is unlikely to decrease markedly, only a radical reduction in the overall costs of female education will encourage more parents to send their daughters to school. To achieve this, governments must certainly minimise the direct costs of schooling, but, with indirect (i.e. opportunity) costs typically comprising well over half of the total (household) costs of female schooling, higher levels of financial support along with other interventions may be necessary. However, with governments being forced to slash public expenditures, the ability to pay for the required policy measures is very limited.
If household incomes are declining, the real cost of education as a proportion of household income increases, which is likely to exacerbate drop-out rates for both male and female children. However, at the intra-household level there are often biases in the allocation of resources that stem from prevailing gender relations. Research in SSA has shown that parents are often more inclined to send their boys to school in the first place and keep them there longer, mainly because of the better labour market prospects for boys compared with girls (Davidson and Kanyuka, 1992). There are additional 'hidden costs' to parents of educating girls given the considerably greater inputs of domestic labour contributed to the household by girls than boys.
Female domestic labour is a key factor that militates against girls' achievement at school and is also a very sizeable opportunity cost for parents when they make a choice about whether to send a child to school. In other words, the need for female domestic labour affects decisions about whether to send a daughter to school in the first place and, once at school, how long she should stay. Not only do women in SSA undertake the major tasks in agricultural production, they also perform tasks associated with biological and social reproduction such as pregnancy, cooking, caring for the young, sick and old and all service work to sustain households (Mbilinyi et al, 1991). Furthermore, in many rural areas of Malawi, Zambia and Zimbabwe, high proportions of households are headed by women who are the sole income earners. In Zimbabwe, for example, 30-35% of households are headed by women (UNICEF, 1994). The task of economic and social reproduction of the household, therefore, falls heavily on women.
In Ghana, recorded time burdens are up to 20% higher for females than males across many/age groups and occupational status (Haddad, 1991). Girls attend school more infrequently and less intensively and that they are not given the same kind of attention by teachers as boys and they take less demanding subjects. These household tasks performed by girls and women are not remunerated and they detract from commitments to education and income generating activities (Haddad, 1991).
An important issue is that conventional measures of opportunity cost based on prevailing wage rates for children may seriously undervalue the real opportunity costs to parents of children (and girls in particular) attending school. This may explain why parental demand for the education of daughters is so much lower than would appear to exist given the allegedly very high private rates of return to primary and secondary education in SSA (see Psacharopoulos, 1994). Generally, parents attach a much higher value to female labour than that of boys but prevailing wages do not generally reflect this (lion, 1992). Whilst virtually all boys and girls participate to some extent in all household activities, gender differences in this participation are substantial. In Zimbabwe, both girls and boys in rural areas spend an average of about 5 hours per week of labour more than their urban counterparts. Girls work an average of 25 hours per week (i.e. the equivalent of a half time job in industrialised countries) which is 7 hours a week more than boys. Similarly, a study in Malawi reports that girls spend 30 minutes per hour of the day on domestic chores whereas boys spend only 12 minutes (GABLE, 1995). The picture of girls' labour being valued more highly than that of boys is fairly universal across the Central and Southern Africa, except in areas (as in rural Botswana) where boys herd cattle. In these situations, girl's enrolment at primary school usually outnumbers boys substantially. And, in poorer households there tends to be a greater demand for domestic labour from girl children. Conversely, in higher socio-economic groups, it has been found that boys tend to contribute more labour to the family than girls (lion, 1992).
It is important to realise that domestic labour is not merely a lost opportunity cost for parents but it is also a loss of energy which could be devoted by the child to school work. The extra labour expended by girls in particular on domestic chores affects school work in terms of time lost, concentration levels in the classroom and available time for home work. Girls also have higher costs in getting to school than boys which are a further drain on time and energy. Walking is by far the most common way of getting to school and rural girls spend more time walking, on average, than other groups; 87% of rural girls in Zimbabwe walk to school and spend nearly an hour walking to and from school. It is notable that girls walk more than boys. Boys in rural areas have more transportation options than girls (by bicycle and bus, for example). Walking involves loss of energy and parents lose the labour of children.
The direct costs for schooling can also differ substantially according to gender. In Zimbabwe parents opted to give their sons more 'discretionary' spending money than girls (Ilon, 1992). In all three countries, the costs of uniforms for girls are generally higher than for boys. A pilot survey of parents and school children in the Machinga district of Malawi found that those families with both boys and girls spent significantly more on sons than daughters in terms of money for schooling (Hyde and Kadzamira, 1994). More detailed research of these direct and indirect costs is needed in order to help explain why parents withdraw their girls from school more readily than boys.
