Research outline: Schooling decisions amongst poor households in Dhaka, Bangladesh (draft) Stuart Cameron November 2007 I. Introduction ........................................................................................................ 2 II. Conceptual framework...................................................................................... 3 III. Literature review .............................................................................................. 9 1. Education in Bangladesh: background .......................................................... 9 2. Costs and expenditure................................................................................. 12 3. Benefits of schooling in the labour market.................................................. 16 4. Other benefits of schooling......................................................................... 20 5. Low-cost private schools in Bangladesh and other south Asian countries.... 21 6. Private tuition............................................................................................. 27 7. Use of English as a medium of study .......................................................... 29 8. Who makes the decisions about schooling? ................................................ 31 IV. Research questions and some propositions ................................................... 32 V. Methodology and methods .............................................................................. 34 1. Methodological approach ........................................................................... 34 2. Methods ..................................................................................................... 39 VI. Contribution to the field ................................................................................. 44 References.............................................................................................................. 45 Appendix: research questions and methods of data collection and analysis....... 52
I. Introduction In the attempt to reach educational goals in developing countries, such as Education For All, the demand for schooling has been recognised as a key component. There is also a large and varied tradition relating demand for schooling to the (expected) costs and benefits associated with it. However, there is little research examining at a micro level how these costs and benefits really impact on families’ decisions about whether (and for how long) to send a child to school, in developing countries. What research there is tends to focus narrowly on the financial costs (school fees and lost wages) and benefits (increased future productivity or wages). This research will examine a wider range of economic, social and cultural costs and benefits of schooling, as they are seen by children and parents themselves, and how these relate to their decisionmaking. It will focus on primary schooling in urban Bangladesh. A key aspect of schooling decisions is likely to be the type of school available and the type (if any) chosen. In Bangladesh a wide range of types – government, religious (madrassah), NGO-run and possibly private schools – exist to serve the poor in urban areas. As part of initial scoping the study will see what types are available in particular areas, given their likely significance for the costs and benefits of schooling.
2
II. Conceptual framework This research will examine the costs and benefits of attending primary school in urban Bangladesh. In weighing up benefits it will consider the general question “What aspects of schooling do people value?”, focussing especially on linkages between primary schooling and children’s lives afterwards. It will relate these to relevant insights from a diverse set of theories including: economic rates of return; different forms of capital; livelihoods in the context of changing macroeconomic circumstances; and capabilities. Household decisions about whether to send a child to school, and how long to keep him or her there, can be broadly said to depend upon the costs and expected benefits of doing so. The costs include the financial costs, the opportunity costs of the child’s time, and perhaps the effort that is required by both parents and the child to stay in school. A starting point for this type of analysis is the returns to education literature, in which income stream from future increased earnings (or increased productivity in non-waged work leading to economic benefits) is compared to the direct and opportunity costs of schooling, and typically found to represent a good investment when studied at a national level (e.g. Psacharapoulos and Patrinos, 2002). The returns can vary, though, depending on each household’s standing in relation to external circumstances such as technological innovation (Foster and Rosenzweig, 1996). Even if returns to schooling in the labour market are high, there are many reasons why this might not translate into effective demand. These include credit constraints, preventing parents from borrowing to pay for a child’s education; and problems of imperfect information and incomplete contracts, which mean that parents cannot be sure of getting back the full return on their investment in a child’s education. In the presence of incomplete information, parents and children make choices on the basis of “mental models” of the benefits and costs of schooling, which they acquire partly through experience and partly through interactions with people around them, and which are thus specific to a particular time, place and socioeconomic context (Srivastava, 2006; North, 1994).
3
In addition, investment in schooling has to be placed within the context of a household’s various livelihood opportunities and, in some cases, attempts to diversify. Schooling is a risky investment, especially when parents have limited information both about the school and about the inherent ability of their child, and the way this risk affects household decisions will also depend on how it compares to the risk and expected returns of other forms of investment.1 In examining the importance of labour market conditions for returns to schooling (and consequently, demand for schooling), it can also be useful to draw on models of the labour market with more than one sector such as the Harris and Todaro (1970) model of rural-urban migration and Fields’s (1974) models of how employers respond towards a surplus of educated labour. The latter paper suggests that for a sufficiently large education system, educated workers have to compete with other educated workers to find skilled jobs; consequently the returns to entering the market for skilled labour decline and demand for schooling lessens. But, given continued growth of the education system, at some point educated workers will start entering the unskilled labour market. If employers in the unskilled market prefer educated to uneducated workers, then uneducated workers are pushed out of this market, and the incentives for education increase again as workers compete to stay in the unskilled market. Another area of economic theory which is likely to be relevant is the literature on screening. The classic models of screening (Stiglitz 1975; Spence 1973) have used the assumption that a time-invariant characteristic of workers, labelled ‘ability,’ is both inversely related to costs of schooling (schooling costs less for more able individuals) and directly related to productivity in the workplace (more able individuals are more productive). The effects of educational screening depend on the model used, but commonly include that “[a]dditional education obtained by individuals of a given ability raises the education needed by the more able if they are
1
Few, if any, studies have considered risk in household education decisions in
developing countries. Weiss (1972), Levhari and Weiss (1974), and Chen (2001) look at risk in decisions concerning higher education in the United States. 4
to signal their talents” (Riley, 1979, p.229). Consequently, the amounts of schooling chosen by different individuals are more spread out relative to the outcome under perfect information (Lang, 1994). Thus factors affecting the household decision process may include risk in labour markets; the possibility of an entirely different decision process for girls than for boys; and the possibility that schooling is valued for providing signals as well as, or instead of, for improving cognitive skills. This study will stick with the assumption that people are rational in making decisions about schooling, and so in some sense assess the costs and benefits, while admitting that these costs and benefits may be much wider in nature than the narrow range usually considered; that a lack of information may add an element of risk to the decision; and that the household may not act as a single utility-maximising unit in making these decisions. The study aims to make some headway through both quantitative and qualitative techniques into understanding and analysing these wider costs and benefits. Amongst these wider costs and benefits, sociological insights such as ways in which schooling acts as a marker of status are also likely to be important; as are insights from the social reproduction and social capital literatures.2 Schooling can also be seen as leading to expansion (or potentially contraction) of the “capabilities” people have to lead the kind of lives they value (Sen, 2000; Robeyn, 2006); and the capabilities approach also ties in with consideration of costs, since these can be considered as the flipside of the freedom (or otherwise) that children have to attend schooling. As well as drawing on Sen’s work and the framework of capabilities and functionings, the study will build on work such as Alkire (2002) in which capabilities are made measurable.3 One possibility of these approaches, along with the signalling
2
Examples in the Indian context include Ray (1988) and Jeffery et al. (2003)
3
However, the reasons for doing this in the present research are somewhat different
than elsewhere. The aim here is to use the capabilities literature to identify potential valued benefits of schooling which might otherwise be overlooked, rather than as a tool for evaluation of programmes and policies – although the results may have policy implications which can be described in terms of capabilities. 5
literature and some of the research emphasising peer effects on learning outcomes, is that they may highlight a propensity for education to perpetuate or increase inequality; some of the benefits that some individuals gain from increased schooling may be at the expense of others. The study will make reference to pedagogic and policy work which provides a normative framework of what schools should be like, but also aims to step back slightly from these and understand what function the school system serves as it is (and for whom it serves this function), rather than automatically seeing it as dysfunctional. The study also aims to look at children’s perceptions of school and aspirations about work. The possibility that children might play an active role in participating in, creating or reproducing institutions, rather than simply being receivers of services, will be explored. Drèze and Kingdon (1999) note that cost-benefit approaches tend to assume that schooling decisions are made by parents, whereas in practice school participation is effectively a joint decision of parents and children, whose interests may not entirely coincide. Similarly Morrow (1996) notes a tendency amongst social scientists to see children as “dependent, passive and nonproductive” and Doss (1996, p. 1605) notes that children are not generally incorporated as decision makers in household economic models: Although children may provide labor and earn an income, the implicit assumption is that children are an investment or a public good, with mothers and fathers having different preferences over the quantity and quality of children and possibly different preferences over the treatment of sons and daughters. Iversen (2002, p. 817) suggests that this failure to perceive children as potential economic agents “may distort analyses of child labor supply, educational attendance and intrahousehold allocations in developing countries.” Exceptions tend to be from participatory research and include the articles collected in Johnson et al (eds., 1998), and Williams (2005), which looks at ways in which participatory research might have
6
influenced policy relating to child poverty, using a case study of working children in India.4 A related and important but under-researched area is the interaction between education decisions for different siblings in the same family. For instance, a family might decide to concentrate its ‘investment’ (of both money and children’s time) on one child – perhaps the oldest boy or the child seen as ‘brightest’ – or to spread educational spending over all of the children in a family, depending upon the returns to different levels of education. A situation where the returns are poor to lower levels of education, or where returns are much better to private than government schooling, would tend to favour concentrating investment on one child – if it is not possible for them all to reach higher levels of education or for them all to go to private school. A relevant branch of household economics here is the literature on “quality versus quantity” of children (e.g. Becker and Lewis, 1973; Hanushek, 1992). The gender of each child is likely to be important in affecting the outcomes of these interacting decisions. However, Srivastava (2006) notes that the way this happens will depend on the institutional context, and warns against seeing this context as fixed and immutable. For instance an increase in the value assigned to women’s education in “marriage markets” could lead to educational expenditures on girls and boys becoming more equal. Whilst economic models of the household (e.g. the individual maximising model of Becker, 1964, and the ‘family’ model of Behrman, Pollak and Taubman, 1982 ) provide some insights here, they are liable to overlook intricacies of the process which are of interest in themselves. Drèze and Kingdon (1999, p. 5) argue, in particular, that “children are unlikely to be motivated by cost-benefit calculations” – although this might be better stated as “children may not be motivated by the same range of costs and benefits as their parents”. A model of household decision making Each family member has a mental model of the benefits and costs of a child’s education5 and how these benefits and costs are distributed amongst the child and
4
See also Williams (2004); Witter and Bukoke (2004) 7
other family members (see diagram 1, at the end of this document). They each have preferences over these distributions and these preferences are combined according to some bargaining or collective process to make the decision. Diagram 2 illustrates how each household member models the benefits and costs, both to the child and to other household members, of the child going to school. As this diagram shows, agents are likely to have uncertainty in several areas, but particularly in predicting future labour market conditions and in understanding how an increase in education for the child in question will impact on the rest of the household. Characteristics of the agent may affect the level of uncertainty, for instance those who currently have connections with the labour market may have a better understanding of trends in it. Agents may also differ in their preferences over the risk implied by this uncertainty, i.e. may be more or less risk averse. Diagram 3 attempts to capture the continuous nature of the decision-making process by charting the decisions that will have to be made at different ages and factors feeding into these decisions. The correspondence between ages and particular decisions will be vague because of grade repetition; delayed entrance to school; children dropping out of school but later entering an NGO school or madrassah; etc. ‘Costs and benefits’ in diagram 3 is a shorthand for all of what goes on in the mental model in diagram 2. These calculations are likely to be done on an implicit level: the household members themselves may not see it as a process of calculation, especially in cases where the outcome is seen as obvious (e.g. where the family ‘cannot afford’ to send a child to secondary school). They are likely to be affected by variables including the child’s gender and perceived abilities, and the family’s socioeconomic status. The cost-benefit calculation at each stage also depends on expectations about what factors will influence future decisions; for instance the calculation of costs and benefits at primary level might depend on the child’s
5
In practice this means decisions about: whether to enrol the child in school initially,
and at what age and what type of school; whether at a particular time the child should drop out or continue; and whether the child should take private tuition, and how much. 8
perceived likelihood of passing the secondary schooling entrance exam (and this likelihood is endogenous in the sense that it is itself influenced by current decisions about schooling and tuition). Obviously, some of the decisions build on earlier decisions, e.g. the decision to stop schooling or take secondary entrance exams around age 11 depends on having attended primary school in the first place. Importantly, a household with more than one child will be making such decisions simultaneously for each child and the decisions for one child are likely to affect the decisions for another child. Simplifications of this deliberately very broad model, and possible propositions stemming from it, are considered in section IV below.
