Findings from the 2010/11 Basic Education Service Delivery Survey - A Collection of Policy Notes. 1. Service Delivery in Basic Schools in Four States of Sudan 2. Student Learning in Four States of Sudan: School Inputs and Teacher Characteristics 3. Challenges and Possibilities for Teacher Management in Four States of Sudan 4. Assessing the Role of Education Councils in School Management in Four States of Sudan 5. Challenges in Basic Education Provision for Internally Displaced Children in North Darfur April 2012 The Republic of the Sudan The World Bank Federal Ministry of General Education Education Unit, Africa Region Copyright Page ii Contents Study Rationale and Survey Description v Executive Summary ix Abbreviations and Acronyms xx Policy Note No.1: Service Delivery in Basic Schools in Four States of Sudan Section 1: Characteristics and Resources of the Average Basic School 2 Section 2: Key Findings and Policy Implications 19 Policy Note No.2: Student Learning in Four States of Sudan: School Inputs and Characteristics of Teachers Section 1: Learning Framework 24 Section 2: Factors Influencing Student Performance 25 Section 3: Results and Discussion 34 Section 4: Key Findings and Policy Implications 37 Policy Note No.3: Challenges and Possibilities for Teacher Management in Four States of Sudan Section 1: Strengths and Challenges of Teacher Management in Sudan 42 Section 2: Teachers and the Teaching Environment in Four States 45 Section 3: Summary and Policy Options 56 Policy Note No.4: Assessing the Role of Education Councils in School Management in Four States of Sudan Section 1: School-Based Management Models 61 Section 2: Characteristics of Education Councils in the Four Survey States 62 Section 3: Key Findings and Policy Implications 72 Policy Note No.5: Challenges in Basic Education Provision for Internally Displaced Children in North Darfur Section 1: Overview of IDPs in Darfur 76 Section 2: Comparing IDP and non-IDP Basic Schools in the Four Survey States 77 Section 3: Key Findings and Policy Implications 88 Annexes 91 Annex 0.1: Sample Characteristics 91 Annex 2.1: Description of Quality Determinant Variables 93 Annex 2.2: The Model Student Learning Function 94 Annex 4.1: Selected International Experience with School-Based Management 95 Annex 5.1: Basic Education Indicators for North Darfur State 97 Annex 5.2: Determinants of Grade 5 Student Performance in Math 98 Bibliography 99 iii Study Rationale and Survey Description The government of Sudan requested World Bank support to conduct the second Basic Education Service Delivery Survey (SDS2) to assess the availability and use of resources in basic education and student learning outcomes. The survey findings were needed for the 2012-16 Education Sector Strategic Plan, under development when these papers were released. The government also wanted to strengthen its capacity to administer future education service delivery surveys and student learning assessments, not least ahead of a national assessment of student learning planned for the 2012/13 school year. The Status of the Education Sector in Sudan report (World Bank, 2012) identified several key challenges in basic education in Sudan: inequity in access to and quality of basic education, inefficient teacher deployment, poorly prepared teachers with inadequate on-the-job training, insufficient and inefficient spending in the education sector, and indications of weak student learning outcomes. Following the request of the Ministry of General Education (MoGE) the SDS2 therefore focused on providing more information on the following five key areas: (i) resource availability and use in basic schools and rural/urban differences; (ii) the characteristics and resources of schools for internally displaced persons (IDPs); (iii) the role of education councils; (iv) teacher characteristics, deployment, and management in basic schools; and (v) grade 5 student learning levels and determinants. The SDS2 builds on a previous education service delivery survey, the SDS1, conducted in 2009 in three states (Kassala, River Nile and North Kordofan). The SDS2 extends the analysis of service delivery and student learning in Sudan to four additional states selected by the government: Blue Nile, North Darfur, Red Sea and South Kordofan. Unlike the SDS1, which was based on a convenience sample due to time and cost constraints, the SDS2 is based on a representative sample. 1 A steering committee was established by the MoGE to provide overall guidance for the SDS2. The ministry also appointed representatives from each surveyed state to be trained on conducting surveys of this type and to build institutional capacity in administering learning assessments. Data collection and entry was conducted by the Sudanese Organization for Educational Development, a local nongovernmental organization, alongside ministry representatives. 1 The 2008 Baseline Survey of Basic Education was conducted by the MoGE with support from the European Commission, UNICEF and the World Food Programme to assess the overall status of basic schools in all 15 states in Sudan in 2006/2007. Overall, it found that school infrastructure was weak and a shortage of resources in basic schools. v The SDS2 was conducted between December 14, 2010 and January 9, 2011, prior to the independence of South Sudan in July 2011. The inclusion of North Darfur as one of the survey states also provided a challenge for the fieldwork given the unstable security situation in the region. Sample Selection and Characteristics The survey covers the states of Blue Nile, North Darfur, Red Sea and South Kurdufan, which were chosen for the following reasons: (i) They provide examples of states that differ in terms of their education inputs, outputs, and outcomes: Retention is low in Blue Nile, North Darfur and Red Sea at 40 percent, 56 percent and 31 percent respectively, but is high in South Kordofan at 92 percent. In terms of per student spending, Blue Nile and Red Sea spend SDG 304 per student annually compared to SDG 201 for South Kordofan and SDG 171 for North Darfur. Basic Education Indicators in the Four Survey States, 2009 or MRY Blue North Red South Nile Darfur Sea Kordofan Sudan Number of Students Enrolled 122,786 346,779 110,707 273,102 4,870,464 Gross Enrollment Ratio (%) 65 67 37 82 72 Gross Intake Rate in Grade 1 (%) 81 80 66 80 80 Grade 8 Completion Rate 32 44 20 73 54 Retention Rate (%) 40 56 31 92 68 Per Student Spending (SDG) 304 171 304 201 240 % of Girls in Total Enrollment 44 46 44 45 46 % of Students Enrolled in Nomadic Schools 6 5 5 2 3 % of Students Enrolled in IDP Schools 0 18 0 0 0 Source: World Bank, 2012. Red Sea state is one of the weaker performing states in terms of the selected basic education indicators. 2 The gross intake rate in Grade 1 is 66 percent, compared to 80 percent on average for Sudan, and completion is very low at 20 percent, compared to the national average of 54 percent. 2 There are issues with the accuracy of population data for Red Sea state, which may affect the basic education indicators (World Bank, 2012). vi (ii) South Kordofan and Blue Nile border with South Sudan, which has been given special attention by the government with the signing of the CPA; and (iii) North Darfur was included on the request of the MoGE given the nature of ongoing conflict and large community of internally displaced people (IDPs). 3 The SDS2 focuses on government basic schools, which constitute approximately 97 percent of all basic schools in Sudan. The different school types in Sudan are: (i) regular (Grades 1 to 8); (ii) village (Grades 1 to 4); (iii) nomadic (typically only Grades 1 to 4 but there are exceptions, as some urban nomadic schools are being made permanent); and (iv) IDP schools (examined separately in Policy Note No.5 on IDP students). In the sample, regular schools form by far the largest group (89 percent), followed by nomadic schools (7 percent) and village schools (3 percent). The survey is designed to be representative of the four states and the findings represent the characteristics of an average government basic school in the four states, but are not representative at the individual state level. The survey design also enables the comparison of rural and urban average school characteristics, which is important in the context of Sudan due to substantial differences across the two. A multi-stage sampling strategy was used to select a representative sample. The lists of localities and schools constituting the sampling frames were generated by the World Bank team based on information collected from the state education ministries in the four survey states. At the first stage, localities were selected using probability proportional to size sampling where the measure of size was the number of schools in each locality and 30 to 50 percent of localities in each of the four survey states were selected. At the second stage, 68 schools in each sample locality were selected using the same sampling method where the measure of size was the number of students. At the final stage, 20 Grade 5 students and five teachers were selected in each sample school using simple random sampling. If a school had fewer than 20 Grade 5 students, all Grade 5 students were tested. Similarly, for teachers, if a school had fewer than five teachers all teachers were interviewed. During the SDS2, 253 of the 272 sample schools were visited (93 percent), which constitute about 9 percent of all government basic schools in the four states. All 68 sample schools in both Blue Nile and South Kordofan, 65 in Red Sea and 52 in North Darfur were visited. The number 3 Fieldwork in North Darfur was significantly more costly and provided substantial challenges due to the volatile security situation. There are 65 schools in the North Darfur state, of which 39 are IDP schools. Of these, 25 were selected for the sample; however, six of the schools could not be visited due to violence. Consequently, 19 out of the 39 IDP schools (49%) were surveyed. vii of schools visited is lower for North Darfur because several sample schools could not be visited for security reasons. The survey tested 4,542 Grade 5 students (an average of 18 students per school), interviewed 253 head teachers and 1,118 teachers (an average of 4.4 per school) and classrooms were observed in 251 schools. 4 If IDP schools are excluded, 234 of 247 (non-IDP) sample schools were visited (95 percent). Details on the samples used for each policy note are provided in each. The SDS2 used four questionnaires: (i) A head teacher questionnaire that captures information on school location, type, infrastructure, enrollment, head teacher characteristics, supervision, teachers, financing and education councils; (ii) A classroom questionnaire based on direct observations that recorded classroom resources, student attendance and availability of chalkboards, desks, and learning materials in Grade 1, Grade 5 and Grade 8 classrooms; (iii) A teacher questionnaire with information on teachers’ academic qualifications, training, appointments and transfers, supervision, salary, attendance, other jobs, leave and teacher unions; and (iv) A student questionnaire and learning assessment that gathers information on Grade 5 students’ characteristics and learning levels. 4 Not all schools offer all grades; in some, only Grade 1, Grade 5 or Grade 8 classrooms were observed. viii Service delivery and student learning in basic schools in four states in the Republic of Sudan? – Executive Summary Background and Objective Enrollment in basic education in the Republic of the Sudan has risen continuously since 2005. However, service delivery in basic schools and student learning outcomes are generally weak (MoGE 2008; World Bank 2012). The papers included in this book use survey data to describe how basic schools function in four states of the Republic of the Sudan. The main objective is to provide representative data on basic school resources and student learning levels to allow targeting of resources to priority areas and increase the efficiency of resource use in order to promote student learning. Source: World Bank. ix The source of data is the second Basic Education Service Delivery Survey (SDS2) that covered 253 schools in the 2010/11 academic year across the states of Blue Nile, North Darfur, Red Sea and South Kordofan in Sudan. These states represent three different regions: North Darfur in the western part, Blue Nile and South Kordofan in the Nuba mountain area in the south and Red Sea in eastern Sudan. All four states were affected by one of the several pre 2005 and 2006 peace agreement conflicts. Basic education is the fundamental cycle of education in Sudan, combining primary and lower secondary education into one eight-year cycle. Enrollments in basic schools have surged in recent years, from 3.3 million students in 2000/01 to 4.9 million in 2008/09, with particularly rapid enrollment growth in post-conflict states, including the four states examined in this book. This book complements the Education Sector Status Report published in 2012. 5 It contributes to the knowledge of basic education in the four states by providing a systematic account of service delivery and student achievement in basic schools. These two reports, along with other analytical work 6, are being used to inform the preparation of a new national Education Sector Plan for 2012-16. Key Topics Discussed Key contributions of this book include the measurement of learning outcomes, comparisons between urban and rural schools, information on the role of the education councils in school financing and specific data on schools for internally displaced persons (IDPs) in North Darfur. For the four states, the information provided in this book can help answer such questions as: (i) Are students learning in school? (ii) Are basic schools adequately resourced? (iii) Are basic schools well managed? (iv) Are teachers well prepared and motivated, and put to good use? (v) Are children in IDP camps being adequately served? 5 In cooperation with the Government of National Unity, the World Bank recently published a sector-wide Education Status Report, The Status of the Education Sector in Sudan. The report provides details on school participation, student flows, teacher policies, the financing of education, the management of resources and equity, for all levels of schooling. It reviews both national trends and state-level performance, given the prominent role of Sudan's 15 states in the provision of education. The status report also describes the results of the first Service Delivery Study, which provides a foundation for the work reported here. 6 Here you can reference the higher education work, etc. x The book's key findings in each of these five areas are presented below, and discussed in detail in the five papers that follow. Are students learning? Test scores in mathematics among Grade 5 students are very weak. On average, students tested in mathematics across the four states only answered 28 percent of questions correctly. To put this in perspective, the test was composed of multiple choice questions that each had four possible answers, so a randomly completed questionnaire would be expected to obtain a score of around 25 percent. The test questions were from the Trends in International Mathematics and Science Study (TIMSS), designed for Grade 4 pupils. A comparison of the results from the four states with results from Morocco and Tunisia shows that average test scores in Grade 5 in the four states of Sudan are roughly comparable with those of Grade 4 pupils in Tunisia, but not as good as those of Grade 4 pupils in Morocco. However, given that the test was administered in Grade 5 in the four states in Sudan, this likely means that the four states are well behind both Tunisia and Morocco in terms of primary school mathematics achievement. The first service delivery study found similar results for reading. 7 Thus, most students in Sudan are not learning enough, by any measure. Test scores were generally lower in rural areas, for children from poorer backgrounds and for boys. Not surprisingly, urban students had higher scores than rural students (averages of 28.7 percent and 27.1 percent) and students from richer households achieved better scores than students from poorer households, in both urban and rural areas. Perhaps more surprisingly, girls in both urban and rural areas obtain better test scores than boys (with respective averages of 29.2 percent and 28.3 percent in urban areas and 27.6 percent and 26.6 percent in rural areas). These and other results from the assessment of mathematics are presented in Policy Note No.2 (Student Learning). The note also presents the correlates between test scores and various characteristics of the learning environment, and find that school closures, poor infrastructure, less experienced teachers and a greater lack of textbooks are associated with worse student performance on the test. The student assessment shows that basic schools are not performing as well as they should. This has implications for all levels of education, as students who are not performing at the basic level are unlikely to perform well at higher levels of education. Improving service delivery in education and improving student learning outcomes therefore ought to be one of the main areas 7 The second Service Delivery Study also administered a reading test, but the results did not correlate well with the variables typically expected to be good predictors of reading scores, suggesting either that the test was not appropriate or that it was not administered correctly. xi of education policy focus in Sudan. This requires an understanding of how schools function, why service delivery is failing and of the critical constraints to improving learning that can be addressed by public policy in a resource-constrained environment. Are schools adequately resourced? Overall, pupil-teacher ratios are quite acceptable, but teacher allocation favors urban schools. The study found average pupil-teacher ratios of 36:1 in rural schools and 32:1 in urban schools in the four states (compared with a national average of 33:1, based on education statistics for 2009). Rural schools typically have lower enrollment and greater difficulty in optimizing class sizes. Rural schools would therefore be expected to have slightly lower pupil-teacher ratios than urban schools. In fact, the opposite is true in the four states. This suggests that rural schools are at a disadvantage in terms of getting their fair share of teachers. The majority of teachers both at rural and urban schools are official (government appointed - 84 percent). In terms of other resources available at the school level, schools rely on parental contributions to keep them running. In Sudan, responsibility for the operation of public basic schools is decentralized to states and mahalyas (localities). Most of the public funding for basic education comes in the form of federal education transfers to the states. These are mainly used to pay salaries. A significant share of the non-salary operational costs are funded by parents or other members of the community, often through the school’s education council. More than half of councils report providing funding to schools, and more so in urban areas. The main uses for education council funding are school feeding (38% of urban councils and 33% of rural councils pay for school feeding), furniture and learning materials and teacher salaries and training. However, the fees collected by school councils may be a barrier to access school. The involvement of parents and communities in school financing and management may therefore be problematic for the poor, while being helpful and affirmative of parents' interest in education. Among the 95 percent of basic schools that have education councils, 63 percent reported charging registration fees, averaging SDG 10-13 per student each year. Of the five percent of schools that do not have education councils, just 48 percent reported collecting school fees, averaging SDG 9 per student per year. As per official policy, basic schooling is free in Sudan, so these fees are informal. Basic schools also receive funding from NGOs, businesses, religious institutions and local communities. In urban areas, 15 percent of basic schools receive financial support from the local community (apart from education council funding), 12 percent from NGOs, 5 percent from businesses and 2 percent from religious institutions. In rural areas, 15 percent of schools in the four survey states receive funds from the local community, 3 percent from NGOs, 2 percent from businesses and 1 percent from religious institutions. xii Despite both public and private funding, the resources available to basic schools are insufficient. As documented in Policy Note No.1 (Service Delivery), service delivery indicators are weak and below minimally acceptable levels in many, if not most, schools. Further, resource shortages are more severe in rural than in urban schools: (i) School infrastructure in Sudan is generally poor or missing: 12 percent of rural schools have no school building and students sit under trees. Another 13 percent of rural schools have a roof but no walls. Other studies have found that many schools are inaccessible during the rainy season. 8 (ii) Many schools lack drinking water and toilets, particularly in rural areas. Half of rural schools and 42 percent of urban schools have no source of drinking water. (iii) There are severe shortages in schools of basic inputs such as desks, usable chalkboards and textbooks. These shortages are particularly severe in rural schools and in the early grades. Usable chalkboards, for example, are missing in Grade 1 classrooms in 39 percent of urban schools and 47 percent of rural schools. Grade 5 classrooms in both urban and rural schools are a little better endowed. (iv) The availability of textbooks is highly insufficient in most basic schools in the four states. In urban schools, about two thirds of children in both Grades 1 and 5 have no mathematics textbook. A similar result was found for reading textbooks. In rural schools, as many as three quarters of Grade 1 students have no textbooks, compared with two thirds of Grade 5 students. Public spending on education has increased rapidly between 2000 and 2009, with most of the new funding being managed at the state level, as shown in the Education Status Report (World Bank/GoNU 2012). Given the survey results reported here, it appears that salary expenditures are largely reaching schools, but that there is either insufficient non-salary spending available from public sources, or there are bottlenecks hindering that spending reaches schools. Further, non-public spending such as that provided by education councils clearly does not compensate for the funding gaps in schools and is perhaps not sufficiently directed at essential school inputs (chalkboards, textbooks and so on). The result is that schools have reasonably generous pupil-teacher ratios and offer school lunches, but students have no textbooks and classrooms have no chalkboards. Are schools well managed? The survey documents several aspects of school management, which has to do with how well schools are leveraging available resources and transforming them into quality instruction and 8 World Bank/GoNU. The Status of the Education Sector in Sudan, 2012. xiii learning: the survey documents such things as the availability and role of schools' education councils, the frequency of school closures and aspects of school record keeping and the administration of resources. On a positive note, education councils are widespread and a central feature of Sudanese basic schools. Education councils are found to play a central role in providing administrative support and coordinating community financing to schools. As many as 96 percent of schools have active education councils (99 percent of urban schools and 94 percent of rural schools). Education councils consist of parents, teachers and community members and serve several purposes, largely similar across rural and urban schools: contribute funds, help with administrative duties, follow up on drop-out students and provide inputs to locality inspectors’ teacher supervision reports (See Policy Note No.5 on Education Councils). Role of Education Councils 70 63 60 Urban 48 Rural 50 44 45 42 43 Percent 40 30 27 23 20 14 9 10 0 Funding Administration Follow-up Supervision Teaching on Drop-outs Source: SDS2 data. Overall, however, school management was found to be weak, resulting in the loss of instructional time and poor record-keeping of student and teacher attendance: (v) School closures account for a considerable loss of instruction time. The average length of the school year reported by sample schools was 171 days, or 39 fewer days than the official length of the school year. School closures were more frequent in South Kordofan and Blue Nile states, perhaps because of their long rainy seasons, and in schools without a roof or walls. Similarly, urban schools generally operate more days per year (173) than rural schools (165). School closures thus appear to reflect a combination of school infrastructure, climate and school management. xiv (vi) School records are often out-of-date, where kept. A relatively large share of urban schools, 82 percent, has up-to-date school enrollment records. By contrast, only 47 percent of rural schools have up-to-date enrollment records and 35 percent keep no records at all of student enrollment. When it comes to monitoring student performance through continuous assessments, 42 to 58 percent of rural and urban schools have up- to-date records. But 41 percent of rural schools and 27 percent of urban schools do not keep any such records, making it difficult to target resources at weaker schools, and within schools, at students who are under-performing. (vii) Teacher absenteeism is rarely monitored. Most rural schools (77 percent) do not keep any records on teacher leave compared to 35 percent of urban schools. Among the rural and urban schools that do keep records of teacher leave, the records were up-to-date in only 16 percent of rural schools and 35 percent of urban schools. Are teachers well prepared and motivated, and put to good use? The average teacher is female and government appointed, but rural schools have a considerable number of paid volunteer teachers, perhaps as a result of teacher shortages. Most teachers in Sudan's basic schools are female, particularly in the urban areas. The majority of teachers in the basic schools are official, government appointed teachers. Rural schools report greater shares of paid volunteer teachers (12 percent) than urban schools (4 percent), perhaps as a result of more frequent teacher shortages in rural areas. Volunteer teachers are typically paid by education councils. Only urban schools seem to operate with subject specialization, which results in larger class sizes. Urban and rural schools seem to use teachers quite differently. Urban schools have more teachers relative to enrollment (urban schools in the four states have an average student-teacher ratio of 32:1 compared with 36:1 for the rural schools), but urban schools organize their students into larger classes and organize the instruction by subject-specialized teachers, at least in part. Rural schools, which make up around two thirds of all basic schools in Sudan, run with smaller class sizes and do not appear to use subject-specialization to the same extent. The average rural school has 1.1 teachers per class relative to 1.8 teachers per class in the average urban school, despite the fact that class sizes averaged 56 in the urban schools and only 40 in rural schools (See Figure below). More information on these patterns is included in Policy Note No.3 (Teacher Management). xv Class Size and Student-Teacher Ratio (School Averages) 60 56 Urban Rural 48 50 Number of Pupils 40 36 40 40 32 30 25 20 10 — 0 Regular Regular Village Nomadic Average STRs Class Size Source: SDS2 data. Teachers' background generally includes 10 to 12 years of formal education and another 10 years of teaching experience, but few have professional training. The vast majority of rural and urban teachers surveyed in the four states (nearly 90 percent) have 10 to 12 years of formal education, which is generally deemed sufficient to teach basic education. Further, many teachers are highly experienced (the average teacher has 15 years of experience in urban schools and 10 years in rural schools). On the other hand, more than 70 percent of teachers report not having received pre-service teacher training. The more experienced teachers appear to have more appropriate classroom practices. The survey contained questions on whether teachers give their students exercises to complete in class as opposed to just having students read or copy from the chalkboard. The expectation is that teachers who provide exercises generally spend more time preparing their classes and interact more with their students. One in five students reported that their teacher provided them with exercises to complete in class. Pedagogical practice seems to be related to experience as in both urban and rural schools, more experienced teachers are significantly more likely to give their students exercises. Are children in IDP camps being adequately served? IDP schools serve a considerable proportion of the population in some parts of Sudan. The study surveyed a sample of schools located in camps for Internally Displaced Persons (IDPs) in North Darfur, where they account for 18 percent of basic school enrollments. A separate analysis of IDP schools is provided in Policy Note No.5 (IDPs in North Darfur). xvi Despite evidence of past support to IDP schools, these schools are currently overwhelmed by student numbers, resulting in weak service delivery. IDP schools have benefited from assistance from aid organizations and as a result, tend to have better school buildings than average for Sudan, often including a roof, walls, access to water and latrines. Their teachers, although few in numbers, tend to have higher academic qualifications than average, although they are often not trained as teachers. All of the surveyed IDP schools have an active education council, and these are more often involved in financing the school than councils in