4.3.1 Parental attitudes and influence
4.3.2. Initiation and early marriage
4.3.3. Sexuality and sexual harassment
The limited evidence that is available suggests that socio-economic status, region of residence and religion are more important factors influencing girls education than among boys (Hyde, 1989). However, in Malawi, some research indicates that family background variables are of lower explanatory power than either school variables or other variables such as urban or rural residence. In the literature on gender and education in SSA, there is little serious examination of social class as a factor in the educational participation of girls. However, it has been widely observed that the parents of female students at primary and secondary school are more educated than those of males (see for example, ILO, 1990 and de la Gorgendiere, 1995, Hyde and Kadzamira, 1994). It is also well known that better educated mothers are likely to educate their daughters.
A recent survey of teachers, university students and secondary students in Ghana which looked at the educational background of parents of students illustrates graphically the multiplier effect of female education. Female students who had mothers with higher levels of education were themselves given opportunities and funding to promote their own education to higher levels. Only a very small minority of female students at university and secondary school levels had mothers with no education at all or education limited to primary school. The implication of this is that for the multiplier effect to work it is important for females to be educated beyond the primary level (de la Gorgendiere, 1995). Other studies have also shown that the impact on daughter's schooling to be greater if the mother has secondary education.
In Malawi, more educated parents tend to send their children to school at an earlier age and this tends to be more pronounced for girls (Hyde, 1992). However, the better educational background of girl student's in Malawi does not seem to make much difference to academic performance, although it is not clear whether this has been adequately researched by existing studies. A study of social class and academic performance in the Cameroon (Cammish and Brock, 1994) suggests that girls from elite families overcome the disadvantages of their sex and have a higher pass rate at secondary school than boys from almost all other occupational backgrounds. A survey of exam performance in primary schools in Kenya showed that social and economic status is strongly correlated with terminal examination performance. Offspring of professional fathers scored an average of 60 more marks than farmer's children. The Kenyan study also found that students who ate more than three meals obtained higher scores, as did girls of higher than average height (Appleton, 1994). These factors can also be related to the better financial background of parents. However, in both these cases the girls attended better schools than boys. More attention needs to be paid to social class as an element in gender disadvantage.
Parental attitudes concerning education are clearly a key factor since they often help influence which children go to school and how long they stay there. Clearly, the attitudes of parents reflect those in the society at large and are deeply embedded in prevailing cultural norms and values. However, the role of parental attitudes in shaping the educational aspirations of their children is unevenly documented for Africa.
Much of the primary research on gender and education in Malawi in the 1990s has taken place in schools and communities in the south of Malawi, which has the lowest socio-economic indicators and literacy rates. A major assumption underlying the Girls' Attainment in Basic Literacy (GABLE) project (financed by USAID) in Malawi, which seeks to promote the higher participation and achievement of girls at primary school, is that the attitude of parents has an important impact on the schooling of girls. The fact that drop out is heavily concentrated in the primary school years suggests that it occurs when girls are still largely under parental influence (Hyde and Kadzamira, 1994). Furthermore, girls from poor rural communities have a higher degree of social control exerted over their activities than boys. Males tend to be more independent than females and they are often able to make choices that lie outside any obligations to family and household welfare (Evans, 1991).
Attitudes that reflect gender structuring are equally present at home and school, serving to circumscribe the opportunities available for girls compared to boys for education. Many of the parents who cited negative parental attitudes towards educating girls as a key factor for female drop out rates in one district in Zambia were not formally educated. Mothers seemed to recognise that perhaps their own attitudes towards education in general and towards educating girls especially might negatively impact on their daughter's persistence at school (Hyde and Kadzamira, 1994).
In an ethnographic study of factors influencing persistence of girls in comparison to boys at the primary school level in the Zomba district, 68% of fathers/male guardians and 70.3% of mothers and female guardians stated that boys are more intelligent than girls. This is despite the fact that the vast majority of respondents asserted that it was important to educate both girls and boys (ABEL, 1992). Furthermore, only 1.4% of female guardians and 8.8% of male guardians thought that it was more important to educate boys because they are more likely to return the educational investment whereas girls do not concentrate because they are preoccupied with being mothers and getting married (Ibid). The results of a survey of parental attitudes at primary schools in Kenya suggested that the relative performance of girls and boy children is self-fulfilling. If they think women are less able, then their daughters will indeed be less able (Appleton, 1994).
Presumably, because they consider the primary role of girls to be 'wives and mothers', parents consider that their sons should reach a higher educational level than girls (ABEL, 1992 and Davidson and Kanyuka, 1992). By Standard 8 in Malawi, the impact of parental attitudes as a reason for girls dropping out is almost negligible (Ibid)). This is probably because most girls who are going to drop out have already done so. Research sponsored by USAID in Malawi in 1992/3 revealed that the decision to leave school is frequently made by the girls themselves rather than their parents. 'Lack of interest' was cited by parents to explain their daughter's decision to withdraw from school. This took place in a context where the average classroom to pupil ratio in Malawi averaged 117 (GABLE, 1995).