III. Literature review 1. Education in Bangladesh: background Studies in Bangladesh, as in other South Asian countries, have suggested that as enrolments have risen without a corresponding increase in the number of schools, the number of students per school has risen and quality has declined (World Bank, 2002, cited in Rahman, 2005); Nath and Chowdhury (2002) suggest this has particularly been the case in urban areas due partly to large increases in urban populations. Gross enrolment in primary schooling stood at 97 per cent (and was the same for both boys and girls) in 2000, but the the completion rate was only 67 per cent – although this seems to have risen rapidly during the 1990s (Ahmed, Ahmed, Khan and Ahmed, 2007). Rahman (2005) finds that, in urban areas in 1999/2000, over one-quarter of the male labour force and 45 per cent of the female labour force had no education. As well as being eliminated at primary level, gender disparities in gross enrolment ratios have been reduced at secondary level (Kabeer and Mahmud, 2004) – raising important questions about the extent to which greater equity in the education system is starting to lead to real advantages for girls and women after leaving school. Major policy initiatives in recent years include: – the 1990 Primary Education (Compulsory) Act, which has led over the years to policies such as the elimination of school fees, provision of textbooks free of
9
charge, and incentives to encourage the participation of vulnerable children (Tietjen, 2003) – the 1993 Food for Education (FFE) programme, through which 40 per cent of children in targeted poor areas received a monthly allocation of wheat or rice for their family if they attended primary school regularly. Ultimately this programme covered 27 per cent of the country. (Tietjen, 2003) – the Primary Education Stipend Project, initiated in 2001, through which eligible pupils received 25 taka per month – In 2003 FFE and the primary stipend project were replaced by a single cash-based stipend programme for the whole country, the Primary Education Stipend Programme (PESP). This targeted children in rural areas with a stipend of 100 taka per month. The programme was broadened to include urban areas in 2004. The programme applies to Pupils enrolled in government primary schools, registered nongovernment primary schools, community schools, satellite schools, complete (5 grades) NGO-run primary schools, and GOB-recognized Ebtedayee Madrassah; non-formal schools such as the BRAC schools, and other unregistered or unapproved schools may not participate. To be eligible children have to meet minimum criteria for attendance (85 per cent) and achievement (marks of 45 per cent). (Tietjen, 2003) – a programme giving stipends to girls who enter secondary school, which was rolled out nationally in 1994 (Raynor, 2005); – the 2000 National Education Policy; – implementation of the Primary Education Development Programme – which attempts to move beyond access issues to focus on quality; and – preparation of the National Plan of Action on Education for All (Unterhalter, Ross and Alam, 2003). The public education budget is less than 3 per cent of gross national income (M. Ahmed et al, 2007, p. 73; International Herald Tribune, 2007) – a low figure compared to other countries in the region and developing countries generally; for 10
instance India spends around 4 per cent of its gross domestic product on education (and has much higher per capita income) and Thailand spends around 5 per cent (Mehrotra, 2006). To some extent this reflects Bangladesh having a relatively small public sector, rather than specific neglect of the education sub-sector (M. Ahmed et al, 2007). Sherman and Poirier (2007) explain the structure of the school system in Bangladesh: Compulsory education extends from age 6 to 11 and covers the primary grades. Secondary education is divided into three cycles. Students who pass an examination are eligible to attend the three-year junior secondary cycle including Grades 6 to 8. On completion of this cycle, students qualify for an additional two years of secondary education and may select an academic or vocational track. After completing Grade 10, students take the Secondary School Certificate examination (SSC) to qualify for higher secondary education. (pp. 118-9). According to official statistics reported in M. Ahmed et al (2007, p. 5), government primary schools account for 58 per cent of total enrolments at primary level. 24 per cent of students are in (registered and unregistered) non-government primary schools; around 6 per cent in schools attached to high madrassas and 4 per cent in “ibtidayee madrassas”; 7 per cent are in other types of school6. However these authors note that: Official primary education statistics do not include over 30,000 one-room, one-teacher schools run by NGOs, serving more than a million children.
6
Nath and Chowdhury (2002, p. 78) report somewhat different figures dating from
1999: over two thirds of total enrolments in state-owned schools; around 9 per cent in NGO-operated non-formal schools; and 15 per cent in non-government primary schools. It is unclear whether this discrepancy represents a shift in recent years towards government and non-government schools, and away from NGO-operated schools; or is just an artefact of different ways of counting. 11
Exclusion of these numbers introduces distortion in officially reported gross and net enrollment ratios.” (p. xviii). Given some similarities between the cases of Bangladesh, India and Pakistan, it might be speculated that official statistics could also overlook substantial numbers of lowcost private schools. Section 5 below considers this possibility. 2. Costs and expenditure Table 1 below, from Rahman (2005), shows education expenditures in urban Bangladesh. For comparison, household income levels in slum areas in Dhaka are on average around 3500 taka (US$ 50) per month, according to a survey conducted in 2005 (MEASURE / NIPORT, 2006, p. 48). Table 1. Monthly expenditure (in taka) by sex, type of expenditure, in urban areas, 1996. From BBS: Household Expenditure Survey, 1995-96 (Dhaka, GoB, 1998), cited in Rahman (2005, p. 20) Type of expenditure Girls Boys Tuition fees
8
11.8
Books, papers, etc.
14.4
21.0
Private tuition fees
11.2
18.9
Hostel boarding charges
1.3
3.2
Other charges
2.1
1.2
Total
37.0
55.9
Note: exchange rate in 2005 was approx. US$1 = 64 taka
12
Table 2. Annual expenditure for class 5 students in 2004/2005. From FMRP (2006, p. 82) Type of expenditure Government Registered Attached non-
ebtedayee
government
madrasahs
43
37
87
Stationery
337
282
278
Textbooks
166
116
147
School clothes
238
211
58
Tiffin
317
265
11
Private tuition
549
337
362
Other indirect costs
89
20
7
Total
1746
1266
1053
Direct costs (including exam, sports, tuition fees)
Note: exchange rate in 2005 was approx. US$1 = 64 taka
Table 2 shows more recent data for all of Bangladesh. The average figures shown hide considerable variation between households within each type of school. Although private tuition formed the single largest component of mean expenditure, it was only paid by 44 per cent of household in government schools (and a smaller proportion in the other types of school studied). Among these households that paid any money on private tuition the average amount paid was over 1000 taka. Transport costs were only counted by a small proportion of households (12 per cent for those with children in government schools) but for those that did have to pay for transport, this formed a large expenditure item. Chart 1 shows how these expenditures varied by socioeconomic status (as indicated by the household’s total expenditure quintile) and gender, according to a representative sample from 2005. Expenditure was generally more than 50 per cent higher for those in the middle expenditure quintile to the poorest (bottom quintile), except for female students in registered non-government schools, for whom 13
expenditure did not vary as much by household income. Expenditure for male students was on average 38 per cent higher than female students in government schools and 10 per cent higher in registered non-government schools, although this varied greatly according to how well-off the household was. For the poorest households sending children to registered non-government schools, expenditure was actually 25 per cent higher for female than male students. Focusing on government schools, expenditure was much higher for males than females in the second-poorest quintile but only slightly higher in the poorest and middle quintiles. These results are suggestive of a complicated pattern in terms of which educational options are available to households of different socioeconomic status, and of how these options are valued for girls and boys. Chart 1. Annual expenditure for class 5 students in 2005, for government (GPS) and registered non-government (RNGPS) schools. Based on FMRP(2006, p. 83)
Education expenditure (taka)
3000 2500 Male (GPS) 2000 1500
Female (GPS)
1000
Male (RNGPS)
500
Female (RNGPS)
0 Q1
Q2
Q3
Q4
Q5
Household expenditure quintile
Maitra’s (2003) study of households in rural Bangladesh finds that children in households with higher income and those with fewer siblings tended to have higher educational attainment, and suggests that this supports the notion of resource constraints within the household. An increase in permanent income contributed more to the educational attainment of boys than that of girls – suggesting that boys’ 14
education is seen as a better investment than girls’. Both mothers’ and fathers’ education had an effect on the attainment of both boys and girls. Sources such as Delap (2000) suggest that child work, even at young ages, is common in urban Bangladesh, meaning that the opportunity cost of attending school is likely to be high. Delap (2000) finds that for both male and female children, participation in income generating work increases with age, with boys participating in income generating work from an earlier age than girls, while girls were more likely to be engaged in housework. Amongst her sample of ten households in a slum in Dhaka, all of the boys aged 13-15 were in income-generating work, while the girls of the same age were involved in a mixture of household and income-generating work. While the direct financial value of children’s work, such as firewood gathering, may be low in itself, Delap notes some of the social and cultural reasons it is likely to continue: “many bustee residents felt that the insults and suspicion generated by adult firewood collection would mean that adult participation in such activities would act as a barrier to network formation. Such networks are important for access to resources including loans and employment information…” (pp. 731-2). Chowdhury, Nath, Chowdhury, and Ahmed (2001) report that though primary education is theoretically free, 90 per cent of parents reported incurring expenditures of some sort. On average parents spent 1000 taka per child in school, equivalent to about two per cent of average household income. These authors report no difference between girls and boys, but wide variation between different school types, with expenditure highest for children attending primary schools attached to secondary schools, and lowest for those attending non-formal education. The largest expenditure item, accounting for over a third of the expenditure, was stationery, while private tuition accounted for 25 per cent (see below). As of 2003, the primary education stipend programme (PESP) allotted 1200 taka per year to families with one child in a registered school and 1500 taka per year to families with more than one child in a registered school (Tietjen, 2003) . Tietjen notes that this figure would cover the direct costs of schooling for a single child in a registered school in a rural area, but would cover less than half of these costs in urban areas. Tietjen suggests that it is unlikely that parents will enrol more than one child in a school covered by PESP if cheaper NFE schools are available. 15
3. Benefits of schooling in the labour market There is a broad literature on the benefits of schooling, but in the economics literature the main benefit studied is the expected returns in future productivity or earnings. Rates of return to primary education are conventionally found to be very high, when compared to physical capital investment, or to secondary or tertiary education. Psacharapoulos and Patrinos (2002) estimate the average private return to primary schooling in low income countries at 26 per cent. However there are also doubts about whether the returns really are so high. Bennell (1996) provides a critique of the methods used in the conventional returns to education literature, pointing out the problems caused by omitting variables such as unemployment amongst the relevant youth group and the failure to account properly for the costs of education. A further criticism focuses on the frequent use of outdated figures in such studies. Bennell and others have argued that the returns to primary schooling are likely to have dropped since the 1970s as individuals with a primary level education or higher become less scarce. The return structure of different levels of schooling is also likely to depend upon what types of job opportunities are available. Foster and Rosenzweig (1996) suggest that the returns to education depend upon the pace of technological change, and so can be low in areas where new technologies are not being taken up. Primary education is usually considered to be the most important level for agricultural productivity, whereas post-primary education may be more important for nonagricultural wage employment (Appleton and Teal, 1998)7. Consequently countries with increasing industrial or service sectors and shrinking agricultural sectors could find their returns to primary schooling dropping, as more advanced skills become relevant8.