regular schools. On the other hand, the survey clearly shows that IDP schools are overwhelmed by very large student enrollment, for which they do not have the resources. Thus, on average, IDP schools in North Darfur have much larger student enrollments (nearly 900), higher student- teacher ratios (47:1) and larger class sizes (76) than other schools included in service delivery study. Also, they are severely lacking textbooks and basic classroom equipment, such as desks and chalkboards. Further, IDP schools were found to be open only 164 days a year on average, compared with 171 days per year in other schools in the survey. Average scores in the Grade 5 math test in IDP schools were below the levels recorded in other schools, even when controlling for household characteristics, such as income and urban/rural location. The deficits in the learning environment may not be the only explanation for this. Ethnic minorities are often the most likely to inhabit IDP camps, and if children's first language is different from the language of instruction, they will likely benefit little from schooling. The low performance of IDP schools underscores their need for additional resources. Although schools have benefitted from capital expenditure for school construction, schools appear to be too few, leading to over-crowding. Also, they seem under-resourced in terms of recurrent funding for their operations, including for teacher salaries and teaching and learning materials. Although most education councils in IDP schools report providing funding to schools, it is clearly insufficient, namely given the textbook shortage. Conclusions and Next Steps The SDS2 survey showed that overall, students in basic schools are not learning what they should. This is true for most children in both urban and rural areas, boys and girls and children from different socio-economic backgrounds. This leads to question what steps can be taken to begin a gradual improvement of the quality of schools and instruction so that students learn more. xvii Establish a national system of student assessment and a feed-back loop A key building block of an education system geared towards quality is to have a national assessment system that regularly measures, most commonly through sample-based national assessments, the ability of the school system to produce learning outcomes. Such a national assessment system must also feed the results back to policy makers and school system administrators, those responsible for developing school curricula and teacher training colleges. When choosing test instruments for national assessments, policy makers and test administrators may want to consider applying tests that are developed specifically for developing education systems, and that measure the extent to which students have acquired the foundational reading and numeracy skills that are needed for more advanced learning. Since the measurement of learning requires using scientific methods, policy makers may consider involving universities in national assessment work. Better articulate the minimum standards for schools and identify the most cost- effective service delivery options Based on the information provided in this book, it is clear that many schools are lacking either infrastructure or basic school inputs and some have class sizes or pupil-teacher ratios that exceed acceptable levels. International evidence shows that bringing all schools up to a minimally acceptable standard is helpful in school systems that are at this level of development. For this to happen, a first step is to define realistic minimum service delivery standards for basic schools, against which funding formula could be developed and against which states and localities can be compared and progress measured. This can also help bring clarity to the question of whether Sudan is spending enough on basic schooling and where the gaps are. It is already clear is that current expenditure is not equitably allocated across schools, localities and states. When developing these standards, it is opportune to think about the types of school models that work well in different settings, to find cost-effective options of delivering services in a context of limited resources. To some extent, Sudan has already developed different school models, including the small village schools, but more can be done. Unfinished business: increase basic education access and completion The focus on delivering better learning does not mean Sudan should forget about the Millennium Development Goals of universal primary education and gender equity. In fact, the learning agenda can help improve access, as improved schools are more attractive to students and parents xviii and can help improve attendance and reduce dropout. For some children, however, particularly those from very poor families, additional incentives may be needed, such as take home rations. Leveraging education councils for stronger accountability International evidence shows that merely increasing spending on education will likely not be enough to improve service delivery. More resources would be helpful, but more importantly, there is a need for greater accountability in the use of resources and the achievement of results at all levels. Parental or community involvement in school management has the potential for strengthening accountability at the school level. Sudan's education councils can potentially play this role, but the current low learning results suggest that education councils are not very effective at the moment. Education councils will need to be strengthened and empowered to have greater influence in school and teacher affairs, and they may also need to be sensitized and educated in what constitutes a good school (since they often choose to fund school lunches, but not textbooks). At the same time, the findings suggest that a lack of essential school records make it difficult for education councils and communities to effectively engage in school management. xix Abbreviations and Acronyms CPA Comprehensive Peace Agreement DHS Demographic and Health Survey EFA Education for All IDPs Internally Displaced Persons MoGE Ministry of General Education MRY Most recent year with data available PTA Parent Teacher Associations PTBR Pupil-Textbook Ratio, or Student-Textbook Ratio SBM School-based Management SDG Sudanese Pound SDS2 2010/11 second Basic Education Service Delivery Survey STR Student-Teacher Ratio xx Policy Note No.1 Service Delivery in Basic Schools in Four States of Sudan - Findings from the 2010/11 Basic Education Service Delivery Survey Abstract Enrollment in basic education in Sudan has risen continuously since 2005. However, service delivery performance and student learning outcomes are generally weak. The second Basic Education Service Delivery Survey was conducted in four states (Blue Nile, North Darfur, Red Sea and South Kordofan) in 2010/11 and provides representative data on resource availability and use in basic schools and differences across rural and urban schools. Overall, the average government basic school in the four survey states lacks resources and resource shortages are more severe in rural than in urban schools. They impact directly on: (i) School infrastructure: students sit under a tree in over a tenth of rural schools; a substantial share of schools have no source of drinking water or latrines; and there are shortages of desks, usable chalkboards and textbooks in many schools. (ii) Learning environment: the average actual school year is much shorter than the official school year, indicating a substantial loss of instructional time. Furthermore, more than 70 percent of teachers have no pre-service teacher training. (iii) Resource allocation and use: there are inefficiencies in the use of existing resources, in particular in the allocation and use of teachers. Differences in student-teacher ratios across rural and urban schools and the relatively large shares of volunteer teachers in many schools reflect teacher management policies that frequently do not allocate teachers based on school enrollment and needs. School record keeping is weak in many schools, which makes the targeted allocation and more efficient use of resources a challenge. Given the resource constrained environment in the four survey states, a combination of additional, targeted investment and more efficient use of existing resources is required to improve service delivery and student performance. 1 Introduction Enrollment in basic education in Sudan has risen continuously since the signing of the 2005 Comprehensive Peace Agreement. However, service delivery performance in basic education and student learning outcomes are generally weak. 9 Poor performance is driven both by supply and demand-side factors. This note focuses on the supply-side of the basic education system: the inputs necessary to provide education services of acceptable standards to school age children in Sudan. 10 The main objective of the second Basic Education Service Delivery Survey (SDS2) is to provide representative data on resources in basic schools of the Blue Nile, North Darfur, Red Sea and South Kordofan states, in 2010/11, to: (i) inform the 2012-16 Education Sector Strategic Plan (ESSP); (ii) address identified supply-side challenges by targeting resources to priority areas (school infrastructure, supplies and learning materials); and (iii) increase the efficiency of resource use especially with regard to teacher allocation and training (See Teacher note). Section 1 of this note presents and discusses the characteristics and resources of the average rural and urban basic schools in the four states. Section 2 summarizes the priority areas based on the findings and outlines possible next steps to address the identified weaknesses in basic education service delivery. Section 1: Characteristics and Resources of the Average Basic School in the Four Survey States School infrastructure and the learning environment directly affect student learning. If schools lack the minimum necessary resources such as appropriately trained teachers or textbooks, little learning will take place. Even when resources are available, how they are managed also influences learning outcomes. This section provides a snapshot of resource availability and when possible resource use, at the average basic school in the four survey states. 9 Basic education is eight years long in Sudan and the official school starting age is six years. 10 Quality as such is not measured, so the analysis focuses on the combination of school inputs generally considered necessary to provide a quality education. 2 School Location and Infrastructure One key determinant of school characteristics and resources is location: rural or urban. Therefore, to gain a more precise understanding of differences in resource availability across basic schools in the four states, rural and urban schools are compared throughout this section. Table 1.1 presents selected education indicators for average basic rural and urban schools in the four states. The differences across rural and urban schools are substantial: urban schools are generally much larger, with 462 students on average, compared to 230 students for rural schools. They also have larger class sizes, of 56 pupils on average, compared to 40 students in rural schools, although the student-teacher ratio is lower. Urban schools also have proportionately more girls enrolled than rural schools, although fewer girls than boys are enrolled in both. The vast majority of schools have education councils, both rural (94 percent) and urban (99 percent). Table 1.1: General School Statistics, by Location (School Averages), 2010 Urban Rural Number of Students 462 230 Student-Teacher Ratio (All Teacher Types) * 32:1 36:1 Class Size (Number of Students) * 56 40 Gender Parity Index (Co-ed. Schools) ** 0.84 0.78 Share of Schools with an Education Council (%) 99 94 Source: SDS2 data. Note: * Regular rural schools only. The different school types in Sudan are: regular (Grades 1-8), village (Grades 1-4), nomadic (typically only Grades 1-4) and IDP schools ** North Darfur is effectively not included in this indicator, as it has no non-IDP co-education schools. Distance to School When schools are far from children’s homes the probability of enrollment drops and the risk of drop-out increases. In remote areas in Sudan, many students have to travel far to reach school, according to the SDS2 survey. In the four states just over half of Grade 5 students live within 30 minutes of school (48 percent in rural areas and 59 percent in urban areas). Conversely, a majority of rural students live over 30 minutes away from the nearest school and 17 percent live over an hour away. In urban areas the situation is less problematic, although still 41 percent live over 30 minutes away and 12 percent over an hour away (See Figure 1.1 below). 3 Figure 1.1: Distribution of Grade 5 Students, by Time to School and Location, 2010 Percent 70 59 Rural 60 48 Urban 50 35 Percent 40 29 30 17 20 12 10 0 Less than 30-60 More than 30 minutes minutes 60 minutes Source: SDS2 data. School Infrastructure Even when schools are nearby, student learning is adversely affected if schools are in poor condition and classrooms are decrepit or overcrowded. The status of school infrastructure in Sudan is generally poor. A 2008 basic education survey found that about 50 percent of classrooms in Sudan needed to be rehabilitated or completely replaced (MoGE, 2008). 11 For the SDS2 survey states, except the Red Sea state where roughly 33 percent of classrooms were estimated to require rehabilitation or replacement, the situation was even worse than the national average: 59 percent of classrooms were in inadequate condition in Blue Nile, 61 percent in North Darfur and 68 percent in South Kordofan. The 2008 basic education survey did not provide a breakdown by rural-urban location or examine the construction materials used for schools. The SDS2 survey on the other hand does. It includes questions on school roofs and wall materials, finding substantial variations between rural and urban schools in terms of building materials used and the existence of roofs and walls (See Table 1.2). 11 The 2008 Baseline Survey on Basic Education was conducted by the MoGE with support from the European Commission, UNICEF and the World Food Program to assess the overall status of basic schools in all 15 states in Sudan in 2006/2007 (MoGE 2008). Overall, that survey found that school infrastructure was weak and a shortage of resources in basic schools. The SDS2 complements the Baseline survey and provides more detailed information on several of the key issues in basic education as identified by the 2008 Baseline Survey, the ESR and SDS1. Moreover, the SDS2 provides information for additional areas such as supervision and head teachers. The SDS2 also includes a student learning assessment unlike the 2008 Baseline Survey. 4 The main roof materials for rural schools are zinc (48 percent) and thatch/grass/mud (32 percent), but in roughly 12 percent of schools students sit under a tree. For urban schools zinc is also the most common roof material (72 percent), followed by thatch/grass/mud (17 percent) and bricks/cement (6 percent). Notably, all urban schools have a roof. In terms of walls, 13 percent of rural schools have no walls, 42 percent have walls made of bricks/blocks and 26 percent have thatch/grass/mud walls. Again, there is a large difference compared to urban schools, only 3 percent of which have no walls and the large majority of which, 80 percent, have walls made of bricks/blocks. Table 1.2: Condition of School Infrastructure, by Location, 2010 Percent Urban Rural Roof Material Wood 0 4 Thatch/Grass/Mud 17 32 Bricks/Cement 6 2 Zinc 72 48 Other 6 2 None (Sit Under Tree) 0 12 Wall Material Wood 0 2 Thatch/Grass/Mud 15 26 Bricks/Blocks 80 47 Stone 2 8 Tree Branches 0 2 Other 0 2 No Walls 3 13 Source of Drinking Water Tap 15 1 Well 13 14 Borehole 0 3 River 2 2 Pump 12 20 Other 16 10 No Drinking Water 42 50 Latrines Schools with Latrines 81 63 Share of Operational Latrines 86 90 Source: SDS2 data. 5 Another challenge is the lack of drinking water in many schools. Tap and pump water is generally considered safe to drink, as well as water from wells, when they are covered. In the four states, 50 percent of rural schools and 42 percent of urban schools have no source of drinking water. Where schools have drinking water, the most common sources are pumps and wells in rural areas and taps and wells in urban areas, indicating that safe drinking water is more common in urban areas. 12 Only 63 percent of rural schools have latrines, against 81 percent of urban schools. This finding shows a deterioration in the availability of latrines compared to the 2008 basic education survey, according to which almost all basic schools had toilet facilities but the majority of these were reported to be of temporary nature or in need of repair. About 90 percent of existing latrines are operational, with no significant difference between rural and urban schools. These findings highlight the overall short-falls of school infrastructure in the four survey states and the large differences across rural and urban schools, underlining the importance of developing clear resource targeting strategies. Learning Environment A learning environment conducive to student learning depends on several factors on the supply- side including instructional time; the condition of classrooms; student-teacher ratios (STR); class size; the availability of chalkboards, furniture, textbooks and learning materials; and the availability of trained teachers. Each of these factors is discussed below based on information provided by head teachers, teachers or through direct observation of Grade 1 and Grade 5 classrooms in the four survey states. The Length of the School Year The length of the official school year ranges from 175 days to 210 days in developing countries. In Sudan, the school year is officially 210 days but the actual average school year is much shorter at 171 days. 12 The 2008 basic education survey did not ask for exactly the same sources of water as the SDS2 and provides no rural-urban breakdown but its findings on the lack of access to water are largely in line with those of the SDS2 (MoGE, 2008). 6 Figure 1.2: Length of School Year, by School Type and Location (School Averages), 2010 Days 250 Urban Rural 195 193 200 173 165 160 150 Days 100 50 — 0 Regular Village Nomadic Note: Some urban nomadic schools are being made permanent. Source: SDS2 data. Breaking down the data shows the variation in the length of the actual school year across rural and urban areas, by school types (See Figure 1.2 above). The average length of the actual school year is longer for urban schools than for rural schools. Nomadic schools have the longest effective school years, in both rural and urban areas, at about 194 days on average, approximately 25 days more than the average for regular schools. 13 Rural village schools are those where the actual school year is shortest, at 160 days on average. Box 1.1: State Differences in the Length of the School Year Examining the reported number of days schools are open by state reveals large differences across the four survey states. South Kordofan and Blue Nile states have the shortest average school years at 158 days and 161 days respectively. This may partly be explained by the long rainy seasons in these states that tend to make schools inaccessible and/or unusable (arguably due to weak school infrastructure) for long periods of time. The average school year is somewhat longer in North Darfur at 169 days and much longer in the Red Sea state at 206 days. The climate in the Red Sea state is hotter than in the other three states but this does not generally have an adverse impact on closure, as long as 13 It should be noted that the length of the actual school year is a crude measure of how much instruction students receive since even if schools are open teaching may not always be taking place. 7 Thus, for the average regular school the school year is shorter than the official school year by roughly 45 days in rural areas and 37 days in urban areas. This implies a considerable loss of instructional time with adverse implications for student learning. In some cases the loss of school days is due to adverse weather conditions making classrooms unusable due to the state of school infrastructure, primarily in rural areas. Student-Teacher Ratio and Class Size When the number of students per teacher is high, teachers have less scope to interact with and assist individual students. In 2005, the average student-teacher ratio (STR) was 45:1 for Sub- Saharan African countries and 22:1 in the Middle East and North Africa Region (UNESCO, 2007). The average basic school in the four survey states in Sudan had a lower STR (37:1) than the Sub-Saharan African average but higher than the Middle East and North Africa Region average. Examining basic schools by location, urban schools tend to have lower STRs than rural schools. The average STR for urban regular schools is 32:1 and for urban nomadic schools 25:1 compared to 48:1 for rural village schools, 40:1 for rural nomadic schools, and 36:1 for rural regular schools (See Figure 1.3). 14 Figure 1.3: Class size and Student-Teacher Ratios, by Type of School and Location (School Averages), 2010 60 56 Urban Rural 48 50 Number of Pupils 40 40 40 36 32 30 25 20 10 — 0 Regular Regular Village Nomadic Average STRs Class Size Source: SDS2 data. 14 The average STR for IDP schools in North Darfur is 47:1 (See Policy Note No.5 on Internally Displaced People). 8 The variations in STRs across average rural and urban schools are a reflection of teacher allocation policies and teacher subject specialization. For Sudan as a whole teacher allocation is considered to be ineffective in that the relation between teacher postings and student numbers displays a high degree of randomness (World Bank, 2012). There are also differences in the number of teachers per class across rural and urban schools. The average rural school has 1.1 teachers per class and the average urban school has 1.8 teachers per class, implying that teachers may be less specialized by subject in rural schools. Since 68 percent of basic schools are rural this indicates that subject specialization of teachers is low for the majority of the basic school system. 15 How conducive the classroom environment is to learning also depends on the number of students per classroom. Even when the student to teacher ratio is low, class size can be relatively large, providing a more difficult learning environment. In the four survey states, class sizes are generally larger in urban than in rural schools, in line with the higher population density in urban areas. The average urban regular school has an average class size of 56 students, compared to 40 students for the average rural regular school. Class sizes for rural village and nomadic schools are much smaller still, as is to be expected given the relatively remote location of these schools, at just 27 students and 25 students respectively. Classroom Furniture and Equipment Related to the issue of class size is student seating arrangements. In crowded classrooms where there are not enough desks (or no desks) for students it is more difficult for learning to take place. Generally, students in urban schools are more likely to have a desk than students in rural schools and students in higher grades are also more likely to have a desk (UNESCO 2007). Table 1.3: Classroom Furniture and Chalkboards (Observed Classrooms), by Grade and Location, 2010 Percent Urban Rural Share of Schools with Grade 1 61 53 Chalkboards in Good Condition Grade 5 72 64 Share of Students with Grade 1 39 21 the Use of a Desk Grade 5 70 49 Source: SDS2 data. Note: Only schools with Grade 1 and Grade 5 data for each variable are included. 15 For further details on teacher distribution, STRs and class size implications see Policy Note No.3 on teacher management. 9 In the four survey states students in urban schools and in higher grades are more likely to have a desk. In the average rural school, 21 percent of Grade 1 students have a desk, against 49 percent of Grade 5 students (See Table 1.3). The pattern is the same for the average urban school, where 39 percent of Grade 1 students are seated at a desk, against 70 percent of Grade 5 students. 16 In many developing countries, in particular in rural areas, the only textbook in the classroom belongs to the teacher and students copy material from the chalkboard, making it the most important teaching tool (UNESCO 2007). Among schools in the four survey states a larger share of urban schools has chalkboards in usable condition: 61 percent in Grade 1 and 72 percent in Grade 5, against 53 percent and 64 percent of rural schools, respectively. 17 The lack of usable chalkboards in many schools in the four survey states, when combined with a lack of textbooks, presents a great challenge to effective teaching. Textbook Availability Textbooks are important to student learning and can reduce differences between the performance of rural and urban students as well as improve the performance of students from poorer backgrounds (Jamison et al. 1981; Lockheed and Hanushek 1988; Postlethwaite and Ross 1992). Although the official student-textbook ratio in Sudan is 2:1 (MoGE, 2008), many schools still have a shortage of textbooks and some lack them completely. In Grade 1 and Grade 5 classrooms in the four survey states, the student-textbook ratio is generally higher than the official benchmark, in both rural and urban schools. The ratio also tends to be higher in rural schools than in urban ones (See Table 1.4). For the average rural school, the availability of textbooks in mathematics and reading is higher in Grade 5 than in Grade 1: 25 percent of students in Grade 1 classrooms have a textbook compared to about 32 percent of students in Grade 5. In the average urban school, about 31 percent of students have a math textbook, whereas the share of students having a reading textbook ranges from 31 percent in Grade 1 to 36 percent in Grade 5. However, these averages conceal the fact that in many schools no students have a textbook at all: in 50 percent of Grade 1 and 15 percent of Grade 5 urban classrooms no math textbooks are available. The shortage is worse still in rural areas, where 53 percent of Grade 1 and 30 percent of Grade 5 classrooms have no math textbooks. The corresponding shares are roughly similar for 16 The 2008 education survey provides data on student seating for Grade 1 and Grade 8 students, with similar findings: nationally, 36 percent of Grade 1 students and 79 percent of Grade 8 students were seated at a desk (MoGE, 2008). 17 This is roughly in line with the findings of the 2008 basic education survey where at the national level about 40 percent of chalkboards in basic schools were in need of repair of replacement (MoGE, 2008). 10 reading textbooks. This starkly underlines the severe shortage of textbooks in many basic schools in the four survey states. Table 1.4: Availability of Textbooks and Learning Materials (Observed Classrooms), by Grade and Location, 2010 Percent Urban Rural Share of Students with Grade 1 32 25 a Math Textbook Grade 5 31 31 Share of Students with Grade 1 31 25 a Reading Textbook Grade 5 36 33 Share of Students with Grade 1 84 84 a Pen/Pencil Grade 5 89 85 Source: SDS2 data. Note: Excludes urban nomadic schools. Shortages are less severe for other learning materials, largely because education councils and communities meet the shortfall. The majority of students have pens or pencils: 84 percent of Grade 1 students, without any variation by location, and a slightly higher share of Grade 5 students (85 percent in rural and 89 percent in urban schools). Resource Allocation and Use In a resource constrained environment such as the Sudanese education sector, more effective use of existing resources is necessary to improve service delivery performance. Having a sufficient number of appropriately trained teachers and deploying these teachers efficiently is crucial to improve service delivery performance and student learning outcomes. Sudan does not generally suffer from a teacher shortage (the average national STR is 33:1, below the FTI benchmark of 40:1 for Sub-Saharan Africa) but the distribution of teachers across schools and the use of teachers within schools are often inefficient (World Bank, 2012). In the four survey states, 69 percent of teachers are female compared to the national average of 67 percent (World Bank 2012a).Figure 1.4 shows the distribution of teachers by employment type. The majority (84 percent) of teachers at both rural and urban schools are official teachers (government appointed). The second largest group is composed of volunteer teachers, who constitute 16 percent of teachers at the average rural school and 14 percent at the average urban school, overall. The share of volunteer teachers who are paid is much larger at the average rural school however (12 percent) against 4 percent in the average urban school). 11 Figure 1.4: Distribution of Teachers, by Employment Type and Location, 2010 Percent 90 84 84 80 Urban 70 Rural 60 Percent 50 40 30 20 10 12 10 4 4 1.7 0.5 1 — 0 Government Unpaid Paid National Trainee Contract Volunteer Volunteer Service Source: SDS2 data. Volunteer teachers’ training is not known; this is an issue as they may not be qualified to teach. The use of volunteer teachers may also be a sign that the number of official teachers being allocated to some schools is insufficient, so that communities have to hire and pay for volunteers. There are also some national service and trainee teachers in basic schools in the four states but they represent a minority (0.5 percent to 1.7 percent respectively). When it comes to teachers’ academic qualifications there are also differences across rural and urban schools. 45 percent of teachers at the average urban school have a BA/BSc and 41 percent at the average rural school (See Figure 1.5). 18 A relatively large share of teachers, 12 percent, at the average urban school has an academic diploma compared to only 4 percent for the average rural school. However, a larger share of teachers has a secondary education, 51 percent, in the average rural school than in the average urban school, 40 percent. These results are similar to those of the 2008 basic education survey but are not directly comparable given the lack of a distinction between rural and urban schools in that survey (MoGE 2008). 