The GABLE pilot survey of parents and school-aged children in Malawi's Machinga district found that the vast majority of parents (over 97%), felt that it was a good idea to get more girls into school. The most commonly cited reason was that it was necessary for national development, improved chances of employment, provided literacy and enabled girls to help their parents. For the small minority of those who felt educating girls was not a good idea, the principle reason was the risk of pregnancy (Hyde and Kadzamira, 1994). This survey also highlighted the problem of rigid gender roles ascribed to both male and female children. When asked what specific skills girls gained at school, the majority (76%) referred to domestic skills of various types and less than 4% mentioned job skills. All but one of the focus groups felt that the behaviour of schooled and unschooled girls differed, all of them positive for schooled girls. The most commonly cited reason by the groups was that girls who had been to school were more cultured and disciplined. Four-fifths of the respondents felt that schooling would promote the farming skills of their daughters through improved knowledge as farming is part of the curriculum. However, it was clear that schooling did not significantly enhance suitability as a marriage partner.
This survey reveals a strong desire for literacy on the part of the parents as individuals, with 90% of parents wishing that they had obtained more education. The dominant reason for wanting to be literate was access to employment opportunities.
There is limited research on parental attitudes towards girls education in Zimbabwe and Zambia, although Gordon mentions that teacher's negative attitudes towards the abilities of girls is much the same as those of parents (Gordon, 1993).
Girls' own attitudes to schooling: To a large extent, girls' attitudes tend to reflect those of their parents. In the GABLE survey in Malawi, girls were asked what they liked about school; the great majority cited various school subjects; reading and writing, also games with friends. Most saw primary schooling as terminal. Both parents and children in the GABLE study showed little knowledge of educational opportunities and 82% of girls reported that they had no knowledge of any girls in their villages who had gone further than the desired level. The main constraint against girls reaching a desired level of education was lack of school fees. In many cases, their attitudes reflected a real lack of educational opportunities in the area beyond primary school. Both girls and parents in this area of southern Malawi have very few role models of educated women in their communities. Almost all girls expected to have a job and the most popular occupations were teacher, nanny, nurse and doctor (Hyde and Kadzamira, 1994). What is lacking in this survey of school children is any exploration their views towards sexuality.
Initiation is a cultural practice which in some cases interferes with a girl's schooling. In the GABLE survey in Malawi, 72% of a sample of parents from 10 schools in Machinga District had either sent their daughters for initiation of intended to do so. Initiation in Malawi takes place in the long holidays and supposedly does not interfere with schooling. Often girls are taken away from school after initiation, although due to lack of information it is difficult to gauge precisely its impact on schooling (Hyde and Kadzamira, 1994). In other parts of Africa, girls are withdrawn permanently from school after initiation having been seen by their communities as 'becoming a woman' (Maimbolwa-Sinyangwe, 1994).
Initiation rites in Malawi appear to vary in structure and length in different communities and regions. They may form a series of rites and rituals that will continue for up to a month. It has been noted that amongst the matrilineal groups in Malawi initiation is regarded as a necessity and more emphasis is placed on the education of boys (Kapakasa, 1992). However, Zeitlyn observes that the relationship between schooling and initiation is not entirely clear from existing research. It is obviously important to understand further how communities view adolescence and girlhood in relation to education (Zeitlyn, 1994).
Role conflicts emerge for girls with regard to initiation and the practice is often a financial drain, depleting resources which could have been devoted to schooling. The GABLE Pilot Survey for Malawi recommended that churches should take a strong stand on initiation in order to reduce the rate of female drop out because of the practice (Hyde, 1994).
Early marriage in many instances is decided upon by parents or guardians. In Zambia, as the girl grows older, her family is concerned about the possibility of her getting pregnant out of wedlock and having a child before a girl gets married often implies that the family cannot demand a high bride price if she does get married eventually. Where there are financial constraints, parents may choose to marry a daughter in order to use the bride price to send her siblings to school or simply reduce the burden of having to support too many children. Many parents still believe that they stand to gain more by educating sons than daughters who end up getting married and leaving their husbands to control the family income (Maimbolwa-Sinyangwe, 1994). A conference on 'Girls Education in Africa' held in London in April 1995 and attended by a large number of women from SSA countries, recommended that early marriages for girls should be discouraged. Both early marriage and initiation were seen by the mostly African participants as being serious constraints against girls' education.
The wide age range of students attending primary schools is an important feature of the education system in Malawi. Children three or more years older than their expected age comprise 60% or more of all primary enrolments after Standard 1. This slows down instruction for all and for this reason GABLE has suggested that the Government of Malawi (GOM) adopts a maximum age of entry policy for schools. A high proportion of children attending the last grades of primary school are in their late teens and, according to custom, are ready for marriage. In Malawi, the GABLE pilot survey reveals that children in rural Malawi are expected to start forming their own households at around 15 years (Hyde, 1994). Girls get married at that age and boys often find employment or start farming. As most children do not start school until they are about 10 and about one-third repeat Standard 1, by the time they are 15 they are unlikely to have progressed beyond Standard 4. In this context, it would seem sensible not only to pressure the communities to encourage the later marriage of girls but also to encourage children to start school earlier.