7
Blunch and Verner (1999) note that entry into these different sectors, as well as
wages within them, can be affected by level of schooling. 8
Of course, primary schooling remains a precondition for obtaining the increasing
benefits of higher levels of education. The extent to which this is reflected in household decisions will depend on the perceived likelihood of accessing higher 16
In India, Duraisamy (2000) finds that returns to primary education indeed appeared to drop during the 1980s and 1990s, although the results of Sharma, Kumar and Meher (2002) suggest that literate individuals still enjoy a significant labour market advantage over illiterate individuals in that country. In Bangladesh, Asadullah (2006), using data from 1999/2000, reports that returns to a year of education are around 7 per cent. However, when the assumption that returns are linear across all types of education is relaxed, the resulting rate of return estimates were lower – around 4 per cent – for primary and secondary school, but around 13 per cent for higher education. Returns were higher in urban than in rural areas, and higher for females than for males, perhaps reflecting the relatively low average level of education amongst female workers and the gender divide in the workplace. Shafiq (2007) argues that rates of return analysis should take into account the fact that one level of education gives one the possibility of proceeding to the next level (where the per-year returns may be higher). Incorporating this “option value” of schooling into the model, Shafiq estimates higher rates of return: 14 per cent for primary, 8 per cent for junior-secondary, and 13 per cent for higher-secondary education. Overall, these results would lead one to expect that labour market returns provide a substantial, if not large, incentive for schooling, although this has to be qualified by the evidence suggesting that access to certain parts of the labour market are limited for slum dwellers and dependent on social and financial resources they may lack. Although, nationally, rates of return are higher for women, women may have more difficulty in realising the full returns because they face these labour-market restrictions in a more acute form. Similarly, opportunity costs are also subject to variation between areas and social characteristics, contributing to further uncertainty about what the incentives to schooling will be for a particular household. As well as average cost levels of schooling, the way these costs fit in with the household’s livelihoods strategies have to be taken into account. For instance M.
levels after primary has been completed. Shafiq (2007) examines this “option value” of lower levels of schooling in Bangladesh. 17
Ahmed et al. (2007) report that “poorer parents who sent their children to school often fell into seasonal economic difficulties. They then could not meet different school expense like examination fees and cost of school dress or copybooks…” (p. 38). This study’s focus group and interview data also suggest other factors likely to be seen as costs, or disbenefits of schooling, such as children being verbally and physically abused by teachers. Rahman (2005) explains that although women are generally disadvantaged in labour markets, they offer the attraction to employers that they will usually work for lower wages, and in addition: … there are certain non-wage advantages to employing women, which urban employers in particular regard as important … Female employees: (a) are docile, thus labour management is easier; (b) do not belong to unions; (c) are usually secondary earners and so making them redundant may be easier; (d) have little bargaining power in the labour market, not only because of the large excess supply of female labour but also because of the fear that losing a job will lead to domestic problems (in the form of ill treatment by the husband, children’s frustration, and so on). (Rahman, 2005, p. 34) Thus an important question with regard to education of girls will be the extent to which it interacts with some of these employer perceptions; for instance, is an educated woman seen as more or less docile than an uneducated one? Opel (2000) looks at the role of social resources in the labour market for residents of slums in Dhaka. Although continuous migration from rural to urban areas suggests that an expanding labour market must be attracting these migrants, the author finds that “people are in fierce competition for employment opportunities in a state of scarce resources”; and “[r]apid urban growth has taken place without a commensurate increase in industrialization” (p. 737). The paper suggests that social capital – in the form (for example) of information accessed through social connections or the ability to provide references – can dominate financial or human capital in allowing access to these scarce jobs. Women are particularly hampered by the labour market’s close relationship with social networks, because their mobility is socially restrained and their housekeeping roles leave little time for network building. Opel reports that, 18
while education has limited implications for entry into the labour market, it plays a crucial role in the progression to higher positions within a particular industry, and sometimes in the transition from the informal to the formal sector. Several authors have also attested to the importance of financial capital in getting a job. For instance one of Sweetser’s (1999) respondents said that it was pointless sending girls to secondary school because they would still have to bribe someone to get a job; and Opel (2000) gives an example of a payment being required by an employees union to make a temporary position permanent. Financial and social capital often interact, as in Opel’s (2000) example of apprenticeships in garment factories: often, skills can be learned from relations or neighbours to help new employees learn skills in the factory; but “for those who lack such relations, it is usual for a certain percentage of salary to be surrendered to the supervising operators in exchange for help in gaining skills” (p. 745). Likewise, financial and human capital interact, as evident for instance in the view of Hossain’s (2005) respondent that “You need a certificate, without that, you don’t get a [formal sector] job. Or if you do, you have to give a bribe, but with a certificate the bribe is less.” (p. 19) According to Salway et al. (1998, cited in Opel, 2000, p. 740), predominant male occupations in Dhaka slums were rickshaw puller (37 per cent), other transport labourer (14 per cent), and manual labourer (14 per cent). Predominant female occupations were housemaid/domestic servant (30 per cent); garment factory worker (27 per cent); and manual labourer (13 per cent). A smaller but still substantial proportion were salaried service workers (8 per cent of men and 7 per cent of women) – an occupational category in which education is perhaps more likely to benefit them. While only 4 per cent of men were involved in the garment industry, Opel’s account suggests that these men (unlike the women working in the same sector) were able to move into higher positions relatively quickly, and that education played an important role in easing this mobility. Overall then, it seems likely that education is an important factor in labour market returns for people living in slums in Dhaka, and that the impact of education is not entirely outweighed by the scarcity of jobs in which education would improve productivity, or by the importance of social and financial capital.
19
4. Other benefits of schooling Hossain (2005) argues that material benefits of education via the labour market are outweighed by the social benefits, and by a non-specific idea of schooling for a “beautiful life” (an idea reminiscent, in terms of capabilities perspectives, of education being seen as inherently valuable, rather than as a means to an end): The material pay-off in terms of job prospects does not seem to be the most important motivation, mainly because so few people ever gain access to the formal sector jobs for which schooling is supposed to equip them. … More commonly, children anticipate schooling will provide them with nonspecifically 'better' prospects in the future … An educated child is a valued member of society wheras nobody 'gives value' to a child who cannot read or write; they may even say bad things. … The clearest purpose of education is stated by children to be the learning of appropriate social behaviour and norms the acquisition of modern, polite manners. Children view the lack of education as a source of social exclusion, blocking their membership in general society. (p. 19) Maddox (2005) notes that, in a rural area in the north-west of Bangladesh, literacy practices were often conducted in a secretive way, especially by women. When asked about the social uses of literacy, when men were standing nearby, female respondents “said that they wanted to learn literacy for activities such as ‘reading seed and fertiliser packets’, ‘helping their children’ and ‘not being cheated’” (p. 127). It later emerged that women wanted to learn literacy and numeracy practices for household budgeting, but that the legitimacy of this type of motivation was “fiercely contested” by others in the community (p. 127). These findings emphasise that education is often the subject of contestation and negotiation within a community; it may be valued by some members of the community in ways that are not seen as legitimate by others. Raynor (2005) focuses on the perceived benefits of educating girls, in the context of the country’s female stipend programme for secondary schooling, which became nationwide in 1994. Perceived benefits in this and earlier studies (Das Gupta, Islam and Siddiq, 1993; Sarker et al., 1995, cited in Raynor, 2005; Sweetser, 1999, cited in Raynor, 2005) included getting jobs; educating her own children when she 20
becomes a mother; getting a better husband; managing the home economy better; . Potential costs include that schooling could endanger girls’ morality and reputation. Some respondents in these studies were unsure whether education really improved the chances of a girl getting a job, and it was also unclear whether schooling raised or lowered the level of dowry that a girls’ parents could expect to pay when she married. Raynor interviewed 41 people including six girls and six boys attending secondary school, and reports: Both boys and girls tended to repeat the accepted view that girls’ education makes them better wives and mothers, and benefits society as a whole. None spoke of individual benefits to the girls themselves. Only the mothers envisioned better things for their daughters. (Raynor, 2005, pp. 93-94) In some circumstances education seemed to raise the level of dowry whilst in others it decreased it, and one mother noted that it might even be possible for an educated daughter to avoid marriage altogether, “because a girl who’s educated can stand on her own two feet and look after herself” (p. 95). Most interviewees “linked girls’ education to employment, but for men/boys the stated reason was almost exclusively financial, whereas women/girls linked employment to such things as ‘independence’, ‘confidence’, and ‘worth’.” (p. 95) 5. Low-cost private schools in Bangladesh and other south Asian countries The apparent rise of “private schools for the poor” in developing countries has been documented by a number of authors. For instance a Unicef survey carried out in 1999 revealed that in urban areas of six northern Indian states, private school enrolments accounted for between 17 and 34 per cent of the total (Mehrotra, 2006); in Hyderabad, Andra Pradesh, Smith, Hardman and Tooley (2005) report that 61 per cent of students were enrolled in private school; and Alderman, Orazem and Paterno (2001) find that 58 per cent of children in Lahore, Pakistan, attended private schools. In other parts of India and Pakistan private schools seem less significant (e.g. Muralidharan and Kremer, 2006), and they are generally less prevalent in rural than urban areas, although even in some rural areas there is evidence of high private school enrolments – for instance, Harlech-Jones, Baig, Sajid and ur-Rahman (2005) report that even in the “isolated” Northern Areas of Pakistan, 28 per cent of all students in school went to private schools. This sector appears to be a new phenomenon: for instance 21
Muralidharan and Kremer (2006) report that nearly 50 per cent of the private schools in their sample had been founded fewer years before the study, although this could also reflect high turnover of schools. The extent to which private schools are really serving the poor remains open to question. Muralidharan and Kremer (2006) find that children attending private schools in rural India come from more advantaged family backgrounds than those in government schools, although this aggregate picture does not rule out the existence of a particular sub-section of the private sector that does serve the poor. Alderman et al (2001) and Andrabi, Das and Khwaja (2002, 2005) find that even amongst the poorest families in Lahore and rural Pakistan, respectively, many sent children to private schools: for instance more than half of children from families in Lahore earning less than $1 per person per day, attended a private school. In Bangladesh there is limited evidence about whether a “low-cost” private school sector exists, and if so how large it is. Imam (2005) reports that the private sector is significant at all levels and constitutes the overwhelming majority of secondary schools. This author emphasises the role of English instruction as part of the rise of private schools; this report notes that, as well as an “elite core of English medium schools,” there are: less exclusive schools charging a lower level of tuition fee and providing on the whole a lower quality English medium instruction. These schools, growing in number, are designed to capture the market among non-elite families for English language education. Teachers at these schools are neither required to have completed English medium schooling, nor to have taken English at tertiary level... (p. 478) Ranked below these lower-fee English medium schools, according to Imam, are the public sector Bangla medium schools, with the madrassahs being perceived as even less prestigious. Official data (cited in M. Ahmed et al, 2007) show that 69 per cent of schools and 80 per cent of enrolments are accounted for by government or registered nongovernment primary schools (RNGPS, which are partly government funded). A further 18 per cent of schools and 11 per cent of enrolments come from ibtidayee 22
madrassas and primary schools attached to high madrassas. Less than 2 per cent of pupils attend non-registered non-government primary schools. However, it is possible that, as has been reported in India, low-fee private schools evade official recognition altogether, perhaps because they choose not to declare their existence to authorities for fear of burdensome bureacratic intervention. The factors underlying the rise of low-cost private primary schooling in India and Pakistan appear to have included: declining quality of government schools following a large increase in enrolments which has not been matched by an increase in government funding; continuing gaps in availability of government schools; and a supply of educated young people who are unable to find work elsewhere, and so willing to work as (formally unqualified) teachers for low wages (Muralidharan and Kremer, 2006; Andrabi et al., 2005). To this we might add that a private sector is, on the face of it, more likely to rise in urban areas, where high population density means increases the viability of multiple providers in a single area, perhaps competing with each other or catering to different types of demand. Arguably weak government provision is a factor also pertaining in Bangladesh. It is less clear whether potential private schools would be able to find a ready supply of educated but unemployed young people willing to work for low wages as teachers. The focus of this study being on urban areas makes it more likely, though, that private provision will be significant. On the other hand, the longstanding existence in Bangladesh of a large number of NGO schools and non-formal education centres, as well as madrassahs, suggest that these may have stepped in to fill the gaps in government provision, leaving less room for a for-profit private sector. The recent primary education stipend programme, which applies to both government schools and registered non-government schools, but not private schools, would also provide a disincentive to families choosing private schooling and make the emergence of a private sector less viable. A brief survey of the evidence from India and Pakistan gives some insights into what the likely consequences are for a household’s perceptions of costs and benefits of schooling and the decision about what type, and how much, schooling to invest in. Indicators of quality in terms of educational “inputs” that might influence the perceived benefits of schooling include facilities, buildings, infrastructure, number 23