18 These results are similar to those of the 2008 baseline survey but are not directly comparable given the lack of a distinction between rural and urban schools in that survey (MoGE, 2008). 12 Figure 1.5: Distribution of Teachers, by Academic Qualification and Location, 2010 Percent Basic or Intermediate 45 2 Urban 40 12 Secondary 1 Education Non-Teaching Diploma 1 BA/BSc Rural 4 41 3 51 Postgraduate Degree 0% 20% 40% 60% 80% 100% Source: SDS2 data. In terms of pre-service training, one of the government’s interim EFA objectives was to raise the share of trained teachers to 80 percent by 2007. Nevertheless, in 2010 in the four survey states, 70 percent of teachers, be it in urban or rural schools had no pre-service teacher training (See Figure 1.6 below). Figure 1.6: Distribution of Teachers, by Pre-Service Training Qualification and Location, 2010 Percent Urban 21 9 None 70 Teaching Diploma Rural 7 23 70 Bachelor of Education 0% 20% 40% 60% 80% 100% Source: SDS2 data. 13 Interestingly, in the average rural school, a much larger share of teachers, 23 percent, have an advanced education degree (Bachelor of Education) than their urban counterparts (8 percent). 19 However, teachers in urban schools are more likely to have a teaching diploma (21 percent) than their rural counterparts (7 percent). These findings highlight the substantial differences in teacher distribution and characteristics across rural and urban schools. Finally, teachers at the average urban school have more work experience at 13 years compared to 9 years for teachers at the average rural school. The relatively larger share of teachers at urban schools with no B.Ed degree and the larger share with a teaching diploma may partly be explained by the fact that the B.Ed requirement was introduced in 1993 meaning that fewer urban teachers have, on average, been exposed to the new teacher training system. 20 Teachers in urban schools may also have less incentive to obtain a B.Ed degree compared to rural teachers who may want to obtain a B.Ed degree in order to be able to move to a more attractive urban teaching position. However, further study is needed to establish the main reasons for the relative imbalance in teachers with a B.Ed degree across rural and urban schools. Table 1.5: Head Teacher Characteristics, by Location, 2010 Percent Urban Rural Academic Qualifications Basic/Intermediate 2 4 Secondary 26 43 Non-Teaching Diploma 15 10 Bachelor's Degree 54 42 Postgraduate Degree 3 0.6 Teacher Training None 44 52 Teaching Diploma 48 35 Bachelor of Education 8 14 Average Work Experience (Years) 10 8 Source: SDS2 data. Head teachers are in charge of the daily running of basic schools, and their education, training and experience affect the delivery of education services. 21 Overall, their characteristics are similar to those of other teachers. Table 1.5 below presents information on their academic qualifications, pre-service training and work experience. In the vast majority of schools head 20 The pre-service qualification for basic education teachers was revised in 1993 from a two-year teaching diploma to a four year Bachelor in Education (B.Ed) degree (World Bank 2011). 20 The pre-service qualification for basic education teachers was revised in 1993 from a two-year teaching diploma to a four year Bachelor in Education (B.Ed) degree (World Bank 2011). 21 In two percent of nomadic and village basic schools head teachers are in fact the only teacher. 14 teachers have at least secondary education, although those working in urban schools tend to have higher academic qualifications than their rural counterparts. In rural schools, 42 percent of head teachers have a bachelor’s degree, 43 percent have completed secondary education and 10 percent hold a non-teaching diploma. By contrast, 54 percent of head teachers have a bachelor’s degree in urban schools, 13 percentage points more (26 percent have completed secondary education and 15 percent hold a non-teaching diploma). A large share of head teachers have no pre-service teacher training: 52 percent of rural head teachers and 44 percent of urban head teachers. Out of those who do have pre-service teacher training, 35 percent and 48 percent respectively of rural and urban head teachers have a teaching diploma and 14 percent and to 8 percent respectively have a B.Ed degree. The average duration of work experience is 8 years for rural and 10 years for urban head teachers. Record Keeping One indication of how schools are managed is the extent of record keeping. Accurate and up-to- date information on student enrollment and performance and teacher attendance and leave is an essential component in the management and targeting of school resources to improve service delivery (Lewis and Pettersson 2009). As part of the SDS2 survey, data collectors asked to examine various school records and noted whether these records were available and up-to-date. Table 1.6 below presents the results for rural and urban schools in the four survey states. A relatively large share of urban schools, 82 percent, has up-to-date school enrollment records. By contrast, only 47 percent of rural schools have up-to-date enrollment records and 35 percent keep no records at all of student enrollment. Although student performance in continuous assessments is tracked by 42 percent of rural and 58 percent of urban schools, a high proportion of schools keep no such records at all (41 percent and 27 percent respectively), making it difficult to target resources at weaker schools, and within schools, at students who are under-performing. Another signal of how school resources are utilized, in this case how teachers are managed is the availability of records on teacher leave. Most rural schools do not keep any records on teacher leave (77 percent, compared to 35 percent of urban schools). Among the schools that do keep such records, they were up-to-date in only 16 percent of rural schools and 36 percent of urban schools. 22 22 The 2008 baseline survey also examined record keeping and found that most schools kept attendance records and about 61 percent of schools nationwide kept records of student performance (MoGE, 2008). 15 Table 1.6: Share of Schools that Keep Records, by Type of Record and Location (Average Schools), 2010 Percent Urban Rural Enrollment Records Seen + Up-to-date 82 47 Seen 6 11 Not available 5 8 None kept 7 35 Student Performance Records Seen + Up-to-date 58 42 Seen 6 2 Not available 9 15 None kept 27 41 Teacher Leave Records Seen + Up-to-date 36 16 Seen 10 1 Not available 19 5 None kept 35 77 Source: SDS2 data. Overall, record keeping is weak in the four survey states, especially with regard to student performance and teacher leave. In addition, relatively more urban schools keep up-to-date records than rural schools. Community Involvement through Education Councils The involvement of communities and parents in the running of schools through education councils is encouraged by the Sudanese government (World Bank 2012a). The vast majority of schools (94 percent in rural areas and 99 percent in urban areas) have a council. These consist of parents, head teachers, teachers and sometimes members from the local community. The roles of education councils are largely similar across rural and urban schools. Ranked by order of importance, in rural schools 48 percent of councils contribute funds, 46 percent help with administrative duties, 43 percent follow up on students who have dropped out and 23 percent provide inputs to locality inspectors’ teacher supervision reports (See Figure 1.7 below). 16 Figure 1.7: Key Activities of Education Councils, by Location, 2010 Percent 70 63 60 Urban Share of Councils that: 48 50 43 46 43 43 Rural 40 28 30 23 20 15 9 10 0 Contribute Help with Follow up on Help Supervise Help with Funds Administration Drop-outs Teachers Teaching Source: SDS2 data. Similarly in urban schools, 63 percent of councils contribute funds, 43 percent help with administrative duties, 43 percent follow up on student dropouts and 28 percent provide inputs to locality inspector’s teacher supervision reports. Thus, in many schools, education councils are vital to the functioning of basic schools. Figure 1.8: Key Uses of Education Council Funds, by Location, 2010 Percent 40 38 Share of councils that fund: 35 Urban 29 30 25 Rural 25 23 21 20 20 18 18 15 9 10 7 5 0 School Learning Furniture Teacher Textbooks Teacher Feeding Materials Salaries Training Source: SDS2 data. 17 The main uses of funds received from education councils are to pay for school feeding (29 percent of rural and 38 percent of urban schools), learning materials, furniture and textbooks (See Figure 1.8 above). A major difference is that 20 percent of rural schools utilize Education Council funds to pay teacher salaries whereas only 7 percent of schools in urban areas do. Education Councils are also more likely to pay teacher salaries in schools that have volunteer teachers. In terms of school feeding the 2008 basic education survey found that nationally 24 percent of schools in Sudan had access to school feeding programs (MoGE 2008). However, among the four survey states, the share was much larger: 63 percent of schools in Red Sea, 33 percent in Blue Nile, 33 percent in North Darfur and 64 percent in South Kordofan had access to school feeding programs. Box 1.2: Nongovernmental Sources of School Financing Basic schools do not only receive funds from education councils and the government but also from NGOs, businesses, religious institutions and local communities. In urban areas, 15 percent of basic schools receive financial support from the local community (separate from council funding), 12 percent from NGOs, 5 percent from businesses, and 2 percent from religious institutions. Among rural schools in the four survey states, 15 percent receive funds from the local community, 3 percent from NGOs, 2 percent from businesses, and 1 percent from religious institutions. Share of Schools that Receive Funds from: Urban Rural Education Councils 63% 48% NGOs (Non-Religious) 12% 3% Businesses 5% 2% Faith-Based Institutions 2% 1% Local Community 15% 15% Source: SDS2 data. Note: Schools receive financial support from multiple sources. These findings are consistent with those in The Status of the Education Sector in Sudan report (World Bank, 2012), which found that basic schools rely on non-public sources of financing to keep schools running. The 2008 basic education survey also showed similar results, with 89 percent of non-salary expenditures nationwide coming from education councils and only 1 percent from the government (MoGE, 2008). At the state level, 18 Section 2: Key Findings and Policy Implications This section summarizes the main findings from the SDS2 survey regarding basic school characteristics and resources in the four survey states and proposes steps to address the identified supply-side challenges. To improve the basic education system will require a combination of policy change and targeted investment in order to have the largest possible impact on sector performance, given limited resources. Key Findings The average government basic school in the four survey states, whether rural or urban, lacks resources, with adverse consequences for education service delivery and ultimately student learning outcomes. The SDS2 findings point to priority areas that require investment and policy change in order to improve service delivery in basic education in the survey states: school infrastructure; availability of chalkboards, furniture and textbooks; and teacher allocation and training. School Location and Infrastructure (i) In remote areas, students frequently have to walk very far, sometimes up to two hours, to reach school; (ii) More than one in ten rural schools has no roof or walls. Instead, students sit under trees. A substantial number of school days are lost due to adverse weather conditions combined with the lack of adequate school infrastructure; (iii) A substantial share of schools (50 percent of rural and 42 percent of urban) have no source of drinking water; and (iv) Only 63 percent of rural and 81 percent of urban schools have a latrine, many of which are not in working condition. Learning Environment (v) Although the official school year is 210 days, the actual school year is 45 days shorter for the average rural regular school and 37 days shorter for the average urban regular school; (vi) Class sizes are relatively large in urban schools with an average class size of 56 students, compared to 40 students per class in rural schools; 19 (vii) In the average rural school, only 21 percent of Grade 1 students and half of Grade 5 students have a desk. In the average urban school the respective numbers are 39 percent and 70 percent; (viii) Many schools do not have a usable chalkboard. A larger share of urban schools have chalkboards in usable condition (61 percent in Grade 1 and 72 percent in Grade 5) than rural schools (53 percent and 64 percent respectively); (ix) The official student-textbook ratio in Sudan is 2:1. In practice, in the average rural school the ratio is 4:1 in Grade 1 and 3:1 in Grade 5 and in the average urban school it is 3:1 in Grade 1 and Grade 5. These averages conceal great disparities however, as half of all Grade 1 classrooms have no textbooks at all; and (x) A large number of schools receive funds from education councils, which are primarily used to cover the costs of school feeding, learning materials and furniture, due to shortfalls in public education spending. Resource Allocation and Use (xi) The STR is higher at the average rural regular school (36:1) than at the average urban regular school (32:1); (xii) More than 70 percent of teachers at the average rural school have no pre-service training, with implications for teaching quality and student learning; (xiii) Education council funds are used by 20 percent of basic rural schools to pay for teacher salaries; (xiv) Volunteer teachers constitute 16 percent of teachers at the average rural school and 14 percent at the average urban school. One issue with volunteer teachers it that they may not be qualified to teach. The existence of volunteer teachers also signals the insufficient number of teachers being allocated to schools and communities by the government; and (xv) School record keeping is weak, especially with respect to student performance and teacher leave. Generally, urban schools are more likely to keep up-to-date records than rural schools. 20 Policy Implications To address the weak school infrastructure in the four survey states will require increases in the currently low education development spending, initially targeting schools in rural areas where school infrastructure tends to be weakest. A coordinated effort by the federal government, state governments, communities and donors is needed given the extent of the resources required. The MoGE’s new education management information system (EMIS) will be extremely useful in documenting the exact needs of individual schools, the first step towards strengthening the infrastructure of the basic education system. To deal with the severe lack of resources in many basic schools, and especially rural ones, calls for new investment combined with the more efficient use of existing resources. The widespread lack of desks, usable chalkboards and textbooks, each of which is more severe in rural than urban schools, clearly points to the need for additional public spending on goods and services. Given the importance of textbooks for student learning, specific targeted policies to improve textbook supply also need to be considered, including the centralization of provision or the earmarking of federal transfers for local purchases. This is all the more necessary given the financial support a large number of schools already receive from education councils and communities. Indeed, although this de facto funding arrangement enables schools to stay open, it has implications for equity in access and the quality of education received, as richer communities are generally able to provide more funding. The differences in STRs across average rural and urban schools is a reflection of teacher distribution policies, and for Sudan as a whole there is an issue of ineffective teacher distribution because teachers are frequently not allocated based on the number of students and school needs s (See Policy Note No.3 on teacher management). This policy needs to be reviewed to improve service delivery, especially in an education system with limited resources. The first step may be a systematic study of teacher supply and utilization that will establish a database on teachers, including information on their age, experience, qualifications, subject specialization, teaching load and location. Apart from strategies to promote a more equitable distribution of teachers, an important step to raise teaching quality is to re-prioritize the development of teacher competence and pedagogical skills rather than the current focus on upgrading qualifications. However, there are few quick solutions and in the meantime, the most effective interventions would be to target resources to improve the skills of teachers in the least advantaged areas. This report summarizes the main issues of priority in basic education service delivery in the four survey states. To improve the basic education system will require a combination of policy change and targeted investment in order to have the largest possible impact on sector performance given limited resources. 21 Policy Note No.2 Student Learning in Four States of Sudan: School Inputs and Characteristics of Teachers - Findings from the 2010/11 Basic Education Service Delivery Survey Abstract The importance of schooling and learning to economic growth and development is well documented and at the individual level people with more schooling tend to be more productive and earn more. Recent evidence suggests that learning is more important for earnings and development than educational attainment and this issue is of particular importance in countries where school enrollment has increased rapidly but has in many cases been accompanied by a decline in the quality of education. This paper provides new evidence from a representative survey carried out in four states of Sudan (Blue Nile, North Darfur, Red Sea and South Kordofan) in 2010/11. It assessed Grade 5 student learning levels and factors that influence learning. Student learning outcomes in the four survey states are very weak: the average student answered only 28 percent of the mathematics questions correctly. Rural students perform worse on average than urban students and IDP students tend to perform significantly worse than both. Performance also varies significantly according to school characteristics. Students in schools that are closed more often, have poor infrastructure, have less experienced teachers and lack more textbooks perform significantly worse. Moreover, in rural schools richer students, girls and students whose teachers give them exercises to complete perform significantly better. To improve student learning calls for a combination of targeted investment and more efficient use of existing resources. Priority areas include strengthening school infrastructure, improving textbook availability, making teacher deployment more efficient and adjusting the duration and contents of pre-service and in-service teacher training. 22 Introduction The importance of schooling and learning to economic growth and development is well recognized. At the individual level, people with more schooling and higher learning levels tend to be more productive and earn more. Recent evidence suggests that learning is more important for earnings and development than educational attainment (Hanushek and Wöβmann, 2009). This issue is of particular importance in countries where school enrollment has increased rapidly but in many cases is accompanied by a decline in the quality of schooling, with adverse consequences for student learning. Given the importance of student learning, the ultimate objective of education systems, most countries devote large shares of their national budgets to education. But investments in education inputs and improvements in access, enrollment and completion rates do not necessarily mean that students are learning. When student learning levels are low, this provides a strong indication that education systems are not performing as intended. In the case of Sudan there are two crucial questions. First, are students who are in the education system learning? Second, if not, what are the reasons for the lack of learning? Student learning assessments can be used as a tool to track education sector performance and, when combined with information on school and household characteristics, to understand why the expected learning outcomes are not achieved. Little is known about student learning levels in Sudan. What is known is that basic schools are characterized by a lack of resources and in some cases by the ineffective use of existing resources, which reduces the opportunity for student learning (See Policy Note No.1 on service delivery). 23 This paper provides new evidence from a representative survey carried out in four Sudanese states: Blue Nile, North Darfur, Red Sea and South Kordofan. The survey assessed Grade 5 primary pupils’ learning levels and determinants of student learning, including school inputs and student and household characteristics. The survey tested and interviewed 4,542 Grade 5 students (an average of 18 students per school), interviewed 253 head teachers and 1,118 teachers (an average of 4.4 per school) and classrooms were observed in 251 schools. The sample used for this note consists of 3,365 students due to non-response items for two key variables: household assets and availability of math textbooks. However, a comparison of the mean of the relevant variables across the regression sample and their mean in the full sample shows little significant difference. The findings represent the average score of a typical Grade 5 student in government basic education in the four states (See Study Rationale and Survey Description). 23 Basic education is eight years long and the official primary school starting age is six. 23 Section 1 briefly outlines the framework used to examine the determinants of student learning. Section 2 provides the discussion of student performance and its determinants in the context of the four survey states. In Section 3 the empirical results are presented and discussed. Finally, Section 4 summarizes and outlines priority areas for improving student performance in the four survey states. Section 1: Learning Framework The key factors that determine student learning outcomes are: (i) school inputs, current and past; (ii) family and student characteristics; and (iii) students’ inherent abilities, although these are not measurable (Hanushek, 2003; Hoxby, 2001; Krueger, 2003). 24 At the national or regional levels, student learning achievement is typically measured by standardized tests or internationally comparable learning assessments. In this case, the survey only contains data on current school inputs and household and student characteristics. A contemporaneous specification is therefore used to examine the determinants of student learning in the four states. It is not unreasonable to assume that school inputs have been virtually unchanged over the relevant time period, given the resource constraints and the fact that investment in education has increased only relatively recently (World Bank, 2012). A school environment conducive to student learning depends on several factors. On the supply- side these include among other things school infrastructure, instructional time, the availability of textbooks and learning materials and sufficient numbers of appropriately trained teachers (UNESCO, 2007). On the demand-side, at very least students need to regularly attend school, be sufficiently nourished and have time to do their school work. The next section discusses the sample and factors that influence Grade 5 student performance in the mathematics learning assessment, based on the information provided by the interviewed head teachers, teachers and Grade 5 students, and direct observation of Grade 5 classrooms. 24 Inherent ability influences student learning outcomes but the survey does not contain data that can be used to address this issue. However, previous studies suggest that school and family inputs and inherent ability may be uncorrelated in some contexts and therefore standard education production functions can still yield unbiased estimates (Goldhaber and Brewer, 1997). 24 Section 2: Factors Influencing Student Performance To assess Grade 5 student learning levels, the Second Basic Education Service Delivery Survey (SDS2) provides representative data on student learning outcomes and characteristics, household characteristics and school inputs in the four states. The survey included a learning assessment that tested student learning levels, a measure of what the education system is delivering. The analysis of student performance relies on data obtained through four questionnaires: a head teacher questionnaire that captured information on school characteristics; a questionnaire based on direct observation of classroom resources; a teacher questionnaire; and a combined questionnaire that recorded Grade 5 student and household characteristics and contained the student learning assessment. 25 Student Performance The student learning assessment in math was based on a subset of 30 questions from the 1995 and 2003 Trends in International Mathematics and Science Study (TIMSS) learning assessments, which test Grade 4 students’ knowledge. Students in the four states in Sudan were assessed in Grade 5, given the post-conflict situation. All questions included in the learning assessment were reviewed by a steering committee to ensure their suitability, based on the contents of the national curriculum. The mathematics learning assessment included 30 multiple choice questions with four answer options. The average learning assessment score for the four survey states was only 28 percent, indicating that the average student’s level is low, particularly given that students who randomly guess all answers should score 25 percent. Figure 2.1 below shows the cumulative distribution of math scores, underlining the generally weak student performance and rural-urban divide (See also Policy Note 1 on service delivery). Indeed, more rural students achieve low scores (81 percent answered one third or less of the mathematics questions correctly, compared to 76 percent of urban students), and fewer rural students achieve high scores (two percent answered half or more of the questions correctly, compared to five percent of urban students). 25 The variables included in this analysis are defined in Annex 1.1 (See also Study Rationale and Survey Description). 25 Figure 2.1: Cumulative Distribution of Math Test Scores, by Location, 2010 Percent 100% 2% 5% Percentage of Correct Answers 90% 17% 19% 80% 50% and above 70% 33-50% 81% 33% or less 76% 60% 50% Urban Rural Source: SDS2 data and authors’ computations. The mathematics learning assessment tests different cognitive domains. Grade 5 students in Sudan perform relatively better in the comparatively simpler domains of knowing facts and procedures than at solving routine math problems and using complex procedures. This provides a preliminary indication that teaching is more geared towards routine learning and memorization than towards problem solving and the application of concepts. Simple correlations between mean math scores and student characteristics and school inputs, without accounting for other relevant factors, reveal potential reasons for the differences in student performance (See Figure 2.2 below). The math score for the average rural student was 27.1 percent, compared to 28.7 percent for the average urban student, pointing to the importance of location. 26 Female students on average performed better than male students both in rural (27.6 percent, compared to 26.6 percent) and urban areas (29.2 percent, compared to 28.3 percent). Students from wealthier households perform better on average than students from poorer households, in both rural (28.2 percent, against 26.7 percent) and urban areas (30.1 percent, against 27.6 percent). In terms of school inputs, the mean mathematics score is higher when textbook availability is higher (illustrated by a lower student-textbook ratio). In rural schools the score is 28.1 percent where the student-textbook ratio is lower than 2:1, against 26.6 percent where it is higher, and in urban schools the respective mean scores are 29.3 percent and 28.4 percent. Finally, students 26 All rural differences are statistically significant. For the urban sample all differences are significant except for between boys and girls and between students in schools with high versus low student-textbook ratios. 26 whose teachers give them exercises to complete in class rather than have them simply read or copy from the chalkboard, have higher mean tests scores (27.8 percent in rural and 29.8 percent in urban schools) than students whose teachers do not (26.9 percent and 28.5 percent respectively). Figure 2.2: Mean Math Test Scores, by Selected School and Student Characteristics and Location, 2010 Percent 30.1 32 29.8 29.3 29.2 28.5 28.4 28.3 28.2 28.1 30 27.8 27.6 27.6 Average Score (%) 26.9 26.7 26.6 26.6 28 26 24 22 20 Urban Rural Source: SDS2 data and authors’ computations. Note: STBR – Student-textbook ratio. The exercises category corresponds to class activity; where none are practiced, students copy from the chalkboard and read from books. International Comparison As the math learning assessment consisted of a subset of questions from the full version of the Trends in International Mathematics and Science Study (TIMSS), a direct regional comparison is not possible. It is however possible to compare the shares of students that answered each of the selected 24 questions correctly in the four survey states of Sudan with the corresponding shares for Morocco and Tunisia. 