In Zambia and Zimbabwe, the age range of primary school pupils is less extreme than in Malawi, although girls are usually somewhat younger than boys. However, in rural areas children generally go to school later than in urban areas.
Sexuality is another major area of concern affecting female participation in education. The problems surrounding sexuality inevitably straddle both non-school based and school based factors and they also reflect wider issues of power and control within particular societies. The following discussion will attempt to show how attitudes and norms surrounding sexuality influence girls' educational outcomes. The final recommendations of an extensive report on gender and education in Tanzania produced for SIDA commented that more research was needed on the issue of sexuality, which was described as a 'burning issue for our youth' (Mbilinyi et al, 1991). This is equally true in Malawi, Zambia and Zimbabwe. The whole issue of the impact of sexuality on schooling has been insufficiently considered by educational researchers even though it is well known that the rates of both teenage pregnancy and HIV/AIDS infection are high, particularly amongst young females in all three countries.
The two major reasons for girls dropping out of school are poverty (inability to pay for school fees and uniforms and high opportunity costs) and pregnancy. The way in which pregnancy is dealt with as a policy issue and the attitudes of parents and teaching staff to girls' sexuality reflects the profoundly anti-female bias in these societies. A pilot study in 1994 of schools in Lusaka township reported that one-third of girls had dropped out of school through pregnancy (Maimbolwa-Sinyangwe, 1994). Up until 1993, girls who become pregnant were expelled from school. Since then, the boys responsible can also be expelled but only if the girl's parents insist. The new education policy for Zambia (1995) now makes it possible for the re-admission of girls who have left to have babies.
In Malawi, a similar situation has existed with regard to official policies of expulsion for pregnant school girls. The current policy (as of 1594) at all levels of education is that girls who become pregnant must leave school. They can only be readmitted at tertiary level if they can prove that adequate arrangements have been made for care of the child. Furthermore, up until recently, a medical examination to check for signs of pregnancy was carried out on all school girls each term (Hyde and Kadzamira, 1994). For many of those who leave school because of pregnancy, this is the end of their education.
Davidson and Kanyuka (1992) reported in their survey of households in the Zomba district of Malawi that one-third of families with daughters who had dropped out of schools had done so for reasons of early marriage and 16.7% because of pregnancy. In 8 out of the 10 control groups of parents in the GABLE pilot survey, the most frequently cited reason for female drop out was pregnancy with poverty coming second. It is also clear from this survey of parents attitudes towards the education of girls that fear of pregnancy is a major reason (and a valid one) for their ambivalent attitude to their daughters' education. There is little mention in the literature on Malawi about the role of male teachers in sexual harassment and the problem of HIV/AIDS amongst young people.
Gordon reports that sexual and verbal harassment of girls in Zimbabwe's co-educational secondary schools are common. In a survey of 10 secondary schools, a number of girls interviewed cited teasing, humiliation, verbal bullying and ridicule of girls by boys as a major problem for girls at school. This harassment occurs both within and outside the classroom (often on their way to school). Teachers often collude with male pupils in the verbal harassment of girls in the classroom either directly or by omission (Gordon, 1993). Teachers who were interviewed confirmed the girls' statements about boys harassing them although they thought it best to ignore such behaviour. They also felt that girls should be more assertive and stand up for themselves.
The issue of sexual abuse was seen by female pupils as one of the main problems facing girls at schools (particularly those in rural areas). According to one girl, 'one of the big problems is teachers proposing love to girls' (Ibid: 16). In general, teachers interviewed (both male and female) denied any knowledge of actual abuse of girl pupils occurring at the time of the study. Nearly all teachers said that such cases had occurred in the past and mainly involved untrained or temporary teachers. Much of the blame for sexual abuse of girl pupils by male teachers was placed on the girls, even though many of them are under the legal age of consent (18 years). Similarly, in the cases where girls are sexually abused, the parents also tend to hold the girls responsible. Significantly, the girls thought that there was no point in reporting incidents to teachers (of either sex).
Girls at school face an impossible conflict of roles. One the one hand, they are chided by teachers for not being more assertive but on the other hand they are treated as 'hunters' of men (Gordon, 1994:20). Teachers in this study consider that the main reason for poor performance of girls academically was that they are mainly interested in love, romance and sex, whereas boy pupils do not take such relationships seriously. It would seem that girl pupils are expected to take responsibility for their actions in a way that boy pupils and male teachers are not. Furthermore, girls are expected to take responsibility for boy's sexuality as well as their own (Ann Cotton, CAMFED Conference, 1995).
It is hardly likely that male pupils will refrain from sexual activity in school if their male role models are engaged in the same practices. An examination of the explanations for the dismissal of a random sample of 48 teachers dismissed for misconduct in two months of 1992 in Zimbabwe revealed that 54% were discharged for sexual offenses with students. In 1988 and 1989, 520 and 468 teachers respectively were dismissed for misconduct. This is almost certainly a major underestimate of the number of girls endangered by harassment each year (IBRD, 1992a). Strict codes of practice need to be adopted and enforced in schools around the issue of sexual harassment by teachers.