of teachers, teacher qualification, how often the teacher is present and how much time they spend on actual teaching activity, the number of working days of the school, and the style of teaching. There is a mixed picture of whether private schools provide better facilities, buildings and infrastructure than public schools (see e.g. Mehrotra and Panchamukhi, 2006; Tooley and Dixon, 2007; Muralidharan and Kremer, 2006). Private school teachers in India were more likely to have degrees but much less likely to have undergone teacher training. Government schools in rural areas often had only one teacher, but this was much less common amongst private schools in Mehrotra and Panchamukhi’s (2006) study areas. Teacher absenteeism seems generally, though not always, to be higher in government than private schools, perhaps reflecting teachers’ security of tenure in the former (Muralidharan and Kremer, 2006; Tooley and Dixon, 2007). Combining the effects of pupil-teacher ratio and time spent in teaching activity, Mehrotra and Panchamukhi find that children in private schools had three to four more times “teacher-contact” time than in government schools. A school’s reputation for quality – or for providing valued benefits – may also be influenced by outcomes such as examination results or the advances in cognitive abilities or knowledge of children who attend. Various a priori reasons have been given to expect that private schools would offer a higher quality of education: that they are accountable to parents who pay fees; that they may compete which each other to provide better service; and that decentralised management is conducive to greater efficiency (Kingdon, 1996). The evidence described above about differences between government and private schools in teacher numbers and teacher activity would also lead one to expect better outcomes. Against these it can be argued that parents who are educated to a low level themselves may have difficulty knowing what makes a good school. Some authors, drawing on international evidence mainly from developed countries, have also argued that introducing choice and competition in schooling has not led to innovations in the classroom; instead schools often embrace traditional practices, responding to a “relatively monolithic conception of what constitutes good schooling” amongst parents (Lubienski, 2006). According to Kingdon (2005, p. 13), small-scale Indian studies “share the common conclusion that private school students outperform their public school counterparts even after controlling for the schools’ student intakes”, although 24
Kingdon (1996) is apparently the only study for India that attempts to control for potential endogenous student selection. Muralidharan and Kremer’s (2006) largerscale representative study of rural India finds that test performance was significantly higher in private schools than public schools; when family and other characteristics (but not peer group effects) were controlled for, the difference is smaller but remains significant. However, a review of some of the earlier studies on public and private schools in India and other developing countries (Bashir, 1997) argues that whilst studies using single-level models seem to show that private schools are more effective, studies using hierarchical or multi-level models do not show a clearly positive effect in favour of the private sector. In Bashir’s own study in Tamil Nadu, the inclusion of peer group characteristics and certain school variables reduced, if not eliminated, the private school advantage; as far as I am aware no more recent studies use multi-level modelling to compare public and private school outcomes in south Asia. Thus peer group effects emerge as a key issue. From an individual family’s point of view, it is not important whether peer group effects or the school itself are the source of better educational outcomes; there is a potential conflict between individual interests and the interests of the whole society or community, which is liable to manifest itself in worse outcomes for the poorest who cannot afford private schools. There are few studies that have attempted objective comparisons of teaching styles in government and private schools. Although the greater likelihood of government school teachers being trained might lead one to expect that outcomes would be better – for instance that more active, pupil-centred pedagogic styles would be used, since these are taught in teacher training colleges – Clarke (2003) argues that teacher training has little impact on actual pedagogic practices in state schools. Alexander (2000, cited in Smith et al., 2005) finds that teaching in Indian government schools is highly ritualised, teacher-dominated, and makes heavy use of rote learning and memorisation, and Smith et al. (2005) draw similar conclusions for private schools serving the poor in Hyderabad, Andhra Pradesh. Pedagogy was dominated by teacher-led recitation, little attention was given to securing understanding, and ritual knowledge was an explicit focus of the learning tasks which teachers presented. Thus it may be that there is actually little variation between teaching styles in the two types 25
of school; one can hypothesise that the norms observed by Clarke (2003) as governing teaching behaviour in government schools, are sufficiently strong and widespread that they also encompass private schools. Which of these aspects of private schools were actually valued by parents? The PROBE Report (1999) suggests that the aspects of private schools that led parents to prefer them to government schools included lower teacher absenteeism; better discipline; and use of English. Muralidharan and Kremer (2006) find that in rural India, major attractions for parents of private schools were that they started teaching English early, and that there was more teaching activity. There is a notable lack of evidence on parents’ perceptions of style of teaching, textbooks, materials, etc. or on whether these were factors in parents’ decisions to send a child to a private school. On the costs side, it is widely documented that although government education is supposed to be available for free, households in fact spend sizeable amounts of money on it, in addition to possible opportunity costs. But direct costs appear to be much higher for private schools. For instance, Tilak (2002) finds that expenditure on children in private unaided schools in India was more than three times higher than on children in government schools; Panchamukhi (2006) finds a difference of between 1.4 and 3.7 times. However, it may be important to look at the disribution of these fees, since mean averages for the private (unaided) sector could be biased by a handful of elite schools in the sample. Comparing costs to income levels, Tooley and Dixon’s (2003) study on Hyderabad suggests that fees at low-cost private schools represented 7 to 10 per cent of father’s annual income; Andrabi et al (2005) find private school fees in rural Pakistan of around 1000 rupees per year, equivalent to 4 per cent of the country’s per capita GDP. It may be that the availability of private schooling interacts in interesting ways with factors such as gender and the number of siblings a child has to affect the amounts a family spends on each child’s education. Pratham (2007) reports that for the whole of India, a slightly higher proportion of boys than of girls were enrolled in private schools, while the reverse was the case for government schools. Evidence on this from smaller-scale studies focusing on low-cost private schools is mixed (e.g. Tooley and Dixon, 2007; Srivastava, 2006). Tilak (2002) reports that household 26
education expenditures do not differ much by gender for children in government schools, but gender differences are sizeable in private schools. This suggests that when girls are enrolled in private schools, they may be more likely to be towards the low-cost end of the spectrum. However, Srivastava (2006) found that families that sent at least one child to low-cost private schools were no more likely to send boys than girls. 6. Private tuition Discussion of public and private schools is somewhat incomplete unless consideration is also given to education outside of school hours. In particular, it may be either required, or highly advantageous, for students to take private tuition (Bray, 1999; Nambissan, 2003). Bray (2003) notes that in the worst cases internationally, private tuition involves “a form of blackmail in which the teachers teach only half the curriculum during the school day and then require their pupils to pay for the other half during private lessons” (p. 13). Muralidharan and Kremer (2006) find that in rural India, pupils in private schools were more likely also to take private tuition than those in government schools. Tilak (2002) reports that in India, private coaching accounts for 5 per cent of household expenditure on elementary education in government schools, and 7 per cent in private schools. These findings may reflect either characteristics of the schools, or the fact that those in private schools were from more advantaged family backgrounds. One plausible scenario, that parents might use private tuition to complement poorquality government schooling, is not supported by these figures, although it is not ruled out either. As discussed in section 2 above, private tuition forms the single largest part of household expenditure on education in Bangladesh, amounting to between 27 and 34 per cent of the total according to a representative survey (FMRP, 2006, p. 85). According to that study, the proportions of households paying for private tuition, and the amounts spent, increased according to household income. For children in government schools receiving private tuition, the amount spent on boys was nearly 30 per cent higher than that spent on girls. By contrast for children in registered nongovernment schools receiving private tuition, the amount spent on girls was somewhat higher than that spent on boys, although girls were less likely to receive any private 27
tuition in the first place. The study reports that for those households who do pay for private tuition, it accounts for a “fairly small proportion of total expenditure” (p. 85), although this leaves open the question of whether some households in lower income groups are spending a high proportion of their income because private tuition is seen as highly important. Three per cent of students in this study reported paying for private tuition from a friend or relative. This raises the question of how paid private tuition and education decisions generally relate to the various other forms of learning, paid and unpaid, that go on outside of school, and which depend on the human and social, as well as financial, capital of the household. Educated relatives or friends may in some cases be providing a highly valued service for free, increasing the returns on schooling and/or decreasing the potential costs of private tuition. Imam (2005) reports that private financing of secondary education in Bangladesh “is high but much of it is directed into private educational goods such as individual tutoring, rather than the resourcing of schools themselves.” (p. 475). Chowdhury et al. (2001) report that 25 per cent of primary school pupils pay for private tutoring. Tietjen (2003) comments that private tuition acts “both as means of compensating for poor quality instruction in school and of augmenting teachers’ salaries”, and reports findings from a World Bank survey, in which one quarter of households “indicated that teachers would inflict some sort of retribution (not teach in school, give poor grades) if not engaged for private tutoring” (p. 19). Nath (2007) reports that the number of primary school students getting supplementary private tuition has been rising in recent years and reached 31 per cent in 2005. In urban areas the proportion rose to more than half. In rural areas, private tuition was more common for boys than girls, but in urban areas this was not the case. Private tuition was more common in wealthier families and those where the parents were more educated; and for children at higher grades. Private tuition was highest among students in kindergartens, followed by (in order) primary sections of secondary schools; government schools; non-government schools; madrassas; and non-formal schools – an order roughly corresponding to the ranking of school-types by prestige described in Imam (2005) (see section 5 above) – although it is not clear whether this order would be maintained if one were to control for household income. Students 28
with educated parents were nearly twice as likely as “first generation learners” to have a private tutor, controlling for other factors. Nath (2007) concludes that private tutors for primary school students have become a “well-accepted norm”. In discussions, parents expressed the view that “If a school functions well, private tutoring is unnecessary, but the schools do not function well” and that students were not able to ask teachers questions, but were able to do this in private tuition (p. 19). This study also provides estimates of private education expenditure showing the importance of private tuition: mean expenditure was 419 taka for children not in private tuition but 1923 taka for those having private tuition. The households in which a child received private tuition spent 46 per cent of total private expenditure for education of that child on supplementary tutoring. However the cost of tuition varied widely, from 20 to 18000 taka with a mean of 887 taka (p. 14), and the amounts spent varied by gender, household income and parental education in similar ways to the likelihood of using private tuition in the first place. This study also presents evidence that learning achievement was greater among children aged 11-12 years who had private tutors, and increased with the amount spent on tuition. In urban areas, only 47 per cent of students without tutors aged 1112 years completed tests to a standard that satisfied “basic education” criteria; amongst those with tutors, this rose to 64 per cent. 7. Use of English as a medium of study The use of English as a medium of study usefully illustrates some of the issues surrounding benefits of schooling. Several studies suggest that private schools are more likely to use English as a medium of instruction, or to start teaching English early, than government schools. Tooley and Dixon’s (2007) New Delhi study found that 47 per cent of recognised private unaided schools, 21 per cent of unrecognised private unaided schools, 37 per cent of private aided schools used English medium, whereas less than 3 per cent of government schools did. However many private unrecognised schools were also Hindi medium or said they used both Hindi and English. Mehrotra and Panchamukhi (2006, p. 424) claim that although non-elite private schools often advertise themselves as English-medium, most of the time they are not genuinely English-medium.