27 According to this measure, Grade 5 students in the four states in Sudan perform worse than their Grade 4 counterparts in Morocco and better than their Grade 4 counterparts in Tunisia (See Figure 2.3 below). 27 Not all 30 math questions were used in the learning assessments in Morocco and Tunisia. 27 Figure 2.3: International Comparison of Math Performance, by Question, 2010 Percent 70 Sudan 4 States Morocco Tunisia Share of Correct Answers 60 50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Question Number Source: SDS2 data for the four states of Sudan; IAEEA, 2008 data for Morocco and Tunisia. Summary of School and Student Characteristics Student learning outcomes depend on several factors on the supply-side such as instructional time, school infrastructure, class size, the availability of textbooks and the presence of trained teachers in adequate numbers. In the four survey states in Sudan the availability of school resources differs, sometimes substantially, by rural and urban location. The school inputs and characteristics used in the analysis of determinants of student learning levels are described below in the context of rural and urban schools in the four states. Table 2.3 presents selected indicators for basic rural and urban schools of average characteristics in the four survey states. Urban schools are generally much larger than rural schools, with respective average enrollments of 473 students and 294 students. 28 The average student-teacher ratio is lower in urban (33:1) than rural schools (38:1), related to the challenge of attracting and retaining high-quality teachers in rural and remote schools. Fewer girls than boys are enrolled in both rural and urban schools, but urban schools have comparatively more girls than rural schools (gender parity index of 0.84, compared to 0.76). In terms of school infrastructure, 75 percent of rural schools have a latrine, against 83 percent of urban schools. Given the substantial differences across rural and urban schools and students, the sample was divided into two sub-sets of rural and urban students to examine the relative importance of each factor for their respective learning outcomes. 28 IDP schools, both rural and urban, tend to be considerably larger. Those located in rural areas in North Darfur raise the rural average for that state. 28 Table 2.3: Average School Statistics, by Location, 2010 Urban Rural Time to Locality Administrative Center (Minutes) 18 104 Enrollment (Number of Students) 473 294 Student-Teacher Ratio (All Teacher Types) 33:1 38:1 Gender Parity Index (Co-ed. Schools) 0.84 0.76 Schools with Latrines (Percent) 83 75 Source: SDS2 data. The summary statistics for the sample are provided in Table 2.2 and each of the variables used to examine factors that may influence student learning levels are discussed below in greater detail. Table 2.2: Summary of School and Student Statistics, by Location, 2010 Urban Rural Mean S.D. Mean S.D. Math Score 28.7 11.0 27.1 9.4 School Characteristics School Closure (Number of Days) 6 9 9 13 Drinking Water (% of Schools) 63 48 66 48 Pupil-Textbook Ratio > 3:1 (% of Schools) 59 49 66 47 Teacher Experience (Years) 15.4 5.2 10.2 4.8 Exercises Offered in Class (% of Schools) 0.21 0.41 0.21 0.41 Student and Household Characteristics Age of Student (Years) 11.7 1.6 12.4 2.1 Girl (%) 51 50 49 50 Household Assets (Rural) * — — 2.0 2.0 Household Assets (Urban) * 5.3 5.2 — — IDP Students (% of Total) 1 10 3 17 No. Of Observations 1,658 1,717 Source: SDS2 data. Note: The estimates are of the population means and standard deviations. Teacher and school resources are school averages. * To capture the role of family background the model uses asset indexes. The indexes consist of assets, weighted according to the share of students in the sample whose households own each type of asset. The urban asset index includes car, refrigerator, television and computer and the rural asset index includes car, television and mobile phone. In the sample, the mean asset index is 2.0 for rural students and 5.3 for urban students. Rural-Urban Location One of the main determinants of school resources in Sudan and other developing countries is the rural-urban location. Moreover, Monitoring Learning Achievement (MLA) assessments conducted in Sub-Saharan Africa find that urban students generally have better learning outcomes than rural students in all countries and all learning areas: literacy, numeracy and life skills (Chinapah, 2003; Majgaard and Mingat, forthcoming). 29 In the SDS2 survey sample, 52 percent of students attend rural schools and 48 percent attend urban schools. 29 The classification of schools as rural or urban is based on the reporting of the head teacher at each school and cross-checked against the rural-urban classification contained in the school records of the state ministries of education in the four survey states. Finally, the distance from each school to the relevant locality administrative center was checked. Instructional Time (School Closures) The official school year in Sudan lasts 210 days but the mean actual school year in the four survey states, as reported by head teachers, is just 171 days (World Bank, 2012). This suggests a considerable loss of instructional time, of 39 school days per year on average, with adverse implications for student learning. In many cases the loss of school days is the result of adverse weather conditions combined with poor school infrastructure, particularly in rural areas where school infrastructure is generally weaker. Schools close due to bad weather (heat or rain) for nine days on average in rural areas and six days on average in urban areas (See earlier Table 2.2). Among rural schools, those with more days of closure than the mean frequently have very poor infrastructure. They tend to lack walls or other protection from the heat and rain, latrines, drinking water, or a combination of the three. To test whether students enrolled in schools that close more often when they should be open have lower learning levels, the number of school closure days is included in the empirical model. Drinking Water Existing evidence points to the importance of the condition of school buildings and the availability of sanitation and drinking water for student learning outcomes. School infrastructure in Sudan is generally poor: in 2007 half of classrooms in basic schools were estimated to need rehabilitation or complete replacement (MoGE, 2008). The vast majority of schools with drinking water also have walls and latrines. Among the sample schools in the four survey states, 66 percent of rural schools and 63 percent of urban schools have drinking water (See earlier Table 2.2). However, the most common water sources are pumps and wells for rural schools and taps and wells for urban schools, indicating greater access to safe drinking water in urban areas. To take the state of school infrastructure into account and determine if it affects student learning levels, the model included a variable indicating the availability of drinking water. 29 For details of rural-urban school characteristics and resources see Policy Note No .1 on service delivery. 30 Textbook Availability When textbooks are available students generally perform better. Moreover, having textbooks can reduce differences in the performance of rural and urban students and improve the performance of students from poorer backgrounds. However, which students gain most from the increased availability of textbooks is not clear. Some evidence suggests that textbooks mainly improve the performance of stronger students and other evidence that students from impoverished backgrounds benefit more (Heyneman, Jamison, and Montenegro 1984; Glewwe, Kremer, and Moulin 2009). The official student-textbook ratio (STBR) in Sudan is 2:1 (MoGE 2008), but the majority of schools have a higher ratio and many lack textbooks completely (World Bank, 2012). In the rural sample, 66 percent of schools had student-textbook ratios of 3:1 or higher in Grade 5, compared to 59 percent of urban schools, again underlining the rural/urban differences in school resources. To assess the relationship between textbook availability and student performance in the four survey states the model included a variable for students in Grade 5 classrooms with a 3:1 or higher student-textbook ratio. Teacher Qualifications, Training and Teaching Skills Adequate numbers of appropriately trained teachers at schools are crucial to improve and sustain education service delivery performance and student learning outcomes. Sudan does not generally have a teacher shortage. The national STR is 33:1, below the Sub-Saharan African benchmark of 40:1. However, there are issues of teacher allocation, education, pre-service and in-service training and teaching methods, each of which affect student learning (World Bank, 2012). Teacher experience tends to affect student performance, with students of more experienced teachers generally achieving better learning outcomes. In the four survey states, teachers in urban schools have more experience on average (15.4 years, with a range of 4 to 27 years) than teachers in rural schools (10.2 years, with a range of less than a year to 24 years - See earlier Table 2.2). The analysis includes teacher experience (measured as the mean number of years teachers at a given school have worked as teachers) to examine whether students of more experienced teachers perform better. Another major determinant of student learning is the quality of teachers and what they do in the classroom. Studies examining the importance of teacher quality for student performance have found that students of higher quality teachers tend to learn significantly more, but that teachers’ education is not important in explaining differences in teacher quality (Hanushek, 2002; Harbison and Hanushek, 1992; Rivkin et al., 2003; Rockoff, 2004). However, it is difficult to assess the quality of teachers. 31 In an attempt to capture the effect of pedagogical practice, the survey included a variable for whether teachers give their students exercises to complete in class as opposed to having students read or copy from the chalkboard. 30 About one in five (21 percent) of students reported that their teacher provided them with exercises to complete in class, there being no significant difference by location. What teachers do in the classroom seems to be related to their experience, as in both urban and rural schools more experienced teachers are significantly more likely to provide exercises to their students. To examine if what teachers do in the classroom affects student performance the model includes an indicator variable that takes the value one if teachers provide exercises. Many studies of student achievement assess teachers’ educational attainment and its relation to student learning. Overall, the evidence suggests that while 10 to 11 years of general education is needed to teach at the primary level, additional years of education do not have any significant effect on student learning outcomes (Harris and Sass, 2008; Majgaard and Mingat, forthcoming). Nearly 90 percent of rural and urban teachers in the sample have 10 to 12 years of formal education (ranging from 8 to 14 years). However, no variable capturing teachers’ academic qualifications was included in the model due to its high correlation with teacher experience. Neither does this paper examine the relationship between in-service training and student performance, mainly because of the large variation in the contents and duration of in-service training, which make it impossible to construct a comparable measure with the available data. Student and Household Characteristics This section discusses the student and household variables included in the analysis that affect student learning outcomes. Student Age Students who start school earlier or later than the official school age may have different learning outcomes than students who start at the official school age, but the effect of being an early or late starter on subsequent achievement is ambiguous (Cascio and Whitmore Schanzenbach, 2007; Taylor et al., 2010). In Sudan, a large share of children (more than 40 percent) start school at the age of eight or older, which is later than the official entry age of six years (World Bank 2012b). A student who starts at age six will turn 11 years old in grade 5 assuming no repetition. The mean age of Grade 5 students is higher for the average rural student (12.4 years) than for the 30 Teachers who provide exercises rather than just have students copy from the chalkboard or read aloud are generally thought to spend more time preparing their classes and to interact more with their students, improving the quality of learning outcomes. 32 average urban student (11.7 years), indicating that children in rural areas start school later or are more likely to repeat. To account for the possibility that the school entry age and/or repetition affect student learning outcomes, the model incorporates the age of each student. Gender The relationship between gender and student achievements is ambiguous; some studies find that girls tend to perform better than boys and others find the opposite, or no effect (Barrera-Osorio et al., 2010; Chinapah, 2003; Cueto, Ramirez, and Montenegro, 2006; Mullis et al., 2000). The relationship between gender and student learning is pertinent in the four survey states in Sudan because fewer girls than boys are enrolled overall and there is anecdotal and emerging quantitative evidence that girls drop out earlier than boys due to marriage. In the urban sample 51 percent of students are female, compared to 49 percent in the rural sample, suggesting that there is some sample selection. 31 To capture any gender differences in student performance a variable is included for female students. Household Wealth Family background also influences student achievement. Some family characteristics are beyond the control of households, such as ethnicity; some are partly controlled by households, such as parental education and income; and some are under the control of the household, including the amount of time parents spend helping their children with homework, the level of encouragement provided and interactions with the school. To capture the role of family background the model uses asset indexes. The indexes consist of assets, weighted according to the share of students in the sample whose households own each type of asset. The urban asset index includes car, refrigerator, television and computer and the rural asset index includes car, television and mobile phone. In the sample, the mean asset index is 2.0 for rural students and 5.3 for urban students. Although not directly comparable, these indexes indicate the difference across rural and urban location in terms of household wealth, in addition to the disparities in school resources described above. IDP (Internally Displaced Person) Students In North Darfur a large share of students in basic education attend IDP schools that are typically very large. The mean size of IDP schools for all three Darfur states over 2008-09 was 815 31 Girls enrolled in rural areas may be systematically different from those enrolled in urban areas. 33 students, compared to a mean of 309 students for regular schools (World Bank, 2012). Students enrolled in IDP schools are also different in that they live in large camps and have been uprooted from their place of origin (See Policy Note No.5 on IDPs). Three percent of rural and one percent of urban students are IDPs. To detect potential differences in performance of IDP students relative to other students the model includes an indicator variable that takes the value of one for IDP students. The results from the regressions of the impact of the variables mentioned above on student math performance are presented and discussed in the next section. Section 3: Results and Discussion The estimated model (See Annex X) considers student learning outcomes a function of school inputs and student and household characteristics and is estimated by the ordinary least squares (OLS) approach, weighted to represent the underlying student population. The results for the rural and urban samples presented in Table 2.4 show the variables that are associated with student performance (shaded) but do not imply a causal relationship. Table 2.4: Determinants of Grade 5 Student Performance in Math, by Location, 2010/11 Dependent Variable: Urban Rural Math Test Score (%) Coefficient S.E. Coefficient S.E. School Characteristics Days of School Closure -0.607*** (0.009) -0.119*** (0.009) Drinking Water 1.733*** (0.309) 5.303*** (0.306) Pupil-Textbook Ratio > 3:1 -7.852*** (0.781) -6.161*** (0.290) Teacher Experience 0.883*** (0.179) 1.112*** (0.069) Teacher Experience Squared -0.050*** (0.004) -0.059*** (0.003) Exercises Offered in Class 1.011*** (0.583) 1.159*** (0.494) Student and Household Characteristics Age of Student -0.092*** (0.194) 0.039*** (0.126) Girl -0.510*** (0.686) 1.068*** (0.433) Household Assets (Rural) - - 0.337*** (0.100) Household Assets (Urban) 0.044*** (0.078) - - IDP Students -20.920*** (0.541) -5.338*** (0.449) Constant 49.34*** (2.876) 23.800*** (1.620) R2 0.302 0.233 No. of Observations 1,658 1,717 Source: SDS2 data and authors’ computations. Note: Teacher and school resources are school averages. Standard errors in parenthesis: *** Statistically significant at the 1% level; ** Statistically significant at the 5% level; * Statistically significant at the 10% level. 34 It is important to keep in mind that these results apply to enrolled students, but do not provide information on factors that influence learning for students who are out-of-school if these students’ characteristics are significantly different. Moreover, the observable school inputs and student and household characteristics included in the model explain about 23 percent and 30 percent of the variation in rural and urban students’ performance, respectively, underlining the importance of unobservable characteristics such as students’ inherent ability and teaching quality for learning outcomes. 32 School Characteristic Determinants The number of days schools are unduly closed is associated with significantly lower student math scores both for rural and urban students. Among rural students, a one standard deviation (13 days) increase in the number of closure days is associated with a 1.5 point reduction in test scores; among urban students a one standard deviation increase (9 days) in closure days is associated with a 5.5 point reduction, holding all other factors constant. The relatively larger effect of school closures on urban students’ performance is probably due to differences in the quality of teaching: as quality tends to be worse in rural schools, the effect on student performance may be smaller. The measure used to assess the effect of school infrastructure on student learning is the availability of drinking water. Both rural and urban students in schools that have drinking water perform significantly better, underscoring the importance of adequate school infrastructure. The effect is particularly large for rural students: the math score is 5.3 points higher for the average rural student in a school with drinking water. Urban students in schools with drinking water score 1.7 points more than those in schools without conditional on the other variables. The results strongly suggest that both rural and urban students in classrooms with a student- textbook ratio of 3:1 or higher perform significantly worse, and that the student-textbook ratio is the best predictor of student performance. The math scores of students in classrooms where three or more students share a textbook are six points lower in rural schools and nearly eight points lower in urban schools, compared to students in classrooms with two students or less per textbook, holding all other factors constant. 32 A variable capturing education council activity (frequency of meetings) was initially considered in the model but not included in the final analysis as it was not significantly correlated with student performance, but was significantly correlated with some of the other explanatory variables with implications for statistical inference. 35 Students taught by more experienced teachers perform significantly better. 33 For rural students, a one standard deviation increase in teacher experience (4.8 years) is associated with a 5.3 point increase in math scores; for urban students, a one standard deviation increase (5.2 years) is associated with a 4.3 percentage point increase in math scores. One possible reason for the positive effect of more experienced teachers is that teaching skills improve over time and that they are better at classroom management than less experienced teachers. However, there are diminishing returns to teacher experience, meaning that after a certain level of experience has been reached, the benefits of additional years of experience start to decline. Rural students whose teachers provide exercises to complete in class have significantly higher math scores, suggesting that what teachers do in the classroom influences learning outcomes. For urban students the effect of being given exercises is close to significant, indicating that what goes on in the classroom influences student learning in urban schools also. On average, students given exercises have scores that are 1.2 points higher in rural areas and 1.0 point higher in urban areas than for students who simply read aloud or copy from the chalkboard. The effects are relatively small in size but given that providing exercises or not is a rough measure of classroom activity, this is not surprising and underscores the importance of effective teaching for student learning to take place. Together, the findings on teachers’ experience and classroom activities suggest a need to adjust training, to impart to new teachers the pedagogic skills required to interact successfully with students, that more experienced teachers seem to acquire over time. Student and Household Characteristic Determinants Age does not significantly influence learning in the four survey states and the difference in performance between urban boys and girls is not significant. However, rural girls perform significantly better than rural boys, with an average math score 1.1 points higher, holding all other variables constant. One possible reason for the gender difference in performance is that fewer girls than boys are enrolled in rural areas overall and those girls who do attend school may be those more likely to perform well. In rural schools, wealthier students perform significantly better than students from poorer households. A two standard deviation increase in the asset index (4.0) is associated with an almost 3.0 point increase in the math score for the average student. For urban students on the other hand, there is no significant relationship between household wealth and student performance. Several possible reasons for students from wealthier households performing better 33 Teacher academic qualifications were initially included in the model but were excluded in the final analysis since they were not significantly associated with test scores but are correlated with teacher experience with consequences for statistical inference.. 36 exist. For example, wealthier households tend to have more resources in terms of books, computers and newspapers; parents in wealthier households are likely to be better educated themselves, and are therefore better placed to help their children with homework; children in wealthier households are less likely to have to help at home or work than their poorer counterparts and so on. Finally, the estimates show that math performance among both rural and urban IDP students is significantly worse than for other students, and the effects are large. On average, IDP students have scores that are 5.4 points lower in rural schools and 21.0 points lower in urban schools than their non-IDP peers. Thus, after accounting for household wealth, gender, age and school inputs, IDP students still perform worse, indicating that other factors are at work (See Policy Note No.5 on IDPs in North Darfur). Many IDP students are orphans and have been uprooted from their homes, which may partly explain their weaker performance. Section 4: Key Findings and Policy Implications The last section of this note summarizes the results on student performance in the four survey states and discusses priority areas for investment and policy change to improve student learning outcomes, based on the findings. Key Findings Student learning outcomes in the four states are very weak in the four survey states: the average student answered only 28 percent of the learning assessment math questions correctly. As such, results provide a clear sign that there are issues with the delivery of education services in basic schools. Rural students and IDP students on average perform worse in the math assessment. Considering school-side factors of performance, students in schools that are closed more often, with poor infrastructure, with less experienced teachers and greater lack of textbooks perform significantly worse. Moreover, in rural schools richer students, girls and students whose teachers set exercises perform significantly better. Taken together these findings underline the importance of school inputs and teachers for student learning, in addition to individual student and household characteristics. For students to learn in an environment characterized by weak school infrastructure and shortages of learning materials and textbooks is difficult. To improve student learning outcomes to the greatest extent possible will require both additional resources and the more effective use of existing resources, carefully targeting quality determinants. 37 Additional resources are urgently needed in many schools for investments in school infrastructure (including buildings, water and sanitation), to enable them to respect the imparted number of school days and to improve learning conditions. The shortage of textbooks in basic schools in Sudan is well-documented and the findings from the SDS2 survey corroborate the importance of textbooks for student learning outcomes in both rural and urban areas. The official textbook policy stipulates a ratio of 2:1, but many schools are far from meeting this standard, highlighting the urgency of further public investment in textbook provision. Many schools further lack desks and usable chalkboards, equally necessary for effective teaching. That resources be channeled to where they are most needed is crucial, especially in environments of limited public resources, such as the four states in Sudan. Rural students tend to perform worse than urban students primarily because of variations in the resource allocation between rural and urban schools and differences in rural and urban household characteristics. Targeted investments in the infrastructure and equipment of rural schools, possibly combined with conditional cash grants to enable rural parents to send and keep their children in school, can help reduce the disparity between rural and urban learning outcomes. Students in IDP schools also perform significantly worse than other students, calling for resources to target these schools. Evidence suggests that teachers are the area where there is most scope to improve resource use. Both rural and urban students taught by more experienced teachers and by those who provide exercises in class perform significantly better. This underlines the need to make the existing teaching staff more effective through adjustments to pre-service and in-service teacher training, to emphasize pedagogy, feedback and student-teacher interaction. Teacher deployment policies also need review to improve the efficiency of teacher use. Policy Priorities (i) Targeted investment in school infrastructure to enable schools to stay open as planned and improve the learning environment; (ii) Modification of school calendars in areas where weather conditions force schools to close for long periods of time and in rural areas to accommodate students who need to do seasonal agricultural work; (iii) Additional resources for textbook provision to reach the official 2:1 student-textbook ratio or lower in all basic schools; (iv) More efficient teacher deployment and adjustment of the duration and contents of pre- service and in-service teacher training to focus on pedagogy, classroom management, lesson preparation and teacher-student interaction; and 38 (v) Establishment of a national assessment of student learning to assess school and student performance and allow the targeting of resources where the needs are the greatest, both across and within schools. 39 Policy Note No.3 Challenges and Possibilities for Teacher Management in Four States of Sudan - Findings from the 2010/11 Basic Education Service Delivery Survey Abstract Teachers and their pedagogical practices are some of the most important determinants of student learning outcomes. Consequently, how the system for managing teachers, including recruitment, deployment, and performance, works is critical to the performance of the education sector. This paper provides an overview of the strengths and challenges related to teacher management in four states of Sudan: Blue Nile, North Darfur, Red Sea and South Kordofan. The findings are based on the second Basic Education Service Delivery Survey (SDS2) conducted in 2010/11 that provides representative data on school resources, teachers and student performance. Three key challenges are identified: (i) Teacher deployment and utilization is uneven and relatively inefficient; (ii) Teacher professional skills development constitutes a challenge; and (iii) Teacher performance management and career advancement is hindered by the lack of an effective linkage of supervision mechanisms to incentive structures. Based on international experience and best practice five potential policy shifts associated with substantial system improvements that are particularly relevant to Sudan are a shifting from central management of teacher supply to a regulated market approach; from central teacher deployment to incentives driven approaches; from teacher quantity to teacher quality; from teacher qualifications and experience to competence and performance; and from performance management based on inputs and processes to management based on outputs and learning outcomes 40 Introduction Teachers and their pedagogical practices are some of the most important determinants of student learning outcomes and the students of higher quality teachers consistently achieve higher results. Consequently, how the system for managing teachers (including recruitment, deployment and performance) works is critical to the performance of the education sector, as measured by one of its key outputs: learning achievements. In Sudan, basic education service delivery performance and student learning outcomes are weak (MoGE, 2008; World Bank, 2012), driven both by supply and demand-side factors). 