In short, sexual harassment, under-age sex and teenage pregnancies are major problems in all three countries. The costs to girls of these activities are considerable: not only loss of educational potential but also the health risks connected with pregnancy and HIV/AIDS. In Zimbabwe, many children contribute to the costs of their education by finding casual work. In the present economic conditions, poverty coupled with the prevailing view in Zimbabwe that adolescent girls are aids free, causes many young girls to be pressed by men to exchange sex for cash (CAMFED, 1995). This depressing trend is confirmed by Ministry of Health statistics on AIDS cases. A study in Matabeleland showed that girls' chances of contracting AIDS in the 15-18 age group are 6.5 greater than boys. It is often difficult to persuade men to undergo testing or accept responsibility for transmission of AIDS/HIV. At the same time, there is a trend for men to seek out younger and younger sexual partners (UNICEF, 1994). In Zambia, the picture is similar in terms of high rates of confirmed AIDS cases amongst the 15-19 age group which, between 1985 and 1995 were seven times greater for females than males (Sampa, 1995 and UNICEF, 1995b).
Concern about the rates of teenage pregnancy and HIV/AIDS infection has led to demands for the development of sex education in both primary and secondary school curriculum but this has only recently been developed for schools in Zimbabwe by UNICEF in conjunction with the Ministry of Education and Culture. The MEC is now providing training materials for students and teachers in AIDS education for all levels from primary to Grade 4 of secondary (UNICEF, 1994). However, sex and sexuality are often taboo subjects in these countries which makes education in the schools of central importance in addressing this major threat to the health of girls and women in particular.
4.4.1. Girls and the school environment
4.4.2. Teachers: expectations and attitudes
4.4.3. Single sex schools
As noted earlier, negative attitudes towards the abilities of girls are deeply embedded in all cultures and, as in the colonial era, education for domesticity is the norm. In Malawi, for example, some subservient cultural practices such as kneeling to parents and elders are carried out in the schools by girls only. Subordinate status is impressed upon girls in all three countries and this is reflected in the structure of schools. Gender bias in schools is often found in teaching pedagogy, subject streaming, teachers expectations, instructional materials and curriculum content. Inadequate school facilities in Zambia and Malawi mean that many children have to sit on the floor and classes are very crowded, typically with a pupil to teacher ratio of 60:1. In Zambia, lack of water and adequate bathroom facilities have discouraged girls from coming to school when they are menstruating (IBRDb, 1992). Indeed, many schools totally lack sanitary facilities in both urban and rural areas and affects girls more than boys (Kelly, 1994). Similarly, girls from poor rural areas in Zimbabwe will frequently stay away from school during menstruation because of lack of sanitary protection. They are often too embarrassed to explain their absences to the teachers and incur punishments (CAMFED, 1995).
In Zambia the lack of water and poor physical facilities, large class sizes and multiple sessions militate against a secure learning environment (Kelly, 1994). Girls attending day schools in particular, tend to be burdened with domestic work which affects their commitment to schooling. In Malawi, when students are late for school, they are punished by having to do heavy physical labour. Research also indicates that girls at mixed boarding schools in Malawi often end up doing a greater proportion of chores than their male counterparts. Girls at these schools are tightly controlled and confined to their dormitories in the mornings and evenings whilst boys can use classrooms after hours. It is also reported that female students are subject to verbal and psychological harassment when they show signs of good performance (Hyde, 1994). This situation is also found in the other two countries. In the rural areas of all three countries where untrained teachers are concentrated, pedagogical practices tend to be particularly authoritarian and sexist. The atmosphere in mixed secondary schools, is not supportive for girls.
It is clear from research in Malawi that girls in school lack female role models both inside and outside their communities which means the female teacher is of great potential importance. Table 7 shows the distribution of female teachers in Zimbabwe, Zambia and Mozambique. In 1990, the highest proportion of women primary teachers was in Zambia at 45%, although they were mainly concentrated in Lusaka and the Copperbelt. (The data for secondary teachers are incomplete). The preponderance of female teachers in the urban areas is a result of a policy which guarantees employment to teachers when their spouses are transferred. This policy is currently being reviewed. However it does highlight the point that in all three countries, most teachers, but particularly women, would prefer to work in urban or peri-urban areas.
The proportion of women teachers in Zimbabwean secondary schools dropped between 1980 and 1990 - from 37% to 29% (see Table 7). Most women teachers are found in urban areas in Zambia and Malawi which means that most rural schools have only a small proportion (often less than 20%) of women teachers. In Malawi, only 27% of the teachers in rural areas are female compared with 60% in urban areas. This is mainly due to the fact that women do not want to work in isolated areas where their security and housing is not certain. The proportion of women teachers in maths and science in all three countries is particularly low mainly because so few women with the necessary science and maths background get on to teacher training programmes. In Zambia, a special training course for the maths and science education project trained only 20 women in 1988 out of a total of 200 teacher participants (IBRDb, 1992). The lack of female role models in the so called 'hard' school subjects such as science and maths does not improve girls' interest and attainment in these areas.