29
Harlech-Jones et al. (2005), reporting a rapid expansion in the number of private schools during 1993-2003 in the Northern Areas of Pakistan, find that English medium was seen as a key requirement by parents, and some educational associations were founded specifically to provide English-medium education. They note that English remains “the language of power and high status in Pakistan”, spoken natively by less than 8 per cent of the population. Proficiency in English alongside Urdu is required “for advancement in the prestigious armed forces, for appointment in the civil service… and for employment in many of the better-paying commercial companies as well as in most NGOs” (p. 562); it is also the medium at universities. English is “desirable and powerful, both in practical and symbolic terms” (p. 563). This apparent preference for English, and the fact that it appears to be one of the factors driving demand for private schools, is interesting because educationalists often see the use of English medium as a problem in developing country schools. Arthur (1996, cited in Smith et al 2005), drawing on a study of African teachers, argues that the use of English as a medium of instruction is a major cause of ritualised teaching practices. The right to learn and use the language of their family is one of the rights accorded under the United Nations Convention on the Rights of the Child (Webley, 2006), and advocates of mother-tongue education argue that when parents choose education in a dominant language such as English, they are “unknowingly going against scientific evidence about learning and bilingualism, as well as the human right of their own children to education in a language they understand” (Skutnabb-Kangas, 2006). There is evidence that children learn better when taught through a language they know well. It may be the case, as these studies suggest, that parents fail to understand that teaching in English leads to poorer pedagogic outcomes; or it may be that they realise this but still prefer English. For instance if the expected duration of education is short anyway, they may be willing to sacrifice other aspects of learning for the possibility of acquiring some status-enhancing English. Another important consideration is that in many cases government schools also use a language other than the child’s first language and so face the same problems in terms of quality of learning.
30
8. Who makes the decisions about schooling? As noted above, children are often not seen as playing an active role in making decisions about education, but in practice the decision about schooling might be best seen as a joint decision by parents and children. But the form that this joint decision making process takes remains to be specified. From a sociological perspective, Punch (2004) emphasises “negotiation” as a way in which children may exercise some agency in decision making: children (in a UK context) “may not be fully independent, but they negotiate a relative autonomy within the constraints which limit their choices” (p. 96); and they “use their resourcefulness to stretch adult-imposed boundaries to limits more acceptable to themselves” (p. 110). Citing Mayall (2001), Punch relates this negotiation to structure and agency debates, with children struggling “to gain a better deal in their relationships within different structures” (p. 95). Drèze and Kingdon (1999) cite findings from the PROBE survey in India suggesting that it is difficult for parents to make a child go to school if the child does not want to go, and suggest there is a particular asymmetry of interests between parents and children in the case of girls, who are likely to leave their parents to get married. Similarly Singh and Sridhar (2002) note that “Child not motivated” was commonly given as a reason for drop-out in both government and private schools in Uttar Pradesh; although household work, helping parents with work, and taking care of siblings were also commonly cited. On the other hand, Tilak (2002, p. 28) notes that the education level of the head of the household affects expenditure on a child’s education more than education level of other members, and interprets this as broadly consistent with a “hierarchical decision-making process in the households in rural India”. Blanchet (1995) found a strong belief in Bangladeshi families in the “unassailable rights of parents to determine their children’s life path” (p. 9, cited in Delap, 2000, p. 731); yet Delap warns that it is “important not to see children as victims of age hierarchies, unable to shape their daily activities” (2000, p. 731). Whilst some children in Delap’s study were dissatisfied with the amount of household work they were required to do, others did not complain about their workload, and generally they made the most of those activities which they did not enjoy. This 31
suggests that child agency may exist in subtle and hard-to-detect forms. In general Delap argues that “theories that emphasize the economic rationality of household decision making are inadequate for explaining task allocation in Bangladeshi slum households” and that cultural and social forces need to be considered, including age and gender hierarchies (p. 723). Further support for child agency in schooling decisions comes from the survey results of M. Ahmed et al (2007, p. 38), which reports that “Child does not like school” was the second most common reason – after “Scarcity of money” – for dropping out of school and the third most common reason given for not enrolling the child in the first place. (This reason, however, was given much more commonly for boys than for girls.)
IV. Research questions and some propositions This research seeks to gain insight into how primary schooling decisions are made by poor households in Dhaka, by answering the following questions for a sample of these households: – What aspects of schooling are valued by parents and children, and how do these vary between households (for instance by income, wealth, and parental education and occupation) and between types of education provision (government, NGO, madrassah, or private school; private tuition)? – What are the costs of schooling, and how do these vary between households and types of provider? – How do these valued benefits and costs combine, and interact with the expectations and aspirations of the parents and children, to determine schooling decisions? A break-down of these questions into sub-questions, along with the intended methods of data collection and analysis, is given in the appendix. The general model of decision making discussed in section II, and illustrated in diagrams 1-3, provides a starting point for narrowing down the focus of the study. It is reasonable to propose that the following simpler model is roughly true for urban Bangladesh: When exercising decision-making power over a child’s schooling, the only agents involved are the mother and father (assuming both are present in the household) and the child him- or herself. The parents weigh the financial costs to themselves against the future benefits for the child and possibly for themselves. The child is somewhat altruistic so that he or she cares about the possible burden placed 32
on the rest of the household by the fact that schooling takes them away from incomegenerating work or housework; and weighs this against his or her future income and other benefits. The parents exert the largest share of control over the decision, although children may have subtle ways of influencing the decision. The father may exert more control than the mother, although the extent of the difference may depend on their relative levels of income generation and education. A further proposition is that families will in most cases try to get as much schooling for their children as they can afford; in other words they will attach a very high value to it, relative to all other commodities except those needed for survival.9 Increasing social acceptance of at least primary education as a norm, combined with increasing difficulty of finding a job if you are uneducated, and fewer opportunities for child work (decreasing the opportunity cost of schooling), will mean that households seek to take maximum advantage of any free (or nearly-free) education provision, and where possible to boost that advantage with additional investments. (The notions of what is affordable or possible, however, will vary; the joint quantitative and qualitative analysis in the study should provide some insights into each household’s own views of such economic thresholds). However it is also possible that trends towards saturation of the labour market with young adults educated to primary level will have driven down the returns to schooling to the extent that families become very sensitive to costs; if this is the case the study should turn up evidence of households making difficult decisions weighing up the costs and expected benefits. A division may emerge between children going to secondary school and children receiving minimal education (possibly even within the same family). Quality of schooling is likely to be generally low, and parents will
9
In such cases the cost benefit calculation may be perceived as a response to the
circumstances in which the family finds itself, rather than a decision actively made, and the study will need to take this into account. However the notion of affordability may cover up underlying value judgements. In addition, genuine decisions will still arise, for example about which child to send to school if it is only possible to send one, or on whether tuition is a worthwhile investment. 33
probably realise this at least to an extent, but it may play a limited part in their decision making process because of limited choice. In cases where there is a choice of schools, schooling that costs substantially more will have to have an apparent (though not necessarily genuine) quality benefit; and parents will be able to explain the nature of this quality benefit in terms of longer-term benefits (economic or other) to the child and/or the household as a whole. As illustrated in diagram 3, decisions about schooling are not one-off but continuous throughout the school-going ages, and interact with decisions concerning other children in the household. It is anticipated that most children will receive at least some schooling, although some may drop out after a short time, with some perhaps dropping out and re-entering school (or staying enrolled but attending intermittantly) in response to household financial problems or the occasional need for child labour. To turn this ongoing, complicated process into a manageable subject for a survey and interview-based study, the research will use the following initial questions as entry points: – Was the child initially enrolled in school? If so, at what age? – If so, until what age did they stay in school? – Did the child have any private tuition? When? — in each case accompanied by ‘why’ questions to explore the reasoning, motivations and process behind these decisions (see appendix for more detail). Thus the research will be guided by these propositions and the simpler model of the decision making process while keeping an eye out for any refinements needed to the assumptions they are based on. Some such refinements may become apparent during the scoping phase of the study in which adjustments will be made.
V. Methodology and methods 1. Methodological approach This research will use a survey approach and a mixture of quantitative and qualitative analysis to look at the more easily quantifiable costs and benefits of schooling, and how these vary with quantifiable (or categorisable) characteristics such as household 34
income and wealth; parental education and occupation; age and sex of child; and the number, ages and sexes of siblings. In-depth interviews, as well as some qualitative analysis of answers to qualitative questions in the survey, will serve two purposes: 1. To test and reflect on economic models of the role of these quantifiable (mainly financial) costs and benefits in schooling decisions, by seeking insight into the ways they are really valued by parents and children and the role they play in actual decision-making processes. 2. To broaden the scope of the enquiry and look at costs and benefits that are not easily quantified. In this process it is necessary to bear in mind that economic models need not be literal descriptions of the decision making process that goes on within an individual household. For instance, there are probably factors that influence decisions but which the individuals concerned are not consciously aware of while making the decision, or which are considered too obvious to be stated in an interview. Nevertheless, this combination of approaches to the same question will provide insights into what type of models are likely to provide a satisfactory approximation of household behaviour, and what factors overlooked by these models are likely to be important. In asking children and parents how they perceive the education system (in terms of what they value about it) the research reflects the fact that “the actor acts towards his world on the basis of how he sees it and not on the basis of how that world appears to the outside observer” (Blumer, 1972, p. 21, cited in Crossley and Vulliamy, 1997, p. 5). It will not claim to see the world from the participants’ perspectives; but aims to gain better understanding of those perspectives. There may be a gap between what people say about their motivations and the factors that in practice guide their decision making, especially if we wish to enumerate these factors comprehensively in a way that is in some sense “objective”, rather than focusing on those that are most salient to the participants in the study. Greene and Hill (2005, p. 7) note that “People can report on their motivations and 35
emotions only to the extent that they are aware of them and only in the manner that they have come to interpret them”. There is also a danger that the things people say in interviews may reflect received wisdom, especially regarding a topic like schooling which is arguably an important site in which power relations within a community, and between the community and wider society, are played out.10 Communities may seek to project an official view of themselves. While it might not be possible to avoid these pitfalls entirely, the study will be wary of them, paying close attention to the ways in which different contexts or differently-worded questions may elicit different types of answer. It will also take care to analyse the responses in a way that reflects the possible gaps between behaviour, the respondents’ own interpretation of that behaviour, and the way they then explain their interpretation to a researcher. While the research will not seek systematically to test hypotheses arising from economic models of the household, the results are likely to shed light on the idea of a rational household as a theoretical starting point, and on (i) whether the idea of a unitary household is useful and (ii) whether rationality extends over a much wider array of costs and benefits than those usually taken into consideration in economic studies. Even if these two simplifying assumptions are not correct, they may still be useful assumptions if they are correct most of the time. The research may help to elucidate what the dangers are of making these assumptions and whether they outweigh the benefits in terms of providing a simple model that can be used for tasks such as policy analysis. Other parts of the theoretical background of the research emphasise the social and cultural factors that may be important in shaping household decisions, and the research will consider whether it might be possible to evaluate these different factors within a single expanded cost-benefit style framework, even if they are difficult to quantify in practice.