34 This paper examines one particular aspect of service delivery: teacher management. The aim is to provide an overview of strengths and challenges relating to teacher management in Sudan as a whole, followed by a more detailed analysis of teachers in four states: Blue Nile, North Darfur, Red Sea and South Kordofan. The findings are based on the second Basic Education Service Delivery Survey (SDS2), conducted in the four states in 2010/11, that provides representative data on school resources, teachers and student background and performance. 35 The findings are intended to inform the Education Sector Strategic Plan (ESSP), 2012-16 and enable it to better target resources and review teacher management policies to improve service delivery and student learning outcomes in basic schools. The Status of the Education Sector in Sudan (ESR) report identified teacher management as one of the key challenges in basic education in Sudan, in particular, inefficient teacher deployment and utilization, and teacher training. 36 Section 1 begins by providing an overview of the strengths of and challenges in teacher management in Sudan. Section 2 provides a closer look at teachers and teacher management in basic schools in the four survey states. Finally, Section 3 outlines priority areas based on the national and state-level findings and discusses international experience and actions to improve teacher management and education system performance. 34 Basic education is eight years long in Sudan and the official school starting age is six years. 35 A total of 1,118 teachers, were interviewed for the SDS2, equivalent to 4.4 teachers per school. The sample constitutes roughly six percent of all teachers in government basic schools in the four states and the findings apply to the average basic school teacher in these states (See Study Rationale and Survey Description for details). 36 This report also draws on two follow-up World Bank missions to Khartoum in March and June 2011. The two missions included meetings with ministry officials at the federal, state (1) and locality (2) levels, and discussions with officials in 5 administrative units. Meetings were also held with a senior official in the General Union for All Workers in General Education in Sudan, a small group of development partners, and a senior professor from the University of Khartoum Graduate School. 41 Section 1: Strengths and Challenges of Teacher Management in Sudan There are some significant strengths in the teacher management system in Sudan. The labor market for teachers is generally characterized by competition for positions in the teaching service and student-teacher ratios at the basic school level are comparable with international norms, which suggest that the overall supply of teachers is adequate. Sudan has established procedures to determine the demand for new teacher positions yearly and, subject to budget availability, deploy new teachers and manage transfers within and across localities. At the national level, the distribution of education workforce by type and function is reasonably well-balanced: 80 percent are government teachers, 4 percent are national service and volunteer teachers (although this proportion is much higher in some states), 12 percent are government non-teaching staff and 4 percent are central and decentralized education staff. The overall gender balance of teachers is in line with international trends: at the national level 67 percent of basic school teachers are female (World Bank, 2012). 37 When it comes to teacher performance management, systems of regular inspections and head teacher reporting are in many cases in place and provide a basis for teacher performance assessment. The majority of teachers in most states receive their salaries regularly and in full (there are some difficulties in this regard in more remote schools) and are entitled to public service benefits, including pension and health benefits. Moreover, teachers are supported by vibrant and active general workers’ and teachers’ unions, which provide considerable welfare services for teachers and other education employees. Nevertheless, teacher management in Sudan also faces challenges that fall into three categories: (i) teacher deployment and utilization; (ii) teacher preparation and professional development; and (iii) teacher performance management and career advancement. Each of these issues is discussed below at the country level and in Section 2 in more detail for the four states. Teacher Deployment and Utilization Teachers are very unevenly distributed across schools. A comparison between the need for teachers (based on student numbers) and the supply of teachers in each school reveals a very low correlation (See Figure 3.1 below). Nationwide, the correlation at the school level between the number of students enrolled and the number of teachers is weaker than in 28 African comparator countries (and somewhat stronger than in two countries). The distribution of teachers is also uneven across states within Sudan. For instance, in the state of Khartoum the correlation between the number of students and the number of teachers at the school level is relatively weak whereas in the Red Sea state it is relatively strong (World Bank, 2012). This implies that factors other 37 For this section the paper draws on the 2012 Status of the Education Sector in Sudan report and two follow-up World Bank missions to Khartoum, in March and June 2011. 42 than student enrolment (such as subject specialization requirements, family status, living conditions and so on) are relatively more important in determining the deployment of teachers in Khartoum than in the Red Sea state. Thus, while the teaching profession is attracting a large number of candidates, in 2011 for instance, there were 7,000 applications for some 1,400 posts in Khartoum state and in Gezira roughly 21,000 applications were submitted for 1,600 positions, it appears that the regulations and incentives currently in place are not functioning well to promote efficient teacher deployment. Figure 3.1: Teacher Deployment at the School Level, Sudan, 2008/09 Number of Students Source: Reproduced from Word Bank, 2012. Once deployed, teacher utilization is uneven and relatively inefficient, with a basic education student-teacher ratio of 34:1 and an average class size of 48. This gap implies that at any one time, almost one third of teachers in basic education are not in front of a class. Furthermore, actual instructional time is significantly below the official number of contact hours. There are several possible reasons for this, including unplanned school closures, teacher absenteeism, in- service teacher training, strikes, conflict, or the use of schools as polling stations. Another factor is that senior teachers generally spend less time teaching than junior teachers, with their instructional time ranging from 5 to 25 hours per week (World Bank, 2012). 43 The above indicators are relatively crude measures of teacher deployment and utilization and are not sufficient for policy or planning purposes. But they signal an important area for further and more systematic investigation so that policies, procedures and management steps can be put in place to improve teacher deployment and utilization. Any such steps would need to be accompanied by the development of simple indicators so that progress can be measured. International experience indicates that these disparities are relatively easy to identify, but require considerable time, planning and some investment to rectify. Indeed, they usually call for a combination of better regulation and enforcement, the use of incentives to attract teachers to remote schools and/or the targeted recruitment of teachers for remote communities. Teacher Preparation and Professional Development Although the number of teachers with academic qualifications has risen substantially, professional skills development continues to constitute a challenge in Sudan. The pre-service qualification for basic education teachers was revised in 1993 and substantial strides have been taken in the transition from the teacher preparation model based on teacher training institutes (TTIs) to the requirement for all teachers to have a bachelor of education degree (B.Ed) provided by the faculties of education at university level.. Whilst this policy change provided a significant boost to university enrollment and called for a substantial investment of resources and time on behalf of both the government and teachers, there appears to be widespread dissatisfaction with the professional skills that university trained teachers enter the professions with, especially at the basic school level. The major focus of investments in teacher training and professional development over the past decade has been on this massive qualification upgrading effort, at the cost of investment in teachers’ ongoing professional development. The number of short, skills-focused courses for in-service teachers has been significantly impacted by the budget constraints at the state and locality level, where responsibility for this function lies. International experience suggests that while good teacher preparation can help to improve teacher quality, it is the classroom experience that makes the greatest contribution to the development of teaching skills. Moreover, it is clear that novice teachers benefit most from good initial training if they enter supportive school environments. Further, a strong international trend exists to move away from center-based courses to school and cluster-based professional development, emphasizing the role of the school head teacher as pedagogical leader, with senior and experienced staff acting as mentors for less experienced teachers. The expansion of school- based professional development requires a significant investment in school leadership development, involving head teachers and senior teachers, as well as the implementation of school development clusters, especially serving smaller schools, with itinerant mentor teachers. 44 Teacher Performance Management and Career Advancement As noted above, teachers in some states are supported by a performance management system which identifies well and poorly performing teachers, although the type and frequency of supervision varies. The current system, where it exists, is a valuable institutional asset and provides a platform to build on. Nevertheless, given that teacher performance management is a decentralized function, different states use different approaches that reflect their differing institutional capacity and context. For instance, it was noted during field visits in the state of Khartoum that while the teacher appraisal system appears to be working well, the linkages to incentives are not optimal. In addition, one of the rewards for appointment to a higher level post appears to be a reduced teaching load. If this is not linked to significant additional administrative or management responsibilities this may simply serve to channel the services of the best, most experienced teachers out of the classroom. Merit-based teacher career advancement and teacher motivation is constrained by the practice of open cadre promotion, according to which teachers automatically move up through seniority grades, except for a few threshold levels where an examination may be required. Cash promotion is also used for some teachers and officials who display merit but for whom no promotion post is available to help incentivize performance. This could be developed into a parallel career track to keep good teachers in the classroom and good head teachers in schools, rather than channeling them into system management positions for which they may not be as well suited. There is now a substantial and growing literature on teacher performance management and best international practice which may provide ideas on how the teacher performance management system can be optimized and linked to incentives, and best practices be shared across states. International experience also points to some successful initiatives in which parental or community involvement has helped improve teacher performance. At the simplest level this may imply involving community members in monitoring teacher attendance, but as school-based management is strengthened, an increasingly stronger role for parents and communities can emerge to help make teachers downwards accountable to parents and the community and horizontally accountable to their peers, as well as upwards to their employers. Section 2: Teachers and the Teaching Environment in Four States This section examines the main teacher management challenges outlined above in more detail based on the findings on teachers in basic schools in the four SDS2 survey states: Blue Nile, North Darfur, Red Sea and South Kordofan. The findings are broken down by rural-urban location to illustrate the differences across the two groups, which can be significant. 45 A total of 1,118 teachers (4.4 teachers per school) were interviewed for the SDS2. The teacher sample constitutes roughly 6 percent of all teachers in government basic schools in the four states and the findings apply to the average basic school teacher in these states. Teacher Characteristics The vast majority of teachers in the four survey states are regular government teachers (85 percent of rural and 91 percent of urban teachers), compared to the national average of 80 percent - See Figure 3.2). Those teaching as part of their national service (less than 2 percent) and trainees (under 1 percent) represent smaller shares. The largest rural-urban difference is in the share of volunteers: 14 percent of rural compared to 6 percent of urban teachers are volunteers. The relatively larger share of volunteer teachers in rural areas implies that teacher shortages in rural schools are being met by parents and education councils who recruit volunteer teachers to fill the gaps (World Bank, 2012). The use of volunteer teachers also has potential implications for teaching quality if volunteers have no formal teacher training. Figure 3.2: Distribution of Teachers, by Employment Type and Location, 2010 Percent 6 2 Urban Teachers 1 Rural Teachers 14 Government Government Volunteer Volunteer Trainee Trainee 91 National 85 National Service Service Source: SDS2 data. Nationally, 67 percent of teachers are female and female teachers represent over 60 percent of teachers in all states except South Kordofan, West Darfur and North Darfur (World Bank 2012). In the four survey states, there are comparatively more female teachers in urban areas (81 46 percent, compared to 60 percent of rural teachers). The gender imbalance by rural-urban location is partly due to security issues as three of the states are in conflict areas. Teacher Deployment and Utilization As in the country as a whole, teachers in the four states are unevenly distributed across schools. When analyzed by location, it appears that the number of teachers more closely correlate with the number of students enrolled in rural schools (the R2 coefficient is 0.62, against 0.44 for urban schools). The R² coefficient indicates that for rural schools, the number of students enrolled explains 62 percent of the variation in teacher distribution, indicating that deployment is determined by factors other than enrollment as well. Figure 3.3 illustrates the variability in teacher distribution For instance, urban schools that have about 1,000 students enrolled may have anything between 15 and 40 teachers. Figure 3.3: Teacher Deployment at the School Level, Four Survey States, 2010 40 Urban schools 30 Rural schools R² = 0.437 R² = 0.615 Number of Teachers Number of Teachers 30 20 20 10 10 0 0 0 500 1,000 1,500 0 500 1,000 1,500 Number of Students Number of Students Source: SDS2 data. Note: Each dot on the graph represents a school. R2 measures the extent to which the number of teachers in a school is proportionate to the number of students. If the number of teachers in a school is perfectly proportionate to the size of enrollment, R2 will be equal to one. Conversely, an R2 value of zero would indicate that there is no relationship. Determinants of Teacher Postings How teachers learn of and obtain their teaching positions is related to the uneven teacher distribution. The majority of basic school teachers, both rural and urban, learnt about their current teaching position from a colleague, friend or relative (36 percent on average). Other sources of information, in order of importance, include: radio (29 percent); teacher unions, (more 47 important in rural areas at 19 percent, against 15 percent in urban areas) and TV and newspapers (2 percent on average). Officially, teachers are selected through a test and interview process. In the four states 75-77 percent of teachers reported that this was the most important factor in obtaining their current position,. There are also other factors that affect who gets a particular teaching position including colleagues, friends and relatives (considered the most important factor by 9-10 percent of teachers), teacher unions (4 percent of postings) and government officials (2-3 percent of postings). Teachers were also asked the reasons for their appointment to their particular school (See Figure 3.4). There is one notable difference between the perceived reasons for rural and urban teachers: 42 percent of rural teachers thought they had been appointed because of staff shortages, whereas only 22 percent of urban teachers felt this was the main reason for their posting. This is symptomatic of the general difficulty in attracting teachers to rural schools. the other main perceived reason for postings is the annual transfer procedure, cited by 39 percent of urban and 32 percent of rural teachers. Other reasons include proximity to family (23-25 percent), subject specialization (for 11-13 percent) and health reasons, especially in urban areas (cited by 7 percent of urban teachers, against 3 percent of rural teachers). Only about one percent of teachers reported being appointed to their post as the result of a promotion, corroborating that while teacher supervision systems exist in Sudan, the linkages to incentives are not strong. Figure 3.4: Reasons Perceived by Teachers for their Appointment, by Location, 2010 Percent 50 42 Share of teachers who cite: 39 Urban 40 32 Rural 30 23 25 22 20 14 12 13 11 10 7 3 1 1 0 Annual Proximity Staff Subject Health Promotion Other Transfer to Family Shortage Special- Reasons ization Source: SDS2 data. Note: Respondents chose all options that applied to their case. 48 Staffing rural and remote schools is a challenge in Sudan. Some countries have experimented with employing teachers locally under the assumption that they will be more willing to work and stay in a particular rural or remote area because of local ties. Among the teachers in the four states, about half (49 percent of rural and 53 percent of urban teachers) are locally recruited, meaning that they teach in their home town (See Figure 3.5). Rural teachers are much more likely to work close to home than their urban counterparts: almost twice as many work in a neighboring locality (34 percent, against 18 percent of urban teachers) and considerably fewer work in another state (17 percent, against 29 percent of urban teachers). Overall, this suggests that working in urban schools involves greater geographic mobility than working in rural schools. Figure 3.5: Proximity of Teacher Posting to Home, by Location, 2010 Percent 60 53 Urban Rural Share of Teachers Working in: 49 50 40 34 29 30 18 17 20 10 0 Home Neighboring Other Town/Village Locality State Source: SDS2 data. Teacher Preparation and Professional Development Teacher preparation and professional development directly affects teacher performance and ultimately student learning outcomes. Three aspects are examined here: teacher experience, academic qualifications and training. Teacher experience tends to influence student performance and students of more experienced teachers generally achieve better learning outcomes, including in the four survey states (See Policy Note No.2 on learning outcomes). The SDS2 survey show that urban teachers in the four states generally have more teaching experience (15 years) compared to rural teachers (10 years). 49 Figure 3.6: Teachers by Academic Qualification and Location, 2010 Percent Basic or Intermediary 50 2 Urban 36 10 Secondary 2 Education Non-Teaching Diploma 2 BA/BSc Rural 5 48 1 45 Postgraduate Degree 0% 20% 40% 60% 80% 100% Source: SDS2 data. Nearly all teachers in the four states have a secondary education or more (See Figure 3.6). 45 percent of rural and 36 percent of urban teachers have secondary education, 48 percent and 50 percent respectively have a BA/BSc degree, and 1.5 and 2.2 percent respectively have a postgraduate degree (Error! Reference source not found.). Among rural teachers 1.4 percent only have a basic or intermediary education, for urban teachers the corresponding share is 1.7 percent. Thus, the majority of teachers appear to be academically qualified to teach at the basic education level. One component crucial for teachers to enable student learning to take place is appropriate pre- and in-service teacher training. Pre-service training provides novice teachers with the skills and knowledge to effectively plan lessons, manage classrooms and teach. In-service training is needed to sustain and upgrade the skills of practicing teachers over time. As discussed earlier, there appears to be some dissatisfaction with the professional skills that newly trained teachers enter the profession with, in particular at the basic school level. Starkly, 73 percent of rural and 68 percent of urban basic school teachers in the four survey states do not have any pre-service teacher training at all (See Figure 3.7 below).. Of rural teachers 10 percent have a teaching diploma and 18 percent a B.Ed degree compared to 22 percent and 10 percent respectively of urban teachers. The findings for the four states highlight the lack of pre-service training, with the implications this entails in terms of teacher quality and performance. 50 Figure 3.7: Teachers with Pre-Service and In-Service Teacher Training, by Location, 2010 Percent 80 73 76 73 Urban 68 60 Rural Percent 40 22 18 20 10 10 0 Teaching Bachelor of None (In-Service Diploma Education Training) Pre-Service Training Note: In-service training refers to any type and duration of training received after having started working as a teacher. On the other hand, about three-quarters of both rural and urban teachers have received some kind of in-service training. The training received does however vary in duration, from one day to a year, and in terms of content. Common examples of in-service training course contents include: teaching methods for Grades 1/2/3 and Grades 4/5/6, how to prepare lessons, how to teach the Arabic language, health and nutrition and corporal punishment. Overall, there appears to be scope to review the in-service training system to ensure that all teachers receive the training they need, and to increase the focus on pedagogy and classroom management. Teacher Performance Management and Career Advancement For student to learn teachers need to be present in school. This section therefore examines teacher arrival and absenteeism, and school tracking of these in the four states. Once teachers are in school, for learning to take place requires that teachers are in the classroom rather than doing administrative work and that they teach in a manner conducive to learning. Therefore, this section also presents data on whether teachers prepare for teaching their classes, certain aspects of how they engage with their students, and the extent of teacher supervision. Instructional Time Instructional time is mainly affected by three factors, school closures, teacher (and student) absenteeism and tardiness, each of which influence student learning. Unplanned school closures 51 shorten the effective school year by 45 days in rural areas and 37 days in urban areas, on average in the four states (See Policy Note No.1 on service delivery). When teachers fail to show up for class students are subject to a loss of instructional time. n the course of the SDS2 survey, Grade 5 students were asked whether their teacher was present for their last math and reading classes. In rural schools, 19 percent reported that their teacher had been absent for math and similarly f reading). Relatively fewer urban grade 5 students reported their teacher absent: 16 percent for mathematics and 15 percent for reading. 38 Lateness was evaluated in the SDS2 survey by interviewing teachers. Most teachers (85 percent of rural and 83 percent of urban) reported that they had arrived on time the day of the survey, with most of those who were late having arrived within 30 minutes of the set time (88 percent of rural and 73 percent of urban teachers). Moreover, 3.3 percent of rural and 1.3 percent of urban teachers that they arrived more than one hour late. Regardless of the reasons for teachers arriving late, unless there are substitute teachers, instructional time is reduced. All urban schools reported that teachers have to report upon arrival at the start of the day, compared to 84 percent of rural schools. Teaching Practices Even if teachers are present in school and arrive on time, student learning does not necessarily take place. Teacher preparation for classes, student feedback, and choice of classroom activities also influence whether students learn and how much. The SDS2 data provide some insights into how teachers prepare for teaching. In the four states, 92 percent of urban teachers and 74 percent of rural teachers have a study plan. The difference may reflect the lower availability of resources for rural teachers. Most teachers also develop and use a lesson plan : 86 percent of urban teachers report always using a lesson plan, against 75 percent of rural teachers (See Figure 3.9). On the other hand, 3 percent rural teachers report never using a lesson plan compared to 2 percent of urban teachers. These data suggest that there are differences in how rural and urban teachers prepare for their classes. The fact that some teachers never use a lesson plan, although the shares are relatively small, may indicates to who to prioritize in the provision of in-service teacher training. Giving homework and marking it provides a further indication of teachers’ effort. As different teachers may give different types of homework that require different amounts of time to prepare, and because grading homework can vary in scope and quality, this indicator is imperfect. To be most conducive to student learning, grading should include the provision of correct answers for 38 Some of these teachers may have been absent for valid reasons. 52 questions students answered wrongly. Some recent evidence in fact shows that frequently teachers provide incorrect feedback, which has an adverse effect on student performance (Cueto et al., 2006). Whether teachers give and grades homework provides an indication of teacher effort. However, it is an incomplete indicator as different teachers may give different types of homework that requires different amounts of time to prepare, and grading of homework can vary in scope and quality. To be most conducive to student learning grading should include provision of correct answers for questions students answered wrongly; some recent evidence shows that frequently teachers provide incorrect feedback with adverse effects on student performance (Cueto et al., 2006). According to reporting by Grade 5 students, 10 8-9 percent of teachers never give math homework and 10-11 percent never give out reading homework, indicating a severe lack of opportunity for students to learn, to assess their performance and to provide feedback. More than half of teachers (60 percent of rural and 58 percent of urban teachers, on average) always grade the math homework they give (and similarly for reading), but at the other end of the spectrum, 7- 8 percent of teachers never do, and the picture is similar for reading homework. To gain an understating of what teachers do in the classroom the SDS2 asked whether teachers give their students exercises to complete in class as opposed to simply having students read or copy from the chalkboard. The expectation is that teachers who provide exercises spend more time preparing and interacting with their students. Among urban Grade 5 students, 21 percent reported that their teacher had provided them with math exercises during their most recent class and 18 percent that they been given reading exercises to complete in class. Among rural students in grade 5 relatively fewer completed exercises provided by the teacher: 18 percent for mathematics and 12 percent for reading. This suggests there may be differences in classroom activities across rural and urban schools, which has implications for learning and teacher training. Teacher Supervision Many basic schools in the four survey states have some form of supervision to assess teacher performance, although systems vary across states both in terms of the type of supervision and its frequency (See Table 3.10). Supervision is provided by locality inspectors, education councils, head teachers, or a combination of these. 