Girls have few role models to emulate in the area of education management. For example, in Zambia, out of 200 senior education officers in post in 1992, only 17 were women. For this reason, the World Bank funded Education Rehabilitation Project in Zambia which began in 1992, is attempting to incorporate gender training into all teacher training and education management courses and to address the logistical constraints that reduce female participation in such courses.
Bentry Nkhata from the University of Zambia in 1994 submitted a proposal based on an OU/ILEA 'Girls into maths' pack to run a 'Girls into maths and science study group' in order to raise the achievement of pupils in maths and science through increased teacher awareness of issues in gender and maths and science teaching (Nkhata, 1994). The progress of this course is not known. Special training for teachers of maths and science could be organised as part of INSET programmes.
The advantages of women teachers: International evidence on the impact of women teachers and headteachers on girls and boys is limited. More systematic research is needed which controls for key factors such as size and type of school and teacher and head teacher characteristics. However, positive correlations between the number of female teachers and parity of male/female enrolment have been made using international cross sectional data (Herz, 1991 and Rockerfeller, 1995). And, in socio-cultural settings where male and female contact is strictly limited, as in parts of South Asia, the evidence suggests that recruitment of female teachers does enhance girls' enrolment. On the basis of such information, the World Bank is making a major effort to promote female teacher recruitment by supporting the training of women teachers.
In Zambia, the provinces with a high proportion of women teachers also have high primary school completion rates for girls while the provinces with the lowest proportion of women teachers have the lowest completion rates (Kelly, 1994). In the context of Zambia and Zimbabwe, the positive impact of female headteachers on girls is considered stronger than that of female teachers (Kelly, 1994 and UNICEF, 1994). A recent study of reading advantage found that, in a sub-set of countries (including Singapore, Trinidad and Tobago), girls performed better when their teacher was also female. Similarly, a recent US study revealed that women who teach feel more efficacious in their work when the head is also female (Fuller et al, 1994).
Appleton's study of factors affecting the exam performance of boys and girls at primary schools in Kenya raises an unusual question: the impact of teacher gender on boys. He finds that women teachers had a generally negative influence on the performance of boys but had no effect either way on girls' performance. His explanation is that male students in Kenya - typically aged 14 plus by the end of primary school see themselves as 'young-men' and they may, therefore, be more ready to accept the authority of a male rather than a female teacher. The effect on boys of a woman teacher was significant: on average, a boy taught by men would receive 69 marks more than one taught by a woman. On the basis of this result, Appleton challenged the logic of the Kenya government's current programme to train more women teachers! Clearly, however, more needs to be known about the construction of the survey before one is inclined to accept this conclusion. Designing policy in this area requires an understanding of the relative impact of teacher gender on pupils of both sexes as well as the extent of the gender gap in performance.
The Kenyan study also showed that the gender of the headteacher was significantly related to school performance. Performance in the Kenya Certificate of Primary Education in 9 schools with a female head out of his sample was significantly better than in the remaining 41 (Appleton, 1994).
Teacher attitudes: Although female teachers are clearly important as role models for girls, there is little evidence to suggest that negative attitudes towards the abilities of girls are any less prevalent amongst female teachers than male. For example, in a survey of 12 secondary schools in southern Malawi, Hyde showed that there was little difference between male and female teachers with respect to their attitudes towards female students. 77% of female teachers responded and 80% of males felt that boys were more interested in school work than girls (Hyde, 1994). In fact, more male than female teachers in the study felt that teaching methods for boys and girls should differ. Both evaluated girl students as more passive and referred to the greater questioning of boys. However, an important consideration in the Malawi context which tempers these conclusions is that female teachers are in general less educated than males due to lack of university and other training opportunities when they were young. Consequently, they are probably less confident than their male colleagues. As noted earlier, it is well known that pedagogical practices tend to be particularly authoritarian amongst poorly trained teachers.
The Hyde study also found that male head teacher respondents at mixed secondary schools, commonly referred to the 'laziness' and 'sense of inferiority' among female students (Hyde, 1994:21). She concluded that headteachers (the large majority of whom were male) displayed a general awareness of the educational problems that girls faced in trying to complete their secondary education but varied in their awareness of the role they could play in ameliorating such problems. However, in the four single sex schools that were surveyed, female teachers did indicate that they had introduced innovations to improve the performance of girls. It would appear, therefore, that differences exist in teacher behaviour and practice according to whether the culture of the school is predominantly male or female. Hyde also found that the attitudes of teachers closely reflect those of the parents with regard to girls innate abilities and potential. Girls' concepts of gender roles and related natures, abilities and aptitudes in turn tend to mirror those of their parents and teachers. As elsewhere, the socialisation process has a profound effect on African girls at both home and school.