10
See Maddox (2005), which notes that women in rural Bangladesh were sometimes
reluctant to admit to wanting to acquire education for reasons, such as managing household budgets, that were controversial aims for women to have in that community. 36
A broad economic model may help to clarify how different methods will try to deliver insights into different aspects of the situation. The household chooses a combination of schooling and tuition that maximises the total benefits of schooling, minus the total direct costs, and minus the opportunity costs: Maximise U = L (K0, S1, S2, … , SN) – C (K0, S1, S2, … , SN) – w(K0)T with respect to S1, S2, …, SN, where -
K0 is the household’s capital in the “first period” (when schooling decisions are being made), seen as a vector of different forms of (human, social and financial) capital that may have distinct effects on the benefits and costs of schooling
-
S1, S2, …, SN are the number of years spent in each type of school
-
L is the array of benefits in terms of different types of capital, weighted to reflect how highly each benefit is valued by the decision maker (and bearing in mind that the decision maker may not be the one who directly receives these benefits);
-
C is the total financial cost of schooling, which is seen as potentially depending on the costs of each year of schooling but also on the capital (including social capital) of the household
-
w is the (implicit or explicit) wage paid for child labour, for a family with capital K0, and
-
T is total schooling, ΣSi
The solution chosen will be where the net benefits from each type of schooling are equalised: ∂L / ∂S1 – ∂C / ∂S1 – ∂T / ∂S1 = ∂L / ∂S2 – ∂C / ∂S2 – ∂T / ∂S2 = etc. An important constraint on this in practice will be the difficulty in attending more than one type of school, because of constraints on registration etc. In practice an individual child is likely to attend no more than one or two schools, even if a larger number of schools are available, plus (in some cases) private tuition. But families with more than one child can be seen as investing in a mixture of types of schooling. The survey data collected will indicate the amounts of each type of schooling or tuition chosen, the costs, the household’s wealth and the potential value of child’s labour, although there are issues about what measure of wealth to use and how to 37
measure implicit wages for child labour, especially in the case of girls11. From this it is possible to conduct quantitative investigation on the form of the benefits and costs functions L and C. For instance, do costs actually vary according to the household’s initial capital? This may be the case if social capital is defined broadly, to encompass the availability of different school types in the area. But between households within the same area there may be little difference, so that C becomes a uniform function of the amounts of each type of schooling chosen. It may also be possible to make inroads into understanding the relationship between initial capital and the benefits of each type of schooling by quantitatively analysing how the amount of schooling chosen varies with financial wealth, parental education, and (perhaps) measures of social capital. This benefits function also incorporates the family’s preferences, however, and understanding the role of these preferences in the decision; whether the benefits of schooling are valued as means or as ends in themselves; and how they are influenced by the family’s social environment – such as norms regarding the education of girls – calls for qualitative analysis of the survey and interview data. Research with children As explained in the literature review, this research aims to take the possibility seriously that children may play an active role in schooling decisions and to learn more about their own values and aspirations in relation to schooling. Relevant antecedants include Woodhead’s (1998,1999) international studies of child labour, which included asking children what are the good and bad things about school; and Delap’s (2000) study of child housework in a Dhaka slum area, in which children aged 10-15 were interviewed as well as their parents. With Greene and Hill (2005), I assume that “it is possible to learn about children’s experience,” and about their understanding of education and the society they live in, “both by enquiry into their active engagement with their material and
11
One complicaton is that learning outside of school may be seen as more useful (for
some children, at some ages) than that in school; this effect raises this opportunity cost of schooling (i.e. increases w). 38
social worlds, whether the focus is on action or words, and from their own reports on their subjective world” (p. 6). These authors argue that “researchers should not take for granted any adult-child distinction” but “must be open to the use of methods that are suited to children’s level of understanding, knowledge, interests, and particular location within the social world.” (p. 8). Children “are not necessaily less reliable informants than adults” (p. 10) and Greene and Hill do not find any evidence that they are more suggestible than adults. Similarly Kellet and Ding (2000) note that recent studies have challenged claims that children are less reliable respondents than adults, and suggest “Children can and do provide reliable responses if questioned in a manner they can understand and about events that are meaninful to them.” (p. 165). O’Kane (2000) describes how the way we see children informs selction of research methods and techniques, and links participatory research techniques to a view of children as having “different, though not necessarily inferior social competencies” to adults; seeing children in this way makes it easier to develop research techniques which “engage more effectively with children, allowing them to participate on their own terms and thereby enabling us to learn more about their experiences of the world.” (p. 139). Similarly France (2004) notes that children “tend to be portrayed as ‘cultural dupes’ and not as competent to explain and theorise about their own social worlds” (p. 176). The questions asked in the qualitative parts of the research will leave space for children to do this, although language issues and circumstances of the interview may place limits on the extent to which a full picture of children’s understandings will emerge. 2. Methods Initial scoping and piloting Initial scoping will seek to identify suitable areas for the study; what types of educational provision are available in the area, including availability of different forms of private tuition; and preliminary information on what costs and benefits households are likely to face, including whether stipends are available. It will also help to refine the survey strategy discussed below. All other things being equal, an area with a larger diversity of education provision would be favoured over areas with less diversity as this would provide more 39
scope to study the types of variation in decision-making that are the focus of this study. One possible strategy is to select about five areas around Dhaka identified as slums by Measure Evaluation / NIPORT (2006) and other relevant studies and by local informants, covering a spread of characteristics including peripheral vs. central; poorer vs. relatively affluent; and those that are least well-served by public services vs. those that are better served. A sample of around 20 households in each of the areas could then be interviewed. However there is a danger that this strategy could fail to capture variation within particular slum areas, since the number of households sampled within each area would be relatively low. Very large numbers of people live in some of these areas – according to the Measure Evaluation / NIPORT report, over 250,000 in the slums in the Kamrangirchar area, and around 100,000 in Karail, the city’s largest single slum – and so there may be considerable variation in the availability of different types of schooling, socioeconomic status, and other relevant variables within each area. The scoping will thus consist of initial interviews with informants with experience of Dhaka, to determine a strategy that best combines breadth and depth. Information will also be sought on strategies or ideas for interviewing; for instance it may be difficult to access the views of household members other than the head of household, who may try and answer on behalf of the whole household; but local informants may have ideas about how to overcome this difficulty. Pilot interviews will also be conducted with a limited sample (around 5 households) from (one or more of) the study areas. The pilot will consist of the survey interview, followed by follow-up interview going into more depth and exploring “how” and “why” questions. This should make apparent any alterations needed to the survey questions to ensure they make sense to the interviewees and achieve responses that appear open and as comprehensive as possible. The in-depth interview part of the pilot should help to reveal topics that are important to the respondents and relevant to the study’s focus but which have been omitted from the survey; for instance aspects of schooling that are valued. 40
Survey Around 100 households where at least one child aged 11-16 lives will be surveyed. Since the study aims to cover children of secondary school age who are not currently in school (as well as those who are), the most appropriate strategy may be to use a school as a starting point (assuming there is a secondary school which is used by children in the area). Focus groups could be conducted at the school and information on school characteristics gathered, before using a snowball method – asking to be introduced to siblings, relations, neighbours etc. of children who attend the school – to meet and interview children who do not go to school. It needs to be noted, however, that the location of interview is likely to influence children’s responses; for instance “the social meaning children will attach to concepts such as work or honesty may differ depending on whether children are at home or at school.” (Scott, 2000, p. 103). Thus for consistency, children will be interviewed at or near their homes where practicable, and the location of interviews and discussions will be noted in analysing the resulting qualitative data. During survey interviews with parents and children, questions will be asked on the following areas: – Parents’ educational levels, occupations, income and a measure of household wealth. Measures of household wealth could draw on the CASHPOR House index (Simanowitz, Nkuna and Kasim, 2000) which looks at house size, structural condition, quality of walls and type of roof may be appropriate; the CREATE household survey instrument; and Filmer and Pritchett’s (1999) index based on similar DHS survey questions about assets and household characteristics. – Composition of the household (whether female- or male-headed; number and ages of children) – The age and sex of each child and their history of schooling (approximated to number of years spent in school, and what type of school it was). Also whether they have used private tuition and for how much of the time. – How much was spent on this schooling and tuition, and whether they benefited from any stipend schemes in meeting these costs. 41
– Views about schooling: how good was your school? What were the good things about it? What were the problems? What were the most important things you learned? How have you used the things you have learned? What difference does it make to your life now? How do you expect to use them in future? – Expectations and aspirations: what do you think you will be doing in ten years’ time? What would you like to be doing? What are some of the difficulties in getting the type of work you would like to do? It may be possible to use the baseline survey being conducted by CREATE (CREATE, 2007) in rural Bangladesh as a basis for the survey, adding extra questions where necessary. Surveys will be conducted in Bengali via a translator. Where possible the questions concerning views and aspirations will be asked of both parents (if both are present in the household) or, if the child does not live with his or her parents, all adults in the household who are involved in the child’s upbringing – as well as the children themselves being interviewed. It is recognised that there may be difficulty accessing all members of the household in this way. Strategies for overcoming this will be explored as part of the initial scoping and piloting, and may include re-visiting the household at different times of day. However ethical considerations such as potential repercussions for the participants of taking part in the survey will be considered, and may ultimately place limits on what data can be collected. Follow-up interviews and/or focus groups More in-depth interviews will be conducted with a sub-sample of households, in an attempt to gain greater insight into the causal models expressed by parents and children as explanation of their schooling decisions, as well as to pick up potential nuances or qualifications to the answers expressed in the surveys. Households will be chosen for follow-up interviews depending on their survey answers, aiming at a crosssection in terms of school type, gender of the child, and household socioeconomic characteristics (See the appendix for more detail). Focus group discussions may provide a way of exploring issues around aspirations and expectations relating to schooling and the labour market, and may offer advantages over individual interviews. Focus groups may be less intimidating 42
for some children than a one-to-one interview with a foreign researcher (Kellett and Ding, 2000). It may be that a different type of information emerges, for instance O’Kane (2000) suggests that differences between responses in focus groups and individual interviews can help “to acknowledge different types of discourse in the ‘private’ and ‘public’ arenas” (p. 148), and quotes Kitzinger’s (1994, p. 116) view that “focus groups do not easily tap into individual biographies or the minutia of decision making during intimate moments, but they do examine how knowledge and, more importantly, ideas both develop, and operate, within a given cultural context”. Regarding specific methods for interviews and focus groups, Kellett and Ding (2000) list a number of methods for working with children, including Hill, Laybourn and Borland’s (1996) example of adapting focus groups for children; the multimethod “mosaic” approach of Clark and Moss (Clark, 2004); and participatory techniques used by Thomas and O’Kane (2000) to elicit children’s views about decision-making processes. Although many of these methods have been used mainly with younger children in developed country settings, they provide starting points for developing methods suited to the context and research questions of this study. Interviews and discussions will be conducted in Bengali via a translator. With regard to accessing different household members, similar considerations apply to those mentioned above for the survey. Ethics Informed consent will be asked from both parents and children prior to each interview and focus group discussion. Responses will be anonymised, although it is not anticipated that any negative repercussions would be likely to result for participants following publication of the results of this study. Analysis of data Regression analysis will be used to investigate how the number of years spent in schooling, the type of school chosen, and decisions concerning private tuition, relate to household income, measures of wealth, geographical area, and the availability of different types of school and tuition. Non-economic benefits of schooling will be coded and subjected to similar quantitative analysis.
43
As discussed above, qualitative analysis of the interview data will examine the full range of costs and benefits of schooling, and aim to understand how these are weighed up (and by whom) in making schooling decisions. The analysis is likely to proceed by identifying responses under the following categories: benefits of schooling; costs; ambiguous aspects which can be either benefits or costs; insights into how the decision about schooling is actually made; and insights into who makes the decision. However, further unanticipated categories may arise during the course of the interviews. The appendix shows what methods of data collection and analysis will be used for each research question.
VI. Contribution to the field The key contribution of this research will be to provide micro-level understanding of the actual decision making process of families in a poor area of a country striving to achieve educational goals such as Education For All. It will thus complement larger scale studies of household expenditure such as FMRP (2006). For instance, that study reports that expenditure on boys is on average higher than that for girls, but not at all levels of income and all types of school. Micro-level work is needed to suggest explanations for these patterns and to understand their potential implications for educational policy. The study should make clearer what, if any, educational subsidies (such as the Primary Education Stipend Programme) are made use of by families in poor areas of Dhaka and what impact these have on their educational decisions, with possible implications for the design and funding of these subsidies. The study will also reveal how expenditure decisions about different siblings within the same family interact, and how these depend on the gender of each child, an area where there is little detailed knowledge. The initial scoping should be of value in revealing what types of providers are available to a particular slum area, and may pick up on forms of provision that go below the radar of larger-scale studies.