53 Figure 3.10: Incidence of Teacher Supervision, by Type and Location, 2010 Percent Urban 43 12 (35) Council Inspector Teacher Head Rural 39 11 (43) This month Urban 18 10 (53) 1 month ago Education Locality Rural 9 11 (70) 2-3 months ago 4-6 months ago Urban 12 2 (79) (Never) Rural 13 5 (78) 0% 20% 40% 60% 80% 100% Source: SDS2 data. The most common forms of supervision, carried out by head teachers and locality inspectors, are more frequent in urban areas. Most teachers reported to have been visited in the classroom in the course of the preceding six months by their head teacher (57 percent of rural and 65 percent of urban teachers). Of these, 68 percent of rural and 66 percent of urban visits occurred within the last month. Visits by locality inspectors are slightly less common and the rural-urban divide is more contrasted: 47 percent of urban teachers reported having been visited by a locality inspector in the course of the six-month period, against 29 percent of rural teachers, and the remainder that they have never been visited by a locality inspector. Education councils also play a supervisory role in some schools: 21 percent of rural and 22 percent of urban teachers had received a visit by the education council during the preceding six months. Overall, teacher supervision appears to be more common in urban than in rural basic schools and although many schools have some type of supervision in place a large proportion of both rural and urban schools (32 percent of rural and 21 percent of urban schools) have no supervision at all in the form of head teacher, locality inspector or education councils visits. More importantly still, where supervision systems exist, they are not necessarily effective. On the one hand, there is no strong relation between teacher performance and incentives, and on the other, merit-based teacher career advancement and options to increase teacher motivation are limited due to open cadre promotion practice mentioned above. 54 Teacher Motivation The survey of the four states also covered reasons for promotion, teachers’ satisfaction with their salary and non-financial incentives. When asked the reason for their most recent promotion, 78 percent of urban teachers and 56 percent of rural teachers cited years of service (See Figure 3.11). Only 4-6 percent of teachers cited either locality inspector or head teacher recommendation as the basis for their promotion. Figure 3.11: Main Reasons Perceived by Teachers for their Promotion, by Location, 2010 Percent 100 Urban 78 80 Rural 56 Percent 60 37 40 20 11 6 4 5 4 0 Seniority Head Locality Other Teacher Inspector Recommendation Recommendation Source: SDS2 data. The majority of rural and urban teachers (88 percent and 85 percent respectively) reported that they had received their salaries in full the previous month. However, the share of teachers who have a second job (10 percent of rural and 11 percent of urban teachers) and the larger proportions of teachers that complement their income through private tutoring (31 percent of rural and 19 percent of urban teachers) imply that a many teachers seek to supplement their regular salary. Although financial incentives influence teacher performance, non-financial incentives can also be effective in attracting and retaining teachers and improving performance. For instance, teachers may exert more effort if it improves their standing in the local community, or if they are offered a positive and supportive work environment (World Bank, 2012). Other non-financial incentives that affect teacher motivation and performance in Sudan and elsewhere include access to training, the provision of teacher housing, availability of good schools for teachers’ own children and professional recognition. 55 The SDS2 survey provides some indirect evidence on this. The main challenges teachers cite in working in their current school include housing (44 percent of rural and 33 percent of urban teachers), no access to training (35 percent and 29 percent respectively), schooling for their own children (12 and 8 percent respectively) and the lack of community acceptance (8 percent and 6 percent respectively). Figure 3.12: Main Teaching Challenges, by Location, 2010 Percent Low Student Learning Lack of Management Support Housing No Access to Training Access to Safe Water Security Teachers' Childrens' School Urban Lack of Community Acceptance Rural Other 0% 10% 20% 30% 40% 50% 60% 70% Note: teachers were asked for the three main challenges working at their current school. The two main challenges cited by teachers however are the lack of management support (56 percent and 52 percent of rural and urban teachers respectively) and low levels of student learning (53 and 62 percent respectively). Access to safe water (42 and 27 percent respectively), and security (7 and 8 percent respectively) are also concerns. Thus, while some of the challenges are related to school infrastructure and resources one of the most important challenges according to teachers, management support, is not. Section 3: Summary and Policy Options The main challenges in teacher management in Sudan and the four survey states fall into three categories: (i) teacher deployment and utilization; (ii) teacher preparation and professional development; and (iii) teacher performance management and career advancement. 56 Key Findings Teacher Deployment and Utilization (i) Teachers are very unevenly distributed across schools; (ii) Once deployed, teacher utilization is uneven and relatively inefficient; and (iii) Actual instructional hours are significantly below the official number of contact hours. Teacher Preparation and Professional Development (iv) While the number of teachers with academic qualifications has risen substantially, professional skills development continues to constitute a challenge in Sudan; and (v) There appears to be quite widespread dissatisfaction with the professional skills that university trained teachers enter the profession with. Teacher Performance Management and Career Advancement (vi) While teacher supervision systems are in place in many schools and some states and appear to be working relatively well, the linkages to incentives are not optimal; and (vii) Merit-based teacher career advancement and teacher motivation are constrained by the practice of open cadre promotion, according to which teachers are automatically promoted according to their seniority, except for a few thresholds where an examination may be required. Policy Implications Recent trends in international research have focused on the factors that help countries make significant improvements in education system performance. In the teacher quality literature there is a focus on policy shifts, changes in the approach to key aspects of teacher management that are associated with significant improvements in system performance. Most common among these are five policy shifts, each of which may be particularly relevant in the case of Sudan, and some of which are already foreshadowed in steps that are being taken in some higher capacity states and localities. The main challenges in Sudan within each category are listed below, together with the policy shifts used by other countries to raise education system performance. 57 Policy Shift: From the Central Management of Teacher Supply to a Regulated Market More successful systems have been using or are moving towards approaches to teacher supply which depend less on comprehensive planning to determine teacher requirements and production quotas, and more on market forces, to help shape the supply of teachers to the system. Although market forces also operate in planned systems, they are often not taken into account for planning purposes, which instead are based on the assumption that teachers will be recruited and deployed according to an efficiency-based plan. Experience suggests that better information about the teacher market, appropriate regulations to ensure a level playing-field and appropriate incentives can help ensure a more reliable, motivated and responsive teacher supply. Policy Shift: From Central Teacher Deployment to Incentive-Driven Approaches Teacher deployment requires a range of strategies, especially to ensure the more equitable distribution of teachers, and changing teacher deployment patterns can take time. The use of incentives as part of the teacher deployment strategies has had considerable success and some good lessons on what works have been derived, in the area of use of incentives as part of the teacher deployment strategies. Policy Shift: From Qualifications and Experience to Competence and Performance While teacher academic qualifications and experience provide the mainstay of most teacher selection and career advancement systems, successful systems are increasingly also assessing the actual teaching skills of applicants and engaged teachers by measuring teaching performance in terms of outcomes. These assessments are often linked to various forms of professional registration or licensing. Policy Shift: From Teacher Quantity to Teacher Quality Almost all well-performing education systems have shifted their focus from producing the correct number of teachers with appropriate qualifications and specializations to the quality of the teaching staff. A landmark 2005 OECD study makes the point that “of those variables which are potentially open to policy influence, factors to do with teachers and teaching are the most important influences on student learning. In particular, the broad consensus is that ‘teacher quality’ is the single most important school variable influencing student achievement.” 58 Focusing on teacher quality means tending not only to training and professional development but also to motivation and career development, as well as to improving deployment and the effectiveness of teacher utilization. Policy Shift: From Performance Management Based on Inputs and Processes to Outputs and Learning Outcomes Performance management in successful education systems is increasingly linked to a number of outputs, ranging from teacher attendance at the simplest level, to student learning outcomes in more advanced systems. Effective performance management systems that use student learning outcomes as a key indicator require careful research and management to ensure that they operate in a fair and accountable manner. Ill-conceived or implemented schemes can have a negative impact on teacher morale and learning outcomes through teaching-to-the-test, while well designed interventions can have powerful positive outcomes. One thing that emerges from the Sudan context is that there is considerable potential for internal learning across states and localities within the education system. Some states and localities are already implementing best international practices in some areas and can provide strong lessons for other parts of the system. Moreover, the Education Sector Strategic Plan, 2012-16 currently being drafted will provide opportunities to ensure that interventions to improve teacher management are incorporated into consistent targets, strategies and indicators for the next five years. 59 Policy Note No.4 Assessing the Role of Education Councils in School Management in Four States of Sudan - Findings from the 2010/11 Basic Education Service Delivery Survey Abstract Education councils consisting of parents, teachers and community members are an essential component of the basic education system in Sudan. Localized management of schools has a strong potential to increase efficiency in the delivery of education services and ultimately raise student performance. However, successful community involvement in school management depends on the degree of autonomy conferred to schools, the responsibility and accountability arrangements in place and local capacity and empowerment levels. School-level data collected in four states of Sudan (Blue Nile, North Darfur, Red Sea and South Kordofan) in 2010/11 confirm the widespread presence of education councils and their central role in providing administrative support and coordinating the community financing of schools. At the same time, the findings suggest that a lack of essential records and direct authority make it difficult for education councils and communities to effectively engage in school management. For Sudan to realize the kind of benefits that international experience suggests school-based management can bring to improving education service delivery, and ultimately to student learning outcomes, strategies that enhance the authority and capacity of education councils are needed. 60 Introduction Education councils consisting of parents, teachers and community members are an essential component of most education systems around the world, as in Sudan. Many countries have strengthened the role of education councils in school management in the light of the perceived benefits of community involvement in running schools. While the evidence remains limited, evaluations of school-based management (SBM) programmes demonstrate that community participation can improve service delivery and in some cases student performance (See Annex 4.1 for international experiences with SBM). In Sudan, education councils play a major role in financing basic schools, alongside the government: they fund school infrastructure, teacher housing and salaries, and furniture and learning materials (World Bank, 2011). Councils also assist with administrative duties, follow up on students who have dropped out, engage in teacher supervision and, in some cases, assist with teaching. Given the importance of education councils for the operation of basic schools in Sudan, it is critical to understand their activities and how to utilize them to improve service delivery. The second Basic Education Service Delivery Survey (SDS2), 2010/11 in four states of Sudan (Blue Nile, North Darfur, Red Sea and South Kordofan) collected representative data on the characteristics of basic schools that provide insight into the involvement of education councils in their management (See Study Rationale and Survey Description). The aim of this note is to illustrate the responsibilities, practices and diversity of education councils in the four survey states and, when possible, to relate to existing international evidence on school-based management. Section 1 briefly describes different models of school-based management. Section 2 presents new evidence on education councils’ composition, activities and financing, based on SDS2 data. Finally, Section 3 concludes by considering priority issues and policy implications, based on the findings. Section 1: School-Based Management Models The devolution of decision-making powers from central authorities to the school level varies according to the degree of autonomy conferred and to whom autonomy is conferred. There are numerous functions that can be reassigned from the central to the school level including staff management, pedagogy, infrastructure and maintenance and budget monitoring and evaluation. Stronger SBM models include schools that receive direct fund transfers from the government, hire and fire teachers and determine curricula. In weaker models, schools have authority over select functions such as instructional methods and school planning. Autonomy of functions can 61 be conferred to school administrators, teachers, community members (typically parents), or a combination of teachers and parents. Decentralizing decision making to schools and communities can help improve education service delivery. Firstly, the localized management of public services can create a direct channel through which service receivers can provide feedback to providers. When parents are involved in school management they can bypass politicians and hold teachers and principals accountable for school resource management and performance. Secondly, the potential for more efficient resource allocation tends to be higher in locally managed schools. Generally, school staff and parents have more knowledge of school needs than non-local officials and can therefore distribute limited funding for school inputs more efficiently. Finally, introducing SBM is less costly than other interventions aimed at improving school and student performance (Barrera-Osorio et al., 2009). However, SBM is subject to a number of caveats. Head teachers and school staff may not be supportive of lessening their authority and raising their accountability. On the other hand, they might maintain community involvement at a superficial level, to avoid adding complexity to their jobs (Cook, 2007). At the same time, communities risk being subject to elite capture (Bardhan, 2002 and Bardhan and Mookherjee, 2000 and 2005) and may lack the capacity and empowerment to effectively manage schools. 39 In these cases, SBM may not effectively improve service delivery in practice. Education councils are a key component of SBM models. Despite the caveats related to effective community involvement, education councils have the potential to improve service delivery in education. Existing evidence points to three main potential benefits from SBM: (i) changes in school dynamics via increased parent involvement or altered teacher behavior; (ii) reduced repetition, failure and dropout rates; and in some cases (iii) improved test scores after SBM has been in place for a period of time (See Annex 4.1 for an overview of the available evidence). Section 2: Characteristics of Education Councils in the Four Survey States This section presents and discusses the findings on education councils in basic schools in the Blue Nile, North Darfur, Red Sea and South Kordofan states of Sudan, based on SDS2 data. A multi-stage sampling strategy was used to select a representative sample of 234 basic schools for the survey. The findings therefore reflect conditions in a fictitious government basic school of average characteristics for the four surveyed states. 40 The results below draw from the class and head teacher questionnaires, which captured information on school location, type, infrastructure, 39 Elite capture is when privileged local populations use the new powers conferred to councils to further their own objectives, potentially at the expense of the less privileged. 40 IDP schools in North Darfur were also surveyed but were not included in the sample used here (See Study Rationale and Sample Description). 62 enrollment, head teacher characteristics, teacher supervision, teachers, financing and education councils. Box 4.1: Nomadic Basic Schools in Sudan So-called nomadic schools account for 8.7 percent of basic schools and three percent of basic school enrollment in Sudan (World Bank, 2011). How to effectively provide education to nomadic children nevertheless remains a longstanding challenge. The first schools for nomadic children were boarding schools for boys located in central areas. They had limited success with their target population due to (i) low levels of awareness of the opportunities and benefits of education; (ii) the need for boys to travel with families in search of pasture and look after cattle; and (iii) schools’ high running costs. In the 1990s, UNICEF and the federal Ministry of General Education began to promote a mobile school system for nomadic communities. The Mobile Schools for Nomads Inclusive Education Approach has subsequently expanded throughout the country. Under this system, students in Grades 1 to 4 attend mobile multi-grade schools. The classrooms are typically composed of temporary materials, tents, straw, bamboo, or are simply under trees, and students from all four grades typically form one class. After Grade 4, nomadic students attend collective schools until Grade 8 (UNICEF, 2011). The strong cultural preference for early arranged marriages over schooling provides one barrier to girls’ access to education, whereas the need to travel with herds during the dry season prevents boys’ attendance, even though many nomadic populations have become How common and active are education councils? It is evident in the data that education councils are an essential part of the basic education system in the four survey states: 95 percent of schools have one (See Table 4.1 below). The proportion of schools with an education council is somewhat greater for urban regular schools (99 percent) than rural regular schools (92 percent) and rural village schools (94 percent). All surveyed nomadic schools have education councils. 63 Table 4.1: Presence of Education Councils, by Type of School and Location, 2010 Percent Urban Rural All Schools Regular Nomadic Regular Village Nomadic Share of Schools with an 99 100 92 94 100 95 Education Council Source: SDS2 data. In SBM, highly active education councils are most likely to be effective in improving the quality of service delivery (Bruns et al., 2011). The majority of education councils in the four survey states meet frequently. In urban schools 65 percent of education councils meet once a month, compared to 58 percent in rural schools (See Figure 4.1). However, six percent of urban councils and nine percent of rural councils had no meeting during the last 12 month period, indicating a substantial variation across councils with potential implications for their effectiveness in influencing school management. Figure 4.1: Frequency of Education Councils’ Meetings, by Location, 2010 Percent 70 65 58 60 Urban Share of councils that: 50 Rural 40 32 30 30 20 9 10 6 0 Meet once Meet once every Have not met in a month 2-6 months 12 months Source: SDS2 data. Who are the members of the education councils? The average urban council has 11 members, compared to 9 members for the average rural council. The councils are composed primarily of parents: 92 percent of rural councils and 94 percent of urban councils have parents as members (See Figure 4.2 below). Teachers’ level of participation is also high: they are present on 61 percent of rural councils and 73 percent of urban councils. This information furthermore suggests that teachers are more active in school 64 management in urban areas. Very few councils have members that are neither parents or teachers. Just six percent have other members, who are typically non-parent community members and local business representatives. Figure 4.2: Participation of Parents and Teachers in Education Councils, by Location, 2010 Percent 100 94 92 Urban Share of councils with: 80 73 61 Rural 60 40 20 6 6 0 Parents Teachers Other Participants Source: SDS2 data. The composition of education councils can pose challenges and even when roles are clearly defined, they may not be able to operate effectively: (i) From the perspective of school staff, the head teacher must remain accountable for the school, even when relinquishing some authority to parents. As a result, in some cases school staff limit community involvement, to avoid complications; (ii) The lack of empowerment of poor community members can prevent parents from criticizing head teachers and teachers, undermining the communities’ accountability function and putting the council at risk of suffering elite capture; and (iii) Education councils need to possess adequate skills, authority and information to function effectively. What activities do education councils engage in? Education councils’ activities are varied. Overall, the most common activity is financing, carried out by 63 percent of urban councils and 48 percent of rural councils (See Table 4.2 below). Urban councils’ ability to contribute relatively more funding has equity implications for the 65 education system overall, as urban communities tend to be wealthier than their rural counterparts and thus can purchase more resources such as textbooks, furniture and learning materials. A large proportion of education councils also assist with administrative duties (44 percent of urban and 45 percent of rural councils) and play an important role in following up on students who drop out of school (43 percent and 42 percent of urban and rural councils respectively). Table 4.2: Main Activities of Education Councils, by Location, 2010 Percent Share of councils Follow-up on Funding Administration Supervision Teaching involved in: Drop-outs Urban 63 44 42 27 14 Rural 48 45 43 23 9 All Schools 53 45 42 24 10 Source: SDS2 data. In addition, education councils are involved in the supervision of teachers in many schools. They provide inputs into locality officials’ teacher supervision reports in 27 percent of urban schools and 23 percent of rural schools. Finally, another less common activity is for councils to assist with teaching in the classroom (in 14 percent of urban schools and 9 percent of rural schools). Evidence suggests that education councils need real authority and means, such as the ability to hire and fire teachers and the power to allocate funding, to hold school management accountable and influence school performance. Several studies suggest that transferring the management of teachers to communities can be successful in improving student performance. 41 In the case of Nicaragua, the legal autonomy of schools had no effect on student test scores whereas increased decision-making power for education councils, in the form of control over teacher monitoring and staffing, was positively associated with students’ performance (See Annex 4.1). The SDS2 data does not indicate whether education councils have meaningful authority in managing schools in Sudan. It is known from other sources that whereas councils appear to have a widespread presence and role in coordinating community funding, their tangible authority and leverage over schools is weak at best. 41 See Annex 1 for examples of strong SBM interventions in El Salvador, Nicaragua and Nepal as well as of The Extra Teacher Program in Kenya. 66 How much funding do communities provide? Public spending on goods and services is low in Sudan, which creates funding shortfalls for schools’ running costs (World Bank, 2011). As a result, despite basic education supposedly being free in Sudan, households and communities contribute to funding schools, either directly or through education councils. The specific proportions of education council financing that originate from various types of fees and voluntary contributions cannot be determined from the SDS2 data however. The data do show that fees are commonly paid to education councils or directly to schools and that education councils play a key role in raising funds to help cover school running costs. Of those schools with councils, 63 percent reported charging registration fees, of SDG 10 to SDG 13 per student per year, on average (See Table 4.3). Although the schools that do not have education councils (that represent five percent of the sample) also collect fees, of SDG 9 per student per year on average, this is less common (48 percent of schools). Furthermore, although the SDS2 data clearly demonstrates the charging of fees, there is no evidence to suggest that the administration of fees is harmonized in any way. Table 4.3: Share of Schools that Charge Fees and Average Amount, by Location, 2010 Percent and Sudanese Pounds Share who Charge Average Annual Fee Registration Fees Per Student (SDG) Schools with Education Councils Urban 61% 13 Rural 64% 10 All Schools 63% 11 Schools without Education Councils All Schools 48% 9 Source: SDS2 data. In addition to fees, households in Sudan are estimated to spend an average SDG 15 per student towards school running costs, slightly above public spending per student, of approximately SDG 12. Households also pay for school uniforms, textbooks and meals so that average annual spending per student per year amounts to SDG 84 in urban areas and SDG 24 in rural areas (World Bank, 2011). 67 The existence of fees may hinder students from poorer households from attending school, thereby increasing basic education access iniquities. Among children aged 10 to 15 years, cash is the most common reason for not attending school. For those aged 6 to 10 years, it was cited more often than the distance from school, having to support the family and illness and disabilities (DHS, 2009). Moreover, rural households are less able to pay fees than their urban counterparts. Even when they pay lower fees, the burden on their household spending is proportionately larger, which exacerbates the regressive nature of shortages in education funding in Sudan. What do education councils fund? In Sudan, education councils provide a critical source of funding for basic schools. Council funding is mainly used for school feeding, furniture and learning materials and teacher salaries and training (See Table 4.4). School feeding is the most common type of spending for councils (38 percent of urban councils and 33 percent of rural councils support such programmes). Table 4.4: Education Council Spending, by Type and Location, 2010 Percent Share of councils School Learning Furniture Teacher Textbooks Teacher that fund: Feeding Materials Salaries Training Urban 38 18 21 7 8 1 Rural 33 29 25 21 20 5 All Schools 35 25 24 16 16 4 Source: SDS2 data. Education councils also pay for furniture and learning materials in many schools. Councils pay for furniture in one in four rural schools and in one in five urban schools. Learning materials are bought by 29 percent of rural councils and 18 percent of urban councils, and textbooks by 20 percent and 8 percent of rural and urban councils respectively. These findings strongly suggest that public funding for goods and services is insufficient, especially in rural schools. Rural councils are more likely than urban councils to contribute to teachers’ salaries and training. Specifically, 21 percent of rural councils help pay teacher salaries, against 7 percent of urban councils; the proportions of each that contribute to teacher training are five percent and one percent. That a much larger proportion of rural councils provide funding for teachers is partly the result of the uneven allocation of teachers across schools (World Bank, 2011). On the other hand, it is difficult to detach government teachers to work in remote rural areas, so that many rural schools depend on councils and other community funding to recruit and pay local teachers. 68 Record-Keeping and School Accountability Education councils contribute critical funding to schools, although often without the monitoring and evaluation that transparency and accountability would justify. Less than half of urban schools (46 percent) and less than a fifth of rural schools (19 percent) keep up-to-date financial records. 