Gordon's study of teachers and pupils in Zimbabwean secondary schools arrives at similar results and conclusions to those in Malawi. Perhaps her most important finding is that. while there was a general awareness among teaching personnel about poor performance among girls, the causes were not perceived as being the responsibility of the school or the teachers. In short, the stereotyping of gender roles in school by teachers affects their own expectations and treatment of the girls. In her study, there was a remarkable consistency (one assumes among both male and female teachers) in the acceptance of the female role as being primarily domestic and the belief that the man is the breadwinner and provider (over 90% of the teachers interviewed). Many teachers saw it as their duty to prepare students for these roles. There was also a high degree of correlation between occupational aspirations of parents and their daughters. Most parents interviewed wanted their daughters to be nurses, teachers, clerks and secretaries and their sons to be engineers, doctors, drivers and craftsmen (Gordon, 1993). Gendered subject specialisation at secondary school is carried over into gender segregated labour markets.
There is clearly a conflict between good performance in maths and science and the notion of the passive female who specialises in domestic science and homecare. If these roles are to be challenged it is vital that the sex stereotyping of subjects is ended and active attempts are made to encourage the participation of girls in maths, science and non-traditional vocational courses. This can be done in a variety of ways including more flexible timetabling, introducing gender sensitive curricula and teaching materials as well as counselling.
There is widespread evidence from many SSA countries, including the three under scrutiny, that educational outcomes for girls attending single sex schools are better than girls enrolled at mixed schools. The generally negative environment for girls in mixed schools in Zambia. Zimbabwe and Malawi goes a long way to explaining why girls perform better at single sex schools (and single sex boarding schools in particular). Most heads of single sex girls schools are female as are most, although not all, of the staff. These institutions do provide positive role models for girls, and, equally important, an atmosphere that is free from sexual harassment. However, little detailed research is available that explains the processes operating at these schools.
In Malawi, head teachers in all girl schools report placing stronger emphasis on raising 'independence and achievement' of female students (Fuller et al, 1994). As discussed earlier, despite the improvement in performance among girls at single sex schools in Malawi, their pass rates are still 10% lower than boys in all boy schools and 5% lower than boys in mixed schools. Evidence from Malawi and Zimbabwe shows that single sex schools are an advantage for both sexes, but particularly for girls. Nyagura and Riddell's analysis of 48 secondary schools in Zimbabwe pointed to the need for further research to identify which aspects of boarding and single sex education explain the higher levels of academic achievement found in these schools. Even though boarding schools are a very high cost alternative to day schools and therefore cannot become a realistic alternative, it seems important to investigate if boarding schools have certain characteristics which could be replicated in day schools (Nyagura and Riddell, 1991).
One of the main advantages of all girls boarding schools in Zimbabwe is that the pupils are allocated extra study sessions attended by a resident teacher. This extra time for study could be contributing to better performance. Another feature of boarding schools is that children from poor homes receive more nutritious meals than they would at home (CAMFED, 1995). Investigating which factors of advantage associated with single sex schools, particularly with regard to maths and science for girls might help identify strategies that could lessen gender inequalities and improve school effectiveness.
In Zambia, it is well known that girls attending single sex schools perform better than at mixed schools mainly because girls are not socially rewarded for good performance at mixed schools. However, the performance of girls in Zambia at single sex schools only surpasses their performance in regular co-educational schools. Although this effect occurs consistently in each province, in some it is very strong. For instance, in the Central Province, 63% of girls in single sex schools obtained full certificates as against 16.5% of those in co-educational schools. However, the factor of quality must be taken into account as some are mission schools and some boarding schools (Kelly, 1994).
The main vocational training organisation in Zambia received more applicants from girls who had attended single sex schools than mixed for the accountancy course (70% of female applicants are from single sex schools). This would suggest that the standard of maths is higher at these institutions as this subject is a prerequisite for entry to accountancy (ILO, 1990). It might also indicate that girls from single sex schools are more confident and have been encouraged through counselling to apply for training in non-traditional areas. Indeed, the recent education policy paper for Zambia intends to create more boarding places for girls, presumably at single sex schools (ROZ, 1995).
While half of the girls taking MSCE exams in Malawi come from mixed sex schools, 70% of girls who gain admission to Chancellor College come from single sex schools (Kadzamira, 1994 and Hiddleston, 1991). The CAMFED scheme in Zimbabwe is attempting to create a more balanced enrolment between boys and girls in rural secondary schools by the provision of bursaries for girls. 'Girls, already more reticent than boys by the age of entry to secondary school, are more likely to play an active, participative role in the classroom when they form part of a large group' (CAMFED, 1994:4). The experiences of the small number of secondary schools that CAMFED supports in Zimbabwe strongly suggests that transforming mixed sex schools requires a major effort as well as financial and moral support for girls.
It is not entirely clear what it is about single sex girls schools that makes for greater academic success across all subjects. Obviously not all single sex schools work well for girls. Some single sex private secondary schools in Tanzania, for example, are of a very low standard. Quality is a critically important variable in determining whether girls from a single sex school will perform well academically. Hyde found in Malawi that girls in single sex secondary schools tended to be more competitive and presumably less passive than in mixed schools. Also, the female teachers in the four single sex schools in the survey indicated that they had introduced innovations to improve the performance of girls (Hyde, 1994). While some female teachers who were interviewed still held negative attitudes towards girl's abilities, presumably the all-female atmosphere has other positive benefits for the students. The advantages of a protected atmosphere free of male harassment (verbal and sexual) along with female role models are likely to be among the most important. A conference in 1994 held in Malawi on maths education called for a greater number of single sex schools in order to boost girls' performance in maths and science.
There are, however, huge variations in the quality of single sex schools and this should be taken into consideration in assessing their effectiveness. It has been pointed out, for example in Zimbabwe, some of the older single sex schools offer only limited gender defined subject choices that limit girls chances of access to labour markets (Gordon, 1993). It would seem that in others, girls manage to perform much better in maths and science options than in mixed schools. When communities were consulted in Zimbabwe about what kind of schools they wanted for their children after Independence, the majority opinion was in favour of single sex boarding schools (Chung, 1995). The education authorities as part of their massive post-independence expansion campaign preferred instead to construct new mixed day schools because these were seen as being more equitable and cost effective forms of provision.
Official policy in most SSA countries is to encourage co-educational schools, mainly on the grounds of efficiency and cost. Just as in Britain during the 1970s, the received wisdom favours co-education, largely for social reasons. Given this policy preference, gender streaming in mixed schools may well offer a temporary solution to the problems of poor performance and high dropout rates of girls at both primary and secondary schools.
Gender streaming: Three main strategies have been employed to improve girls' performance in Malawi: (1) Tighter control of female students' use of designated study time; (2) some form of separation of the sexes during instruction; and (3) talks and counselling by female teachers. The staff at one boarding school associated girls' poor performance with their poor attendance at homework sessions (prep). Consequently, attendance at prep was made compulsory. The attendance at prep improved although it is too early to assess the impact on performance. Although Kelly in his 'Situation Analysis of Girl Child Education in Zambia' recommends that pilot studies be undertaken to test out streaming by sex, Malawi is the only country (among the three under investigation), where gender streaming within mixed secondary schools has actually been tried. For some time in Malawi, a few schools have experimented with gender streaming within the same school. One example of this separation strategy comes from a school near Zomba which from 1980 until recently initiated an experiment in order to find out whether teaching maths to single sex groups would improve exam results. At JCE level, the pass rate for girls increased from 24% to just over 60% in 1985 and between 1988-1991 the pass rates reached over 80% (Hyde, 1994).
A total of 15 teachers in three schools in the Hyde survey said they had conducted gender related experiments. In one school, the maths teachers (one of whom was an expatriate) were concerned about the significant differences between male and female performance at the JC and MCE levels. They also noticed the negative attitudes girls had towards the subject and the fact that girls who took maths were disparaged by their male classmates. The maths teachers of this school, with the cooperation of the headmaster and Ministry of Education, decided to employ two way streaming by gender and ability, with four groups placed in separate classes (i.e. the top girls and top boys were put in separate classes and the bottom girls and boys had their own classes). Female pass rates at JC went up from just over 20% in 1981 (before the innovation) to over 80% between 1988 and 1991. Boys also benefitted. Their pass rates increased from around 70% to close to 100% in 1988. The MCE results also show a consistently upward trend for both boys and girls. What this streaming experiment achieved was to raise the performance of both boys and girls at JC and MCE and it also narrowed the differential between them at JC level (Hyde, 1994:24-25). An important feature of this experiment was that there had been low staff turnover over the period of the innovation. Also, the striking gains in performance of girls were made at no extra financial cost to the school.
Despite the gains made during these experiments in streaming, they have apparently been recently abandoned because the schools thought it would be a better social experience for boys and girls to mix together during their adolescent period (Kadzamira, 1994). This does seem an extraordinary turnaround which might relate to central government directives on the importance of co-educational schooling. However, the GABLE project is currently financing a pilot project on gender streaming in mathematics classes to test the effectiveness of gender streaming at the primary level. Clearly, streaming of this kind has great advantages in that it can be organised at little or no extra cost (apart from time) to the school. Streaming in subjects like maths and science where girls perform particularly badly would seem a sensible short term option until attitudes towards girls' abilities in the wider society show some signs of change, although its success would seem to depend on a high level of staff commitment and continuity. Single sex schooling is probably not an option for new development due to cost although it does provide a temporary solution to the problems girls are experiencing in mixed schools.