44
Potentially the study could tie in with work done on the CREATE baseline survey being conducted in rural areas, extending its coverage to parts of Dhaka and providing a basis for more detailed future research on the same area.
References Ahmed, M., K.Z. Ahmed, N.I. Khan, and R. Ahmed (2007) Access to education in Bangladesh: country analytic review of primary and secondary education. BRAC University Institute of Educational Development / Consortium for Research on Educational Access, Transitions and Equity (CREATE). Alderman, H., P. Orazem, and E. Paterno (2001) School quality, school cost, and the public/private school choices of low-income households in Pakistan. Journal of Human Resources Vol. 36, No. 2, pp. 304-325. Alkire, S. (2002) Valuing Freedoms: Sen’s Capability Approach and Poverty Reduction. Oxford University Press. Andrabi, T., J. Das, and A. Khwaja. (2002) The Rise of Private Schooling in Pakistan: Catering to the Urban Elite or Educating the Rural Poor?” Mimeo, Kennedy School of Government, Harvard University Andrabi, T., J. Das, and A.I. Khwaja (2005) Private schooling: limits and possibilities. Accessed online, July 2007, at: http://www.economics.pomona.edu/Andrabi/Research/PrivateSchoold_Final_Nov5 .pdf Appleton, S. and F. Teal (1998) Human capital and economic development. Background paper prepared for the African Development Report. African Development Bank. Asadullah, M. N. (2006) Returns to education in Bangladesh. Education Economics, 14: 4, 453-568. Bashir, S. (1997) The cost effectiveness of public and private schools: knowledge gaps, new research methodologies, and an application in India, in: C. Colclough (Ed.) Marketizing education and health in developing countries (Oxford, Clarendon Press). Becker, G. S. (1964) Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education. Chicago, University of Chicago Press. Becker, G. S. and H. G. Lewis (1973) On the interaction between quantity and quality of children. Journal of Political Economy, Vol. 81, No. 2, pp. S279-88. Behrman, J.R., R.A. Pollak, and P. Taubman (1982) Parental preferences and provision for progeny. Journal of Political Economy, vol. 90, no. 1, pp. 52-73. Bennell, P. (1996) Using and abusing rates of return: a critique of the World Bank’s 1995 education sector review. International Journal of Educational Development, 16,3, 235-248. Blunch, N.-H. and D. Verner (1999) Is Functional Literacy a Pre-Requisite for Entering the Labor Market? An Analysis of the Determinants of Adult Literacy and Earnings in Ghana. CLS Working Papers 00-5, Aarhus School of Business, 45
Centre for Labour Market and Social Research. http://www.cls.dk/workingpapers/docfiles/70.pdf Bray, M. (1999) The Shadow Education System: Private Tutoring and its Implications for Planners. Paris: UNESCO/International Institute for Educational Planning. Bray, M. (2003) Adverse effects of private supplementary tutoring. Dimensions, implications and government responses. Paris: UNESCO/International Institute for Educational Planning. Chen, S. H. (2001) Is investing in college education risky? Discussion paper 01-09, State University of New York at Albany, Department of Economics. Chowdhury, A. M. R., S. R. Nath, R. K. Chowdhury and Manzoor Ahmed (2001) Renewed Hope, Daunting Challenges: State of Primary Education in Bangladesh, Education Watch 2001, Campaign for Popular Education and the University Press Limited. Christensen, P. and A. James (eds., 2000) Research with Children. London: Falmer Press. Clark, A. (2004) The Mosaic approach and research with young children, in V. Lewis, S. Fraser, M. Kellett, C. Robinson and S. Ding (eds) The Reality of Research with Children and Young People, London, Sage. Clarke, P. (2003). Culture and classroom reform: the case of the district Primary Education project, India. Comparative Education, 39, 127-44. CREATE (2007) Instrument for CREATE household survey (document HHS1). Consortium for Research on Educational Access, Transitions and Equity (CREATE), University of Sussex. Crossley, M. and G. Vulliamy (1997) Qualitative reseach in developing countries: issues and experience. In Crossley and Vulliamy (eds.) Qualitative Educational Research in Developing Countries, New York: Garland Das Gupta, A., S. Islam, and M.S. Siddiq (1993) Female Education Awareness Program of Female Secondary School Assistance Project - Baseline KABP survey: a report, Dhaka: Associates for Communications Options, prepared for Academy of Educational Development. Delap, E. (2000) Child housework in urban Bangladesh: an exploration of intrahousehold labour deployment. Journal of International Development, 12, 5, 723734. Doss, C. (1996) Testing among models of intrahousehold resource allocation. World Development, 24, pp. 1597-1609. Drèze, J. and G. Kingdon (1999) School participation in rural India. Development Economics Discussion Paper No. 18, London School of Economics. Duraisamy, P. (2000) Changes in Returns to Education in India, 1983-94: By Gender, Age-Cohort and Location. Discussion Paper No. 815, Yale University Economic Growth Centre. Fields, G. (1974) The private demand for education in relation to labour market conditions in less-developed countries. Economic Journal, 84, 336 (December), 906-925. 46
Filmer, D., and L. Pritchett. 1999. "The Effect of Household Wealth on Educational Attainment: Evidence from 35 Countries." Population and Development Review 25(1): 85-120 FMRP (2006) Social sector performance surveys: primary education (draft final report). Financial Management Reform Programme, Bangladesh. Foster, A.D. and Rosenzweig, M.R. (1996) Technical Change and Human-Capital Returns and Investments: Evidence from the Green Revolution. American Economic Review,86, 4 (September), 931-953. France, A. (2004) Young people. In S. Fraser, V. Lewis, S. Ding, M. Kellett, and C. Robinson (eds.) Doing Research with Children and Young People, pp. 175-190. London: Sage, 2004. Fraser, S., V. Lewis, S. Ding, M. Kellett, and C. Robinson (eds., 2004) Doing Research with Children and Young People. London: Sage, 2004. GOPRB (2006) Report on the baseline survey of Second Primary Education Development Programme (PEDP-II). Directorate of Primary Education, Ministry of Primary and Mass Education, Government of the Peoples’ Republic of Bangladesh. Greene, S. and Hill, M. (2005) Researching children’s experience: methods and methodological issues. In S. Greene and D. Hogan (eds.), Researching children's experience: approaches and methods, pp. 1-21. Sage Publications Ltd, London. Greene, S. and D. Hogan (eds., 2005), Researching children's experience: approaches and methods, pp. 1-21. Sage Publications Ltd, London. Hanushek, E. A. (1992) The trade-off between child quantity and quality. Journal of Political Economy, Vol. 100, No. 1, pp. 84-117. Harlech-Jones, B., M. Baig, S. Sajid, and S. ur-Rahman (2005) Private schooling in the Northern Areas of Pakistan: a decade of rapid expansion. International Journal of Educational Development 25, 557-568. Harris, J.R. and M. P. Todaro (1970) Migration, unemployment and development: A two-sector analysis. American Economic Review, 60, 1, 126-142. Hill, M., A. Laybourn and M. Borland (1996) Engaging with primary-aged children about their emotions and well-being: methodological considerations. Children and Society, 10: 129-144. Hossain, N. (2005) Inheriting extreme poverty: household aspirations, community attitudes and childhood in northern Bangladesh. BRAC / Save the Children UK. Accessed December 2007 at www.bracresearch.org/publications/iep_text.pdf. Imam, S. R. (2005) English as a global language and the question of nation-building education in Bangladesh, Comparative Education, vol. 41, no. 4, pp. 471 – 486. International Herald Tribune (2007) ‘Bangladesh reveals budget focusing on education, rural development, transport infrastructure’ (June 7, 2007). Accessed 06/2007 from http://www.iht.com/articles/ap/2007/06/07/business/AS-FINBangladesh-Budget.php Iversen, V. (2002) Autonomy in child labor migrants. World Development, 30, 5, 81734. 47
Jeffery, R., P. Jeffery, and C. Jeffrey (2003) Privatisation by default: the case of secondary schooling in north India. Unpublished preliminary report. University of Edinburgh. Johnson, V., E. Ivan-Smith, G. Gordon, P. Pridmore, and P. Scott (eds., 1998) Stepping forward: children and young people's participation in the development process. London: Intermediate Technology Publications. Kabeer, N., and S. Mahmud (2004). ‘Globalization, Gender and Poverty: Bangladesh Women Workers in Export and Local Markets’. Journal of International Development, 14, 1, 93-109. Kellett, M. and S. Ding (2000) Middle childhood. In S. Fraser, V. Lewis, S. Ding, M. Kellett, and C. Robinson (eds., 2004) Doing Research with Children and Young People, pp. 161-174. London: Sage, 2004. Kingdon, G.G. (1996) The quality and efficiency of private and public education: A case-study of urban India. Oxford Bulletin of Economics and Statistics, 58, 1. Kingdon, G.G. (2005) Private and public schooling: the Indian experience. Paper prepared for the conference: "Mobilizing the Private Sector for Public Education". Lang, K. (1994) Does the human-capital/educational-sorting debate matter for development policy? American Economic Review, 84, 1 (March), 353-358. Levhari, D. and Weiss, Y. (1974) The effect of risk on the investment in human capital. American Economic Review, Vol. 64, No. 6, pp. 950-963. Lubienski, C. (2006) School diversification in second-best education markets: international evidence and conflicting theories of change. Educational Policy, Vol. 20, No. 2, pp. 323-344. Maddox, B. (2005) Assessing the impact of women’s literacies in Bangladesh: an ethnographic inquiry. International Journal of Educational Development, 25, 2, 123-132. Maitra, P. (2003) Schooling and educational attainment: Evidence from Bangladesh. Education Economics, 11(2), 129-153. MEASURE Evaluation / NIPORT (2006) Slums of urban Bangladesh: mapping and census, 2005. Centre for Urban Studies / Measure Evaluation / National Institute of Population Research and Training. Accessed 9 June 2007 at http://www.cpc.unc.edu/measure/publications/pdf/tr-06-35.pdf Mehrotra, S. and Panchamukhi, P.R. (2006) Private provision of elementary education in India: findings of a survey in eight states. Compare, Vol. 36, No. 4, pp. 421442. Mehrotra, S. (2006) What ails the educationally backward states? The challenges of public finance, private provision and household costs. Chapter 1 in Mehrotra (ed.) The Economics of Elementary Education in India: the challenge of public finance, private provision and household costs. New Delhi: Sage. Morrow, V. (1996) Rethinking child dependency. Sociological Review, 44, pp. 58-77. Muralidharan, K. and M. Kremer (2006) Public and private schools in India. Forthcoming in P. Peterson and R. Chakrabarti (eds.) School Choice International
48
Nambissan, Geetha, B. (2003) Educational deprivation and primary school provision: a study of providers in the city of Calcutta, IDS Working Paper 187, Institute of Development Studies, UK. Nath, S.R. (2007) Private supplementary tutoring among primary students in Bangladesh. BRAC Research and Evaluation Division, Dhaka. Nath, S.R. and M.R. Chowdhury (2002) Level and trend of basic education of children in Bangladesh: 1993-1998. Educational Studies, 28, 1, pp. 77-92. North, D. (1994) The new institutional economics and development. Working paper accessed online 24th July 2007 from www.econ.iastate.edu/tesfatsi/NewInstE.North.pdf. O’Kane (2000) Development of participatory techniques. In P. Christensen and A. James (eds) Research with Children, pp. 136-159. London: Falmer Press. Opel, A.E.A (2000) The social content of labour markets in Dhaka slums. Journal of International Development, 12, 5, pp. 735-750. Panchamukhi, P.R. (2006) Universalising elementary education in Madhya Pradesh: can the successes of decentralised governance offset the problems of public finance, private provision and private cost? In S. Mehrotra (ed.) The Economics of Elementary Education in India: the challenge of public finance, private provision and household costs, ch. 5. New Delhi: Sage. Pratham (2007) Annual Status of Education Report, 2006. Mumbai: Pratham Resource Center. PROBE Team (1999) Public Report on Basic Education in India. New Delhi: Oxford University Press. Psacharopoulos, G. and Patrinos, H.A. (2002) Returns to Investment in Education: A Further Update. Policy Research Working Paper No. 2881. Washington, D.C.: World Bank. http://econ.worldbank.org/files/18081_wps2881.pdf Punch, S. (2004) Negotiating autonomy: children’s use of time and space in rural Bolivia. V. Lewis, S. Fraser, M. Kellett, C. Robinson and S. Ding (eds) The Reality of Research with Children and Young People, London, Sage. Rahman, R.I. (2005) The dynamics of the labour market and employment in Bangladesh: a focus on gender dimensions. Employment Strategy Papers, ILO. Accessed 19th June 2007 at http://www.ilo.org/public/english/employment/strat/download/esp2005-13.pdf. Ray, R. (1988) The contested terrain of reproduction: class and gender in schooling in India, British Journal of Sociology of Education, 9(4), 387–401. Raynor, J. (2005) Attitudes to girls’ education in Bangladesh: watering a neighbour’s tree? Oxfam. Accessed December 2007 at www.oxfam.org.uk/what_we_do/resources/downloads/BA_6.pdf. Riley, J.G. (1979) Testing the educational screening hypothesis. Journal of Political Economy, 87, 5, Part 2: Education and Income Distribution (October), S227-S252. Robeyns, I. (2006) Three models of education: rights, capabilities and human capital. Theory and Research in Education, vol. 4, no. 1, pp. 69-84. Sen, A. (2000) Development as Freedom. Oxford University Press. 49