42 This lack of financial records implies limited accountability for resource allocation and use. Available evidence suggests that clear roles and the allocation of responsibilities in school management are crucial to successful SBM. Generally, improvements are facilitated when there is a clear definition of who assumes what authority and all stakeholders support the agreed responsibility arrangement. In the four survey states, no particular person or agency held the responsibility for checking schools’ financial accounts in about half of schools (See Figure 4.3). Education councils are more active in checking financial accounts in urban schools (43 percent) than in rural schools (31 percent). Figure 4.3: Responsibility for the Verification of School Financial Accounts, by Location, 2010 Percent 60 54 Share of school accounts 50 43 Urban 40 40 Rural checked by: 31 30 20 11 8 8 10 4 1 1 0 Nobody Education School Local Other Council Staff Authority Source: SDS2 data. The availability and status of school records is an indicator of the potential for accountability in school management. Overall, record keeping is weak although urban schools are generally better at keeping records than rural schools. 42 These rates seeming particularly low, the survey enumerators double-checked all records. 69 Without accounting for other factors that may influence the extent of record keeping, schools with active education councils are those most likely to have up-to-date school records (See Figure 4.4). 43 Figure 4.4: Probability of Schools Keeping Up-to-Date Records, by Type of Council and Location, 2010 Percent 60 Share of Schools with Up-to- 48 50 Urban 40 Rural Date Records 28 30 23 20 10 1 0 Active Inactive Council Council Source: SDS2 data. Schools with active councils are considerably more likely to keep up-to-date records than schools with inactive councils, both in urban areas (48 percent, against 28 percent of schools with inactive councils) and especially in rural areas (23 percent, against just 1 percent of schools with inactive councils). Of the different types of records that schools typically keep, up-to-date teacher leave records are held at only 35 percent of urban and 16 percent of rural schools (See Figure 4.5 below). In comparison, up-to-date student performance records are available at 58 percent of urban and 41 percent of rural schools. Enrollment records are those most frequently kept and up-dated, by 82 percent of urban schools, but just 48 percent of rural schools. 43 Councils reporting to have met once or more during the last six months are classified as active; those not having met during the six months preceding the survey are classified as inactive. 70 Box 4.2: Village and Nomadic School Education Councils Village schools and nomadic schools face greater challenges in providing education services than regular basic schools. The SDS2 survey identifies some of the differences related to education councils. Village Schools Village school councils are typically less active than councils in regular rural schools: just 31 percent contribute funds to schools, compared to 60 percent of regular rural school councils. They are also less likely to perform most other activities: only 33 percent follow up on drop-out students and none participate in teacher supervision or help with teaching. Among the village school councils that provide funding, fewer pay teacher salaries (13 percent) or buy textbooks (13 percent) than regular rural councils (21 percent and 24 percent respectively). This may be the result of village schools generally being located in poorer areas where the capacity to provide funding is lower although the needs are arguably greater. Record keeping at village schools is weaker than in regular rural schools. For enrollment records the difference is slight: about half of village schools have up-to-date records, against 55 percent of regular rural schools. However, the proportion of village schools with up-to-date student performance records is considerably lower (28 percent keep such records, against 39 percent of regular rural schools) and no village schools have up-to-date teacher leave records, whereas 24 percent of regular rural schools do. Nomadic schools Many nomadic education councils engage in a range of school activities. Many more help with administration (65 percent, against 45 percent of rural schools on average) and contribute to teacher supervision reports (53 percent, against 23 percent of rural schools on average) than regular schools. However, relatively fewer nomadic education councils contribute funds to schools (41 percent) than councils in regular schools (48 percent). It is more common for nomadic councils to allocate funding to teacher salaries (25 percent, against only 21 percent for regular rural schools) and teacher training (12 percent, against 5 percent for regular rural councils). By contrast, a much smaller proportion of nomadic councils allocate funding for textbooks and furniture than councils in regular rural schools. Nomadic schools are more likely to have up-to-date records than regular rural schools. A 71 Figure 4.5: Share of Schools with Up-to-Date Records, by Type of Record and Location, 2010 Percent 100 82 Urban Share of schools with: 80 58 Rural 60 48 41 40 35 16 20 0 Enrollment Student Teacher Records Performance Leave Records Records Source: SDS2 data. The availability of up-to-date and accurate records of vital information is essential for efficient planning, resource allocation and monitoring of school performance. Improvements to school- level record keeping systems also increase transparency and enable communities to hold schools accountable for performance. The current state of record keeping in schools in the four survey states is inadequate to foster evidence-based community involvement. Without substantial improvements in the availability and reliability of school records it is difficult for communities to effectively engage in the management of schools. Section 3: Key Findings and Policy Implications The SDS2 findings provide a snapshot of education councils and their main activities in basic schools in the four survey states. The vast majority of basic schools have an education council and most councils meet frequently. Education councils in Sudan are an important source of funding for schools and also engage in a range of support activities. Key Findings (i) The majority of education councils contribute funds to basic schools to help reduce shortfalls in public education spending; 72 (ii) The main uses of education council funding include paying for school feeding, learning materials, textbooks, teacher salaries and teacher training; (iii) Education council and other community funding provision is more common in urban than rural areas, contributing further to the existing rural-urban resource gap in basic schools; (iv) Education councils frequently help with administrative duties; follow up on students who have dropped out; provide inputs into locality officials’ teacher supervision reports and in some cases also help with teaching; (v) School financial accounts are more likely to be up-to-date in schools with active education councils; and (vi) Enrollment, student performance and teacher leave records at the school level are frequently unavailable or inadequate, making it difficult for education councils to effectively engage in school management. Policy Priorities The SDS2 findings make it clear that education councils play a significant role in providing funding to and managing basic schools in the four survey states. In many cases however, although councils provide administrative support to schools they do not seem to play any role in the allocation and management of school resources, including teachers. Combined with the general lack of school records containing the information required for planning and monitoring purposes, the scope for education council involvement in running basic schools is therefore currently limited. This study does however show that Sudan has valuable existing institutional infrastructure in the form of education councils, in almost all schools in the four survey states. This constitutes a significant asset upon which to build, to enable communities to effectively engage in school management. If Sudan is to fully realize the kind of benefits that international experience suggests school- based management can bring to improving access, retention and the quality of student learning, strategies are needed that will formalize the responsibilities of education councils, build their capacities and transfer authority to them. These may call for a variety of interventions that are each suited to the wide range of contexts in which education councils operate. Such interventions must also fully take the risks highlighted by international evidence into account, such as that of elite capture, or sabotage by officials or school leaders. Finally, the findings highlight the substantial differences in education council activities and capacities across schools within rural areas and across school types. They underline the importance of carefully targeting lagging schools within rural and urban areas, and allocating additional funding to certain school types. 73 Policy Note No.5 Challenges in Basic Education Provision for Internally Displaced Children in North Darfur - Findings from the 2010/11 Basic Education Service Delivery Survey Abstract The delivery of basic education to internally displaced persons (IDPs) poses a unique set of challenges in Sudan. International evidence shows that security, poor quality learning environments, language barriers and the cost of schooling hinder displaced children from accessing basic education. The 2010/11 Basic Education Service Delivery Survey (SDS2) covered four states in Sudan including North Darfur, where IDPs account for 18 percent of students in basic education. Displaced students in North Darfur attend schools that are on average larger, have higher student-teacher ratios and larger class sizes than non-IDP basic schools in the four survey states. However, teachers at IDP schools are generally more academically qualified and school infrastructure is better than in non-IDP schools. After taking into account observable school, student and household characteristics and inputs, Grade 5 refugee students perform significantly worse in mathematics than their non-IDP peers, which underscores the special needs of displaced students that need be addressed to improve education service delivery and student learning. Furthermore, investments are needed in classroom equipment, learning materials and pre and in-service teacher training, just as efficiency in the use of existing resources, including the teacher distribution policies, could be improved. 74 Introduction Conflict in Darfur has resulted in large-scale human displacement. Before the independence of South Sudan in 2011, Sudan hosted the largest number of displaced persons of any country in the world. 44 In the Darfur states alone, the Internal Displacement Monitoring Center estimated that 2.6 million people had been displaced in January 2011, or one third of the total population. In the North Darfur state, 18 percent of students attend IDP schools, compared to only 1 percent of students in Sudan as a whole. 45 The negative social and economic impacts of displacement are widely documented. The most relevant to mention here is the interruption of schooling for children, which heightens vulnerability levels and signifies a further barrier to overcome in the future. On the other hand, education can provide multiple additional benefits to IDPs in particular and society at large. On the individual level, in addition to the economic benefits, schooling can initially provide a sense of normalcy and help children integrate into their new communities; upon their return, it allows displaced children to remain in classes with children their own age. For the benefit of society, schools provide a forum to communicate critical information to IDPs such as on risks of land mines and HIV/AIDS and can provide physical safety from attack, sexual violence and the military recruitment of children. Above all, schooling has the potential to nurture long-term peace building in conflict-afflicted settings. International law places the primary responsibility for addressing the needs and protecting the rights of IDPs with the state. Although Sudan’s national policy on internally displaced persons acknowledges this, the de facto protection of IDP rights in Darfur has not been fully evidenced (Jaspars et al., 2010). The continued vulnerability of displaced persons in the Darfur states warrants attention to the provision of education services in IDP camps as part of the short and medium-term solution to the challenges of displacement. With the aim of enabling an evidence-based education policy for IDPs, this paper presents data from the second Basic Education Service Delivery Survey (SDS2), which surveyed IDP schools in North Darfur in 2010/11 as well as non-IDP schools in this and three other states. 46 Section 1 offers background information on characteristics, governance and service delivery in IDP camps and presents general challenges in providing education to displaced children. Section 2 compares the characteristics of IDP basic schools in North Darfur to those of non-IDP basic schools in the 44 This paper uses “displaced” to refer to Sudanese living in North Darfur who have moved involuntarily from their homes. This is based on the United Nations definition, which differentiates between IDPs and refugees, the latter having crossed a border to escape a conflict. 45 Annex 5.1 provides further education indicators for North Darfur in national perspective. 46 Of the 39 IDP schools in the North Darfur localities, 25 were selected for the sample. However, six of the schools could not be visited due to violence. Consequently, 19 IDP schools (49%) were surveyed (See Study Rationale and Survey Description). 75 four survey states. Finally, Section 3 discusses priorities based on the SDS2 findings and potential next steps to improve education service delivery and learning outcomes in IDP schools. Section 1: Overview of IDPs in Darfur There were an estimated 2.6 million IDPs in the Darfur states in 2011, eight percent of the total population of Sudan (IDMC, 2012). In order to improve their safety and access to basic services a substantial proportion of these IDPs have moved to urban camps. Over the course of eight years of conflict, one third of IDPs in Darfur have become long-term camp residents. The urbanization in camp settings has fundamentally changed the social fabric of Darfur, increased water use, shifted land use patterns and slowed local integration. General Characteristics An assessment of three camps near El Fasher (the capital) in North Darfur in 2007 provides some insight into the characteristics of and social services provision in IDP camps in Darfur (Adam, 2007). In these three IDP camps, the average number of residents is 50,000, over half of which are women and children and the average family comprises five to seven people. Some of the key findings include: (i) Income-generating activities are few and far between and place IDPs at risk of attack, especially outside the camp (farming, firewood collection) and for women. Access to markets varies widely among camps and generally offers few opportunities to IDPs given the limited production of cash crops. (ii) The main form of local government is an IDP committee composed of elected members, responsible for assisting with camp management, donor and NGO coordination, the distribution of assistance, rations and water, camp structure and organization and protect the health and interests of its residents. (iii) Significant disparities exist in the appropriation of scarce resources, exacerbating extreme poverty and difficulties in access to food and social services, especially among new arrivals. (iv) Social services were provided mainly by national and international NGOs and aid agencies, focusing primarily on food distribution, but also offering health and education services, hygiene, sanitation and human rights awareness building. (v) While health care is provided free of charge, health clinics face drug shortages and absenteeism of workers. Other frequently cited gaps in available services include 76 secondary education and vocational training, support for income-generating activities and the poor availability of water. (vi) The expulsion of international organizations in 2009 further damaged service delivery and around 14,000 local aid agency employees lost their jobs (Jaspars, 2010). Although local organizations have better knowledge of and access to local communities and camps, they sometimes lack the capacity and finances necessary to provide reliable services. Access to Basic Education Services Compared to other children, displaced children often face added barriers to accessing education, which include: (vii) Greater economic responsibilities, to help support families struggling with temporary living arrangements, transport water or assist in securing food rations; (viii) Security, especially for students whose schools are outside the camp. Government presence and protection is rare in the Darfur states; (ix) Language, for ethnic minorities and marginalized groups that do not speak the main language and who will benefit little from instruction as a result; (x) Missing documentation, as registration is required in many IDP camps in Darfur for school enrollment and community schools require proof of permanent residence; (xi) The cost of schooling, which is strongly correlated to the demand for education services, can be prohibited for impoverished IDP families; 47 and (xii) Excessively overcrowded school infrastructure and exacerbated lack of basic supplies, although this does not appear to be the case of IDP schools in North Darfur. Section 2: Comparing IDP and non-IDP Basic Schools in the Four Survey States In placing IDP basic schools in perspective, this section reviews data analyzed in each of the accompanying policy notes in this publication, that cover school infrastructure and the learning environment, student learning outcomes, teacher characteristics and teaching practices and education councils. 47 The Sudan Demographic and Household Survey of 2009 found that “no money” was a major reason among households for children not attending basic education. Although basic education is supposed to be free, the cost of attendance typically involves households paying between SDG 58 and SDG 140. 77 School Infrastructure The quality of school infrastructure across Sudan is generally low (See Policy Note No.1 on service delivery). Approximately half of classrooms are in need of rehabilitation or complete replacement (FMoGE, 2008). IDP schools in North Darfur in contrast, tend to have better infrastructure than non-IDP schools in the four survey states, primarily because basic school construction is largely funded by international aid. Indeed, it is estimated that 85 percent of all schools in North Darfur receive external support for construction (Lloyd et al., 2010). All IDP schools have a roof and walls, whereas 12 percent of non-IDP classes in rural areas are held under trees, and even in urban areas 3 percent of non-IDP schools have no walls (See Table 5.1). Although they are mostly built of thatch, grass, or mud (95 percent of roofs and 84 percent of walls) whereas most non-IDP schools are built of more resistant materials such as brick walls and zinc roofs, these materials are suitable to the local context. Table 5.1: Comparative Condition of IDP School Infrastructure, 2010 Percent North Darfur Four Survey State Non-IDP Schools IDP Schools Urban Rural Roof material Thatch/Grass/Mud 95 17 32 Zinc 0 72 48 Other 5 12 8 Sit under tree 0 0 12 Wall material Thatch/Grass/Mud 84 15 26 Bricks/Blocks 11 80 47 Other 5 2 14 No Walls 0 3 13 Source of Drinking Water Pump 18 12 20 Tap 12 15 1 Well 6 13 14 Other 29 18 15 No Drinking Water 35 42 50 Schools with Latrines 95 81 63 Source: SDS2 data. Furthermore, students in IDP schools generally have better access to drinking water and latrines than their non-IDP counterparts. Only 35 percent of IDP schools lack a source of drinking water compared to 50 percent of rural non-IDP schools and 42 percent of urban non-IDP schools. As for rural schools, the most common sources of drinking water for IDP schools are pumps (18 78 percent), although 12 percent have taps and a majority relies on other sources, including delivery by donkey. The majority of IDP schools in North Darfur also have latrines (95 percent), compared to just 63 percent of rural schools in general. Learning Environment The learning environment offered by IDP schools in North Darfur is below the national average in almost all respects. The average number of effective days schools are open for class in IDP schools in North Darfur is similar to that of rural schools in the four survey states generally (164 days and 165 days respectively), significantly lower than the official length of the school year, of 210 days and below the average number of days urban schools open (173 days). School and Class Size IDP schools are generally much larger than non-IDP schools and have higher student-teacher ratios. The average IDP school in North Darfur enrolls 893 students compared to 230 students for the average rural school and 462 students for the average urban school in the four survey states (See Table 5.2). Student-teacher ratios are also typically significantly higher in IDP schools, at 47:1 on average, compared to 34:1 on average for non-IDP schools. Table 5.2: IDP School Size, Class Size and Student-Teacher Ratios (School Averages), 2010 North Darfur Four Survey State Non-IDP Schools IDP Schools Urban Rural Average Enrollment (Number of Students) 893 462 230 Student-Teacher Ratio (All Teacher Types) 47:1 32:1 36:1 Class Size (Number of Students) 76 56 40 Source: SDS2 data. The average class size is also much larger in IDP basic schools. With 76 students, it reaches almost twice the size of an average rural class and a third larger than the average urban class. To put this figure in further perspective, it is worth considering that within a range of 40 to 60 pupils, the relationship between class size and student learning is typically weak, but class sizes of more than 60 students tend to be outright detrimental to student learning (Majgaard and Mingat, forthcoming). 79 Furniture and Learning Materials Fewer classrooms in IDP schools have chalkboards, and desks and textbooks are less common than in non-IDP schools. Only around half of IDP school classrooms have chalkboards in usable condition (47 percent of Grade 1 and 53 percent of Grade 5 classrooms), considerably fewer than in rural and especially urban non-IDP schools in the four states (See Figure 5.1). The availability of desks in IDP schools is worse still: just 17 percent of Grade 1 students and 28 percent of Grade 5 students have one to sit at. Again this is lower than in rural and urban non- IDP schools in the four survey states, where the average rates are about the double (30 percent for Grade 1 students and 60 percent for Grade 5 students). Figure 5.1: IDP School Availability of Chalkboards and Desks in Context, Grade 1 and Grade 5, 2010 Percent 80 72 70 61 64 60 53 53 IDP 47 49 39 Urban Percent 40 28 Rural 17 21 20 0 Grade 1 Grade 5 Grade 1 Grade 5 Share of Classes Share of Students with Chalkboard with Desk Source: SDS2 data. Sudan also faces a general and severe shortage of textbooks in many schools, which is exacerbated in the case of IDP schools in North Darfur (See Figure 5.2 below). Only 10 percent of Grade 1 students have a math textbook (equivalent to an average student to textbook ratio of 10:1), the worst coverage nationwide. About 20 percent have a reading book, a similar level to the supply of textbooks for Grade 5 students (equivalent to an average student to textbook ratio of 5:1). These averages are starkly short of the official student-textbook ratio of 2:1 and significantly worse than the student-textbook ratios for non-IDP basic schools in the four survey states (FMoGE, 2008). The supply of pens/pencils to Grade 5 students shows a similar pattern (77 percent of students have one, below the average for the four survey states of 87 percent), but 80 Grade 1 pupils on the other hand are comparatively better endowed (96 percent have a pen/pencil, against 84 percent on average for the four survey states). Figure 5.2: Availability of Textbooks and Learning Materials in IDP Schools, Grade 1 and Grade 5, 2010 Percent 120 IDP 96 100 84 84 89 85 Urban 77 80 Percent 60 Rural 32 31 31 31 36 33 40 25 25 20 21 21 20 10 0 Grade 1 Grade 5 Grade 1 Grade 5 Grade 1 Grade 5 Share of Students Share of Students Share of Students with a Math Book with a Reading Book with a Pen/Pencil Source: SDS2 data. Student Performance The SDS2 survey contained a learning assessment that tested Grade 5 students in mathematics and reading in the four survey states (See Policy Note No.2 on student learning). The math assessment included 30 multiple choice questions with four answer options. The average learning assessment score for IDP students in North Sudan was only 27 percent, indicating that the average student’s level is very low (students who randomly guess all answers might score 25 percent) although in line with the average throughout the four survey states, of 28 percent. However, the results from the multivariate regression analysis that controls for school inputs, teacher experience and activity, age, gender and household wealth, show that IDP students tend to perform significantly worse than both rural and urban non-IDP students in math (see Annex 5.2). That is, after removing the effect of selected school, student and household characteristics, IDP students on average perform significantly worse in mathematics than their non-IDP peers. Overall, the model including observable school inputs and student and household characteristics as determinants of learning outcomes explained only about 30 percent of the variation in student performance. Although these observable factors are important, unobservable factors such as students’ inherent abilities and effort and household characteristics are even more so, accounting 81 for 70 percent of the variation in performance. Furthermore, in the case of IDP students, many are orphans, have suffered traumas and have lost months or even years of schooling, which affects their learning levels. The weaker performance of IDP students compared to non-IDP students clearly indicates the importance of addressing their specific needs in addition to the general size of classes, lack of learning materials and excessive school closure days. Teacher Characteristics The quality of the teaching staff in IDP schools in North Darfur is generally above the average of those working in regular schools in the four survey states, at suggested by their status, experience and academic qualifications and pre-service training. 48 The majority of teachers in IDP schools are government appointed teachers (95 percent, compared to the average of the four survey states of 84 percent). Volunteers account for only 2 percent, a minute proportion in comparison to both urban and rural non-IDP schools in the four survey states, where they represent 14 percent and 16 percent of basic school teachers respectively (See Figure 5.3). Figure 5.3: Distribution of IDP Teachers by Contract Type, 2010 Percent 100 95 84 84 IDP 80 Urban 60 Rural Percent 40 20 14 16 2 2 2 1 1 0 Official Volunteer National Trainee Service Source: SDS2 data. 48 See Policy Note No.3 on teacher management. 82 On average teachers in IDP schools have 13 years of work experience, compared to 9 years in rural non-IDP schools and 13 years in urban non-IDP schools. Teachers in IDP schools are generally also more academically qualified. The vast majority of teachers in IDP schools in North Darfur (81 percent) have a bachelor’s degree, compared to less than half of teachers in rural and urban non-IDP schools in the four survey states (41 percent and 45 percent respectively – See Figure 5.4). The reason for this large difference is not clear. One possibility is that many teachers in IDP schools are recruited and financed by NGOs, which may have higher hiring standards. Figure 5.4: Distribution of IDP Teachers by Academic Qualification (School Averages), 2010 Percent Basic or 81 3 Intermediate IDP 10 6 Secondary Education 45 2 Non-Teaching Urban 40 12 1 Diploma BA/BSc 41 1 Rural 51 4 3 Postgraduate Degree 0% 20% 40% 60% 80% 100% Source: SDS2 data. Finally, in addition to having better status, experience and academic qualifications than their counterparts in non-IDP schools in the four survey states, more IDP school teachers have some kind of pre-service teacher training, although the overall proportion is still too low. About 61 percent of IDP teachers have no pre-service training, compared to the average for non-IDP schools of 70 percent, and of those with some teacher training, a higher share have a bachelor of education degree (See Figure 5.5 below). 83 Figure 5.5: Distribution of IDP Teachers by Pre-Service Training Qualifications, in Context (School Averages), 2010 Percent IDP 14 25 61 None Urban 21 9 Teaching 70 Diploma 23 Bachelor of Rural 70 7 Education 0% 20% 40% 60% 80% 100% Source: SDS2 data. Head Teachers Head teachers play an important role in providing instruction in addition to managing schools. In IDP schools in North Darfur they are relatively more qualified than head teachers in non-IDP schools in the four survey states, both in terms of academic qualifications and pre-service teacher training (See Table 5.3). Table 5.3: IDP Head Teacher Academic Qualifications and Pre-Service Training, 2010 Percent North Darfur Four Survey State Non-IDP Schools IDP Schools Urban Rural Academic Qualifications Secondary 5 26 43 Non-Teaching Diploma 5 15 10 BA/BSc 79 54 42 Postgraduate Degree 11 3 1 Pre-Service Training Teaching Diploma 58 48 35 Bachelor of Education 0 8 14 None 42 44 52 Source: SDS2 data. 84 Indeed, 79 percent have a bachelor’s and 11 percent a postgraduate degree compared to 54 percent and 3 percent respectively of head teachers in urban non-IDP schools and 42 percent and 1 percent of head teachers in rural, non-IDP schools. Pre-service teacher training is also more common among head teachers at IDP schools, although still insufficient: 58 percent are trained, compared to 56 percent of their counterparts in urban non-IDP schools and 48 percent in rural non-IDP schools. Teaching and Classroom Activities Teachers’ efforts, as measured by the share of them that grade students’ homework and those that set exercises during class, are deemed to be valuably indicative of the quality of teaching, as feedback and interaction with students have a positive influence on student performance . The grading of student homework varies considerably as a practice across IDP schools. Reports of Grade 5 IDP students indicate that 10 percent of their teachers never set homework on average, 11 percent never grade it and 25 percent only grade it sometimes, meaning that only 54 percent of teachers both set homework and systematically correct it. This general profile is similar, although marginally better, to trends in non-IDP schools in the four survey states (See Table 5.4). Table 5.4: IDP Teacher Effort and Classroom Activities with Grade 5 Students, 2010 Percent North Darfur Four Survey State Non-IDP Schools IDP Schools Urban Rural Homework grading Always 54 56 59 Sometimes 25 28 25 Never 11 6 8 Do not get homework 10 10 10 Main activity during last class Read from chalkboard 30 — — Read from textbook 25 — — Copy from chalkboard 26 — — Complete math exercises given by teacher 22 21 18 Complete reading exercises given by teacher 17 18 12 Source: SDS2 data. Note: Based on student survey responses. 85 The SDS2 data on classroom activities indicate that as is common throughout the four survey states, the use of exercises in class is all too infrequent. Only 17 percent of students are set reading exercises and 22 percent are set math exercises. The most common classroom activities are to read from the chalkboard (30 percent on average), read from textbooks (25 percent on average) and copy from the chalkboard (26 percent on average). The relatively small proportions of students doing exercises in class may be explained by the lack of pre-service and in-service teacher training, the contents of teacher training received and/or the shortage of textbooks and learning materials. Education Councils The government of Sudan strongly supports the involvement of education councils in school management, and indeed they are pervasive: 94 percent of rural non-IDP schools and 99 percent of urban non-IDP schools in the four survey states have one (See Policy Note No.4 on education councils). In comparison, all of the surveyed IDP schools in North Darfur have an education council. The key activity they are involved in is the provision of funding for school’s operational costs. Indeed, this is significantly more common in IDP schools than regular rural or urban schools: 84 percent of education councils in IDP schools in North Darfur provide funding, compared to 48 percent of councils at rural non-IDP schools and 63 percent of councils at urban non-IDP schools in the four survey states. Table 5.5: Items Funded by IDP Education Councils, 2010 Percent North Darfur Four-Survey State Non-IDP Schools IDP Schools Urban Rural Share of Schools with an Education Council 100 99 94 Share of Schools where Education Councils Finance: School Feeding 79 38 29 Learning Materials 53 18 25 Textbooks 42 9 18 Teacher Salaries 26 7 20 Furniture 26 21 23 Teacher Training 11 1 4 Source: SDS2 data. 86 The funding is used to finance a variety of activities. The main activity in IDP schools is school feeding, which is important for generating demand for schooling and enabling students to learn. In North Darfur, 72 percent of IDP school councils pay for school feeding, a vast majority. Many also provide funding for learning materials (53 percent) and textbooks (42 percent), whereas less common items include furniture (26 percent), teacher salaries (26 percent) and teacher training (11 percent). Moreover, it is more common for IDP school councils to provide financing for each of these items than for councils in non-IDP schools in the four survey states to do so. This suggests that there is a shortfall in public funding for IDP schools, which is partly met by funding from education councils, which has implications for equity in access to basic education as richer communities can provide relatively more funding than poorer communities to help finance school running costs. School Record Keeping School records of enrollment, student performance and teachers’ leave are a vital tool for monitoring school performance, making informed school management decisions and holding education providers accountable for performance. Overall, record keeping is weak in IDP schools. Table 5.5: IDP School Record Keeping, by Type of Record, in Context, 2010 Percent North Darfur Four-Survey State Non-IDP Schools IDP Schools Urban Rural School Enrollment Records Seen + Up-to-date 58 82 47 Seen 11 6 11 Not available 16 5 8 None kept 16 7 35 Student Performance Records Seen + Up-to-date 47 58 42 Seen 5 6 2 Not available 16 9 15 None kept 32 27 41 Teacher Leave Records Seen + Up-to-date 37 36 16 Seen - 10 1 Not available 11 19 5 None kept 53 35 77 Source: SDS2 data. 87 The proportion of IDP schools with up-to-date records varies by type of record: 58 percent of schools have up-to-date enrollment records, 47 percent have up-to-date student performance records and 37 percent have up-to-date teacher leave records (See Table 5.5 above). The practice is somewhat more common than in rural non-IDP schools, of which 47 percent keep up-to-date enrollment records, 42 percent keep up-to-date student performance records and 16 percent keep up-to-date teacher leave records; they are also more likely to keep no records at all. Urban non- IDP schools on the other hand were generally more likely to keep records than IDP schools: 82 percent keep up-to-date enrollment records, 56 percent keep up-to-date student performance records and 36 percent keep up-to-date teacher leave records. Section 3: Key Findings and Policy Implications IDP children face a number of barriers to accessing and benefitting from education. Displacement-induced challenges include added economic responsibilities for children, administrative barriers, physical insecurity and psychological trauma. Key Findings Many IDP schools in North Darfur, like many regular schools in the four survey states, are characterized by resource shortages on a scale that has serious implications for student learning outcomes. School Infrastructure and Learning Environment (i) IDP schools in North Darfur tend to have better infrastructure than non-IDP schools in the four survey states; (ii) Students of IDP schools more frequently have access to drinking water and latrines than their counterparts in non-IDP schools; (iii) Although the official length of the school year in Sudan is 210 days, the average number of school days in IDP schools is 164 days, lower than the average 171 days that non-IDP schools in the four survey states open every year; (iv) IDP schools tend to be much larger in size, have higher student-teacher ratios and larger class sizes than non-IDP schools; (v) Fewer classrooms in IDP schools have chalkboards and fewer students have desks than in non-IDP schools; and 88 (vi) The textbook shortage in IDP schools is more severe than in non-IDP schools. Teachers and Student Performance (vii) Teachers in IDP schools in North Darfur are generally more academically qualified than teachers in non-IDP schools in the four survey states although the reason for this is not clear; (viii) As for non-IDP schools, the majority of teachers in IDP schools have no pre-service teacher training and little in-service training; (ix) Teacher efforts, as measured by whether teachers grade student homework, varies substantially across IDP schools; and (x) Only a small proportion of students do exercises in class, indicating that many teachers may not actively engage their students in the classroom. Policy Implications The findings from the SDS2 survey of IDP schools in North Darfur enable the identification of the priorities for improving service delivery in basic education for IDP children. Taken together, the findings imply fewer materials to enable effective teaching, less instructional time and less scope for teachers to interact with IDP students, with adverse implications for student learning. To address resource shortages a coordinated effort is needed by the federal government, state governments, communities and donors, especially considering the financial support a large number of IDP schools already receive from education councils and communities in order to meet shortfalls in public spending. The difference in student-teacher ratios across IDP schools in North Darfur and non-IDP schools is a reflection of the inefficiency of teacher distribution policies as teachers are frequently not allocated according to the number of students. Just as this policy needs to be reviewed to improve service delivery, in North Darfur specific policies may also be needed to attract teachers, given the security concerns. The lack of pre-service and in-service teacher training and the contents of teacher training, combined with the shortage of classroom equipment and learning materials poses a significant challenge to providing effective instruction and student learning. To raise teaching quality and student learning it is vital to prioritize the development of teachers’ skills rather than the current focus on upgrading qualifications. This would involve the introduction of well-designed school and cluster-based professional development approaches, using mentor teachers, peer learning and effective monitoring and feedback to support the development of good teaching practices. 89 The weaker performance of IDP students on the mathematics assessment even after accounting for resource shortages clearly indicates the importance of addressing the specific needs of IDP in order to improve student learning. All IDP schools have education councils and these constitute a valuable asset to improve resource use and strengthen school management and performance. 90 Annexes Annex 0.1: Sample Characteristics The unit response rates were high: in Blue Nile and South Kordofan all sample schools were visited and in Red Sea 96 percent of the sample schools. Given the security situation, the response rate was lower in North Darfur: 77 percent of the selected non-IDP schools were visited. Government Basic Education Schools: Population and Sample % of % of schools in % of N. Sudan Schools in Sample schools in sample sample government Government basic Schools in sample Sample schools state localities schools basic schools: state* localities schools visited visited* visited visited schools Blue Nile 429 212 68 68 16% 32% 100% - South Kordofan 1,001 473 68 68 7% 14% 100% - Red Sea 423 323 68 65 15% 20% 96% - North Darfur: non-IDP 835 380 43 33 4% 9% 77% - Total 2,688 1388 247 234 - - - 14,955 Source: SMoEs 2010; FMoGE (2009); authors' calculations. To examine the breakdown of the school sample relative to the school population according to key characteristics (rural/urban, boy/girls/co-ed, and regular/village/nomadic) two sources are used: the 2009 Education Yearbook and the sampling frame information collected from the four State Ministries of Education. Since the education yearbooks do not contain information on whether basic schools are rural or urban the rural/urban classification of schools contained in the sampling frame was used instead. Among the sample schools, 54 percent are rural and 46 percent urban, compared to 68 percent and 32 percent respectively of the basic schools in the sampling frame. By design, the survey over-sampled urban schools and therefore the sample was weighted to obtain estimates that represent the underlying population. To compare the breakdown of basic schools according to boys/girls/co-ed school and regular/village/nomadic school, data from the Education Yearbook are used: 26 percent of the sample schools are for boys only, 28 percent for girls only, and 46 percent are co-ed. Among all government basic schools in the four survey states the respective breakdown is: 21 percent, 20 percent and 59 percent. 91 In all four states, the share of girls enrolled in basic education is smaller than that of boys and the share of children enrolled in nomadic schools is larger than the national average (two percent) in three of the states: Blue Nile (6 percent), North Darfur (5 percent) and Red Sea (5 percent). Main characteristics of the school sample and population % of sample % of all schools3 schools1,3 Government basic schools: Rural 54 68 Urban 46 32 Government basic schools: Regular 89 83 Villlage 3 4 Nomadic 7 13 Government basic schools, regular: 2 Boys 27 21 Girls 28 20 Co-ed 46 59 Notes: 1. The Education Yearbooks do not contain a rural/urban breakdown for basic schools. Instead the % of schools in the samplilng frame that are rural and urban respectively are shown. This information is not available for North Darfur. 2. The Education Yearbooks only contain a boy/girl/co-ed breakdown for regular schools. 3. Totals may not sum to 100 percent due to rounding. Sources: SDS2 (2010); SMoEs (2010); Education Yearbook (2009); authors' calculations. This compares to an overall distribution in the four survey states of: regular schools (81 percent), nomadic schools (13 percent) and village schools (4 percent). Thus, nomadic schools are under- represented in the sample because some nomadic sample schools were unreachable during the field work due to bad road conditions or closure. 92 Annex 2.1: Description of Quality Determinant Variables Variable Definition Math Score Grade 5 student score in the math learning assessment (% correct) (Min=0; Max=100) Days of School School closure during the academic year due to heat or rain. Closure (Number of days) Drinking Water Indicator variable. (=1 if school has drinking water, otherwise =0) High Student- Indicator variable. Textbook Ratio (=1 if student-textbook ratio is greater than 3:1, otherwise=0) Teacher Experience Years of teaching experience (school mean) (Number of years) Exercises Offered Indicator variable. in Class (=1 if student completed math exercises provided by teacher during the last lesson, otherwise=0) Age Age of student. (Years) Girl Indicator variable. (=1 if student is a girl, otherwise=0) Household Assets Weighted asset index based on whether a household owns a car, television and/or (Rural) mobile phone Household Assets Weighted asset index based on whether a household owns a car, television, (Urban) refrigerator and/or computer IDP Students Indicator variable. (=1 if student is an IDP, otherwise=0) 93 Annex 2.2: The Model Student Learning Function The model underlying the estimation considers student learning outcome as a function of current school inputs and student and household characteristics. (1) Tij=Ti (Cij, Hij, Si) + uij, where: Tij is the test score of student i in household j, Cij is a vector of student characteristics, Hij is a vector of household characteristics and Sij is a vector of school inputs. The error term uij in the model includes unobservable inherent student ability, any measurement error, and past inputs. Linearizing equation 1 provides the specification that is estimated by the ordinary least squares approach taking into account the survey design: (2) Tij=β0 + β1Cij + β2Hij + β3Si + uij 94 Annex 4.1: Selected International Experience with School-Based Management Strength Intervention Results Education with Strong Community elected 5-member council Increase in enrollments; higher Community (ACE) receives funds from Ministry probability of staying in Participation of Education to enact ministry and school; higher teaching time, (EDUCO) community policies and for teacher more parent-teacher meetings; El Salvador hiring, firing and monitoring. Further, PTAs visited classrooms three 1991 in exchange for public funding of to four times more than in schools (for tuition, textbooks, traditional schools. uniforms, registration and basic supplies), parents contributed meals and labor to improve school facilities. Autonomous Strong Councils hire/fire principals, maintain Where the policy was fully School Program facilities and ensure academic quality. implemented, student results in (ASP) math and language improved; Nicaragua more autonomy over teacher 1991 issues, rather than pedagogical issues, had a substantial positive effect on student achievements. Nepal Community Strong School management committee of Improvement in access and Support Project parents and others hire/fire community equity in the first two years. Nepal teachers and could refuse/return 2007 government posted teachers. The committees were also given unconditional black grants to allocate and finance schools' discretional funding. Decentralization Medium Three major reforms: (i) schools were Education performance tended Brazil granted financial autonomy; (ii) to be better under all three 1982 principals were either elected by interventions. The financial school officials, parents and students, autonomy option, however, had or appointed by local governments on a substantially larger impact the basis of a competitive exam, or a than the other two options. combination of both; and (iii) councils coordinate and evaluate school activities (pedagogical, administrative and financial). Extra Teachers Medium Community PTAs suggest teacher Contract teachers’ impact on Program promotions and transfers from scores was higher in schools Kenya capitation grants and help design and where the PTA received 2006 implement school plans. The program monitoring training and gave schools financing to hire one remained involved. Students additional contract Grade 1 teacher with government teachers in and provided the PTAs with training schools with strong community on how to assess teachers’ involvement in teacher performance. monitoring performed better. 95 Strength Intervention Results Quality Schools Medium In exchange for five-year grants, Decrease in dropout rate; Program (PEC) schools must create school plans with arguable impact on failure and Mexico parent associations involved in the repetition rates; modest 2001 design, implementation and improvement in students’ monitoring of the plans. results. Whole School Medium Principals, teachers and community Although there was a Development members received training in a variety meaningful reduction in (WSD) of aspects of school management and student and teacher The Gambia developed school management plans absenteeism, three to four years 2007 during the training. Implementation of into the program there was no the plans was subsequently funded significant positive impact on with US$ 500 school grants. student achievements. The WSD’s impact was more often positive in communities with high adult literacy rates. Support to School Low Provided modest funding to parent Significant effect on lowering Management associations. Funding was subject to repetition and failure rates, but Program (AGE) several conditions, including audit and not on dropout rates; increase Mexico not paying teacher salaries, and in student performance. 1996 mainly allocated to school infrastructure. Parent associations, in turn, attended training and committed to additional involvement in school activities. Education Quality Low Cluster committees representing six Slightly lower dropout rates; Improvement schools on average determine the lower repetition rates and Project priorities for where schools will invest higher achievements. Cambodia cash grants. 1998 School Low School committees of varied Electing school committee Operational representation oversee the spending of members and linking school Assistance block grants, allocated to schools committee members with the Program (BOS) based on the number of students. village council had the Indonesia Grant may be used for only non-salary strongest combined positive 2005 operational expenditure, not teacher impact on learning outcomes. salaries or capital investment. Committee training also had a positive impact, to a substantially lesser extent. Source: Elaborated by the authors, on the basis of recent research. 96 Annex 5.1: Basic Education Indicators for North Darfur State North Darfur Sudan Enrollment (Number of Students) 346,779 4,870,464 % of Girls 46 46 % of Students in Nomadic Schools 5 3 % of Students in IDP Schools 18 1 Gross Enrollment Ratio (%) 67 72 Gross Intake Rate (Grade 1 - %) 80 80 Primary Completion Rate (Grade 8 - %) 44 54 Retention Rate (%) 56 68 Public per Student Spending (SDG) 171 240 Source: World Bank (2011b). Note: The gross intake rate is obtained as the ratio of non-repeaters in Grade 1 to the population of children aged six years. The gross enrollment ratio is calculated as total enrollment for a given level of education divided by the population of the official school age for that level (6 to 13 years for basic education in Sudan). The primary completion rate is calculated as the non- repeaters in Grade 8 divided by the population aged 13 years. 97 Annex 5.2: Determinants of Grade 5 Student Performance in Math Whereas student learning outcomes in the four states are very weak overall, rural students and IDP students perform worse on average in the math assessment than their urban counterparts. School-side factors leading to significantly worse student performance include greater school closures, poor infrastructure, less experienced teachers and a greater lack of textbooks. Moreover, in rural schools richer students, girls and students whose teachers set exercises to in class perform significantly better. Together, these findings underline the importance of school inputs and teachers for student learning in addition to individual student and household characteristics (Further detail is provided in Policy Note No.2 on learning outcomes). Determinants of Grade 5 Student Performance in Math, by Location, 2010/11 Dependent Variable: Urban Rural Math Test Score (%) Mean S.E. Mean S.E. School Characteristics Days of School Closure -0.607*** (0.009) -0.119*** (0.009) Drinking Water 1.733*** (0.309) 5.303*** (0.306) Pupil-Textbook Ratio > 3:1 -7.852*** (0.781) -6.161*** (0.290) Teacher Experience 0.883*** (0.179) 1.112*** (0.069) Teacher Experience Squared -0.050*** (0.004) -0.059*** (0.003) Exercises Offered in Class 1.011*** (0.583) 1.159*** (0.494) Student and Household Characteristics Age of Student -0.092*** (0.194) 0.039*** (0.126) Girl -0.510*** (0.686) 1.068*** (0.433) Household Assets (Rural) - - 0.337*** (0.100) Household Assets (Urban) 0.044*** (0.078) - - IDP Students -20.920*** (0.541) -5.338*** (0.449) Constant 49.34*** (2.876) 23.800*** (1.620) R2 0.302 0.233 No. Of Observations 1,658 1,717 Source: SDS2 data and authors’ computations. Note: Teacher and school resources are school averages. Standard errors in parenthesis: *** Statistically significant at the 1% level; ** Statistically significant at the 5% level; * Statistically significant at the 10% level. . 98 Bibliography Adam, E. O. 2007. Interagency Livelihoods Assessment in Abu Shouk, Al Salam and Zamzam IDP Camps, North Darfur - El Fasher. Available online at: http://www.unsudanig.org/docs/El%20Fasher%20IDP%20Camp%20Livelihood%20Assess%20 Report_Nov%202007.pdf Bardhan, P. 2002. “Decentralization of Governance and Development.” Journal of Economic Perspectives, No. 16 (4), pp. 185-205. Bardhan, P. and D. Mookherjee. 2000. “Capture and Governance at Local and National Levels.” American Economic Review, No. 90 (2), pp. 135-139. Bardhan, P. and D. Mookherjee. 2005. “Decentralizing Antipoverty Program Delivery in Developing Countries.” Journal of Public Economics, No. 89 (4), pp. 675–704. Barrera-Osorio, F., T. Fasih and H. A. Patrinos. 2009. Decentralized Decision Making in Schools: Theory and Evidence on School Based-Management. The World Bank. Washington, DC. Barrera-Osorio, F., V. Garcia-Moreno, H. Anthony Patrinos and E. Porta. 2010. Using the Oaxaca-Blinder Decomposition Technique to Analyze Learning Outcome Changes Over Time. World Bank Policy Research Working Paper No. 5584. Washington, DC. World Bank. Bruns, B., D. Filmer and H. A. Patrinos. 2011. Making Schools Work: New Evidence on Accountability Reforms. The World Bank. Washington, DC. Cascio, E. and D. Whitmore Schanzenbach. 2007. First in the Class? Age and the Education Production Function. NBER Working Paper No. 13663. Cambridge, MA. The National Bureau of Economic Research. Chinapah, V. 2003. Monitoring Learning Achievement (MLA) Project in Africa. Presented at the Association for the Development of Education in Africa Biennial Meeting, Mauritius, December 3-6, 2003. Cook, T. D. 2007. School-Based Management in the United States. Background paper prepared for the programmatic study on school-based management. The World Bank. Washington, DC. Cueto, S., C. Ramirez and J. Leon. 2006. “Opportunities to Learn and Achievement in Mathematics in a Sample of Sixth Grade Students in Lima, Peru.” Educational Studies in Mathematics, No. 62(1), pp. 25-55. Currie, J. and D. Thomas. 2001. Early Test Scores, Socioeconomic Status and Future Outcomes. NBER Working Paper No. 6943. National Bureau of Economic Research, Cambridge, MA. Darling-Hammond, L. 1997. Doing What Matters Most: Investing in Quality Teaching. Report prepared for the National Commission on Teaching and America’s Future. National Commission on Teaching & America’s Future, Kutztown, PA. 99 Fehrler, S., K. Michaelowa and A. Wechtler. 2009. “The Effectiveness of Inputs in Primary Education: Insights from Recent Student Surveys for Sub-Saharan Africa.” Journal of Development Studies, No. 45(9), pp. 1545-1578. Fuller, B. 1986. Raising School Quality in Developing Countries: What Investments Boost Learning? The World Bank, Washington, DC. Glewwe, P. and H. Jacoby. 1994. “Student Achievement and Schooling Choice in Low-Income Countries: Evidence from Ghana.” The Journal of Human Resources, No. 29(3), pp. 843-864. Glewwe, P., M. Kremer and S. Moulin. 2009. “Many Children Left Behind? Textbooks and Test Scores in Kenya.” The American Economic Journal: Applied Economics, No. 1(1), pp. 112-35. Goldhaber, D. and D. J. Brewer. 1997. “Why Don't Schools and Teachers Seem to Matter? Assessing the Impact of Unobservables on Educational Productivity.” The Journal of Human Resources, No. 32(3), pp. 505-523. Greenwald, R., L. V. Hedges and R. D. Laine. 1996. “The Effect of School Resources on Student Achievement.” Review of Educational Research, No. 66(3), pp. 361-396. Hanushek, E. A. 1995. “Interpreting Recent Research on Schooling in Developing Countries.” The World Bank Research Observer, No. 10(2), pp. 227-246. Hanushek, E. A. 2002. “Teacher Quality.” In Teacher Quality, Eds. Lance T. Izumi and Williamson M. Evers. Hoover Institution Press, San Francisco. Hanushek, E. A. 2003. “The Failure of Input-Based Schooling Policies.” The Economic Journal, No. 113(485), pp. F64-F98. Hanushek, E. A. and L. Wöβmann. 2009. “Do Better Schools Lead to More Growth? Cognitive Skills, Economic Outcomes, and Causation.” CESifo Working Paper Series, No. 2524. CESifo Group, Munich. Harbison, R. W. and E. A. Hanushek. 1992. Educational Performance of the Poor: Lessons from Rural Northeast Brazil. Oxford University Press, Oxford, UK. Hardman, F., J. Ackers, N. Abrishamian and M. O’Sullivan. 2011. “Developing a Systemic Approach to Teacher Education in Sub-Saharan Africa: Emerging Lessons From Kenya, Tanzania, and Uganda.” Compare: A Journal of Comparative and International Education, No. 41(5), pp. 669-683. Harris, D. N. and T. R. Sass. 2008. Teacher Training, Teacher Quality, and Student Achievement. CALDER Working Paper No. 3. National Center for Analysis of Longitudinal Data in Education Research, Washington, DC. Heyneman, S. P., D. T. Jamison and X. Montenegro. 1984. “Textbooks in the Philippines: Evaluation of the Pedagogical Impact of a Nationwide Investment.” Educational Evaluation and Policy Analysis, No. 6(2), pp. 139-150. Heyneman, S. P., J. P. Farrell and M. A. Sepulveda-Stuardo. 1978. Textbooks and Achievement: What We Know. World Bank Staff Working Paper No. 298. The World Bank, Washington, DC. Hoxby, C. 2001. “If Families Matter Most, Where Do Schools Come In?” In A Primer on America's Schools, Ed. Ted M. Moe. Hoover Institution Press, San Francisco. 100 IAEEA (International Association for the Evaluation of Educational Attainment). 2008. Report from Test Administration in Benin. Hamburg. IDMC (Internal Displacement Monitoring Committee). 2012. Jamison, D. T., B. Searle, K. Galdas and S. P. Heyneman. 1981. “Improving Elementary Mathematics Education in Nicaragua: An Experimental Study of the Impact of Textbooks and Radio on Achievement.” Journal of Educational Psychology, No. 73(4), pp. 556–567. Jaspars, S., M. H. Zakaria, Y. T. El Nour and A. J. Fuddle. 2010. “Coping and Change in Protracted Conflict : The Role of Community Threats to Livelihoods. A Case Study Based on the Experience of Practical Action in North Darfur.” Development (May). http://www.odi.org.uk/resources/docs/5964.pdf. Lloyd, C. B., S. El Kogali, J. Perlman Robinson, J. Rankin and A. Rashed. 2010. Schooling and Conflict in Darfur : A Snapshot of Basic Education Services for Displaced Children. New York. Lockheed, M. and E. Hanushek. 1988. “Improving Educational Efficiency in Developing Countries: What Do We Know?” Compare: A Journal of Comparative and International Education, No. 18(1), pp. 21–38. Majgaard, K. and A. Mingat. Forthcoming. Education in Sub-Saharan Africa: A Comparative Analysis. World Bank and Education for All Fast Track Initiative. Washington, DC. Michaelowa, K. and A. Wechtler. 2006. The Cost-Effectiveness of Inputs in Primary Education. Insights from the Literature and Recent Student Surveys for Sub-Saharan Africa. Presented at the Association for the Development of Education in Africa Biennial Meeting, Gabon, March 27-31, 2006. Available at: http://www.adeanet.org/adeaPortal/adea/biennial2003/papers/2Ac_MLA_ENG_final.pdf Michaelowa, Katharina (2001). “Primary Education Quality in Francophone Sub-Saharan Africa: Determinants of Learning Achievement and Efficiency Considerations.” World Development 29(10): 1699-1716. MoGE (Ministry of General Education). 2008. Baseline Survey on Basic Education in the Northern States of Sudan. Final Report. Khartoum