Shafiq, M.N. (2007) Household rates of return to education in rural Bangladesh: accounting for direct costs, child labour and option value. Education Economics iFirst 1-16. Sharma, R.K., M.S. Kumar, and S. Meher (2002) Education, skills and the labour market in a globalised world: a case of India. Indian Journal of Labour Economics, 45, 4. Sherman, J. D. and J. M. Poirier (2007) Educational equity and public policy: comparing results from 16 countries. UNESCO Institute for Statistics, working paper no. 6. Simanowitz, A., B. Nkuna, and S. Kasim (2000) Overcoming the obstacles of identifying the poorest families: using Participatory Wealth Ranking (PWR), the CASHPOR House Index (CHI), and other measurements to identify and encourage the p articipation of the poorest families, especially the women of those families. Paper prepared for the Microcredit Summit Meeting of Councils in Abidjan, Côte d’Ivoire, June 1999 and updated June 2000 for the Microcredit Summit Campaign regional meetings in Africa, Asia and Latin America. Downloaded October 2007 from http://www.microcreditsummit.org/papers/povertypaperH.pdf Singh, S. and K. S. Sridhar (2002) Government and Private Schools: Trends in Enrolment and Retention, Economic and Political Weekly, Vol. 37, No. 41, 12th October, 2002. Skutnabb-Kangas, T. (2006) Linguistic genocide? Children’s right to education in their own languages. id21 insights education, no. 5. http://www.id21.org/insights/insights-ed05/ Smith, F. Hardman, F. and Tooley, J. (2005) Classroom interaction in private schools serving low-income families in Hyderabad, India. International Education Journal, 6 (5), 607-618. Spence, M. (1973) Job market signalling. Quarterly Journal of Economics, 87 (August), 355-79. Srivastava, P. (2006) Private schooling and mental models about girls’ schooling in India. Compare, vol. 36, no. 4, pp. 497-514. Stiglitz, J. E. (1975) The theory of “screening,” education, and the distribution of income. American Economic Review, 65, 3 (June), 283-300. Tilak, J.B.G. (2002) Determinants of Household Expenditure on Education in Rural India. Working Paper 88, National Council of Applied Economic Research, New Delhi. Thomas, N. and O’Kane, C. (2000) Discovering what children think: connections between research and practice, British Journal of Social Work 30, pp. 817-33. Tietjen, K. (2003) The Bangladesh Primary Education Stipend Project: a descriptive analysis. World Bank. Accessed July 2007 from www1.worldbank.org/education/pdf/BangladeshStipend.pdf. Tooley, J. and P. Dixon (2003) Private schools for the poor: a case study from India. CfBT. Accessed November 2007 from www.cfbt.com/downloads/psp_india.pdf.
50
Tooley, J. and P. Dixon (2007) Private schooling for low-income families: A census and comparative survey in East Delhi, India International Journal of Educational Development, Volume 27, Issue 2 , March 2007, Pages 205-219 Unterhalter, E., J. Ross and M. Alam (2003) A fragile dialogue? Research and primary education policy formation in Bangladesh, 1971-2001. Compare, 33, 1, 85-99. Webley, K. (2006) Mother tongue first: children’s right to learn in their own languages. id21 insights education, no. 5. http://www.id21.org/insights/insightsed05/ Weiss, Y. (1972) The risk element in occupational and educational choices. Journal of Political Economy, Vol. 80, No. 6, pp. 1203-1213. Weidrich, Eva (2005) “Education Vouchers: Is there a model for India?”, mimeo, Centre for Civil Society, New Delhi. Williams, E. (2004) Children’s participation and policy change in South Asia. Childhood Poverty Research and Policy Centre, Report No. 6. Williams, E. (2005) Small hands, big voices? Children’s participation in policy change in India. IDS Bulletin, 36, no. 1, pp. 82-90. Witter, S. and J. Bukokhe (2004) Children's perceptions of poverty, participation, and local governance in Uganda. Development in Practice, vol. 14, no. 5, pp. 645-659. Woodhead, M. (1998) Children’s perspectives on their working lives: a participatory study in Bangladesh, Ethiopia, the Philippines, Guatemala, El Salvador and Nicaragua. Rädda Barnen. Woodhead, M. (1999) Combatting child labour. Childhood, 6, 27-49.
51
Appendix: research questions and methods of data collection and analysis Sub-questions
Methods of data collection
Types of analysis
1. What aspects of schooling are valued by parents and children, and how do these vary between households and types of provider? 1.1. What do parents and children say they liked or disliked about primary school? Why did they choose that type of school?
Survey
Description of the range of answers given and how many people gave each answer.
Interview
Qualitative exploration of nuances and tensions in answers; and why they like these aspects
Survey
Description of the range of answers given and how many people gave each answer.
Interview
Exploration of whether these stated perceptions are really believed
1.4. What do they perceive as other future benefits of schooling?
Survey Interview
,, Qualitative exploration of the benefits, where appropriate relating these to a theoretical framwork in terms of social / human / financial capital and/or functionings and capabilities.
1.5. How does the schooling the child has been through relate to his or her life now? How does it help?
Interview; possibly survey
,,
1.6. Do parents and children value different aspects of
Survey / interview
Quantitative break down of quantitative
1.3. What do they perceive as the future benefits of primary schooling, in terms of work?
primary schooling?
1.7. How do the above indicators vary by household wealth; parental occupation and education; child’s gender; number of children; and type of schooling?
differences between parents’ and children’s answers in the above questions; qualitative analysis of differences in emphasis in answers given. Survey
Regression analysis, possibly using a simple economic model derived from that specified in section V.
Interviews: - values, aspirations, expectations - counterfactual questions e.g. “Would your answer have been different if…”
Descriptive analysis, noting patterns or diversity in interview responses by household characteristics.
2. What are the costs of primary schooling, and how do these vary between households and types of provider? 2.1. How much do parents spend on: fees/payments to Survey school; stationery; uniform; transport; etc.
Basic descriptive analysis; plus these feed into question 3 below.
2.2. What financial support (e.g. stipends) is available (in theory and in reality)?
Survey
,,
2.3. Does the child work in / out of the home as well as / instead of going to school?
Survey
,,
2.4. Would the child be working if he/she was not in school?
Survey
,,
2.5. How much time and money is spent on private tuition?
Survey
,,
53
2.6. Who helps the child with his/her schoolwork? How much time is spent by parents and/or others doing this?
Survey
,,
2.7. How do the above indicators vary by household wealth; parental occupation and education; child’s gender; number of children; birth order; and type of schooling?
Survey
Descriptive statistics and regression analysis
3. How do these valued benefits and costs combine, and interact with the expectations and aspirations of the parents and children, to determine schooling decisions? 3.1. Why did they choose that type of school?
Survey; interviews
Descriptive statistics examining relationships between reasons given and household characteristics. Exploration of nuances, tensions etc. in reasons given using interview data.
3.2. What do they (a) hope and (b) expect that the child will do for work as an adult?
Survey
Description of the range of answers given and how many people gave each answer.
Interviews
Qualitative exploration, e.g. of whether the expectations expressed are genuine and realistic, or more aspirational;
Survey: household income; parents’ occupation; parents’ education; gender of child; birth order of child
Regression analysis, possibly using a simple economic model derived from that specified in section V.
3.3. What household-level and individual-level variables are associated with different levels of schooling and different types of school?
54
3.4. How do schooling decisions relate to the aspirations and expectations of parents and children?
Survey results
Regression analysis of different categories of occupational aspiration / expectation with school type chosen and amount spent
Interviews
Qualitative analysis of the causal models with which parents and children relate their schooling decisions to their aspirations and expectations
55
Diagrams Diagram 1: mental models of household members combine to determine schooling decisions. mother’s mental model father’s mental model child’s mental model mental models of siblings and other interested parties, e.g. relatives who might contribute to costs of schooling
bargaining or collective decision making process depending on distribution of perceived benefits in each mental model
56
jobs available now at different levels of education competition for those jobs trends in employment and education of rest of population
perceived labour market benefits in future
mental model of child’s increased human, social capital from schooling / perceived tuition
risk/uncertainty
perceived risk/uncertainty
school quality
perceived other benefits (e.g. social capital)
model of benefits to both the child and other family members through, e.g. future remittances sent by an educated child who gets a betterpaid job; externalities of having a literate household member
BENEFITS
child’s abilities
COSTS Each agent’s own education and social setting helps determine: - their benchmark of what constitutes a good school; - their ability to evaluate a school with regard to this benchmark; - their understandings about what abilities are important for education; - their ability to assess the child’s abilities. Similarly factors such as their own involvement with labour markets are likely to affect their understanding
current and future stipend schemes financial costs of school (different types)
model of costs to each family member of child going to (a particular type of) school
opportunity costs
of current and future labour market benefits to particular levels and types of education.
Diagram 2. Mental models of costs and benefits of schooling / tuition. 57
= decisions to be taken
Diagram 3. The decision making process over time
Availability (and costs) of secondary school entrance exam
= factors feeding into the decisions
Child passes secondary schooling entrance exam Availability of different types of school (including safe access)
Costs and benefits of primary schooling and tuition (see diagram 2)
Availability of (different types of) secondary school
Costs and benefits of secondary schooling and tuition (see diagram 2)
What type of secondary school? Private tuition? Availability of tuition
Private tuition?
Stop schooling / enter secondary school
Drop out / advance through grades / repeat same grade
Academic / vocational track (grades 9 and 10) Drop out / advance through grades / repeat same grade
What type of primary school? There may be the possibility of switching schools after having entered, although there are likely to be costs associated with this. Enrol in school? NGO schools, madrassahs may admit children up to a later age than government schools 5
11
16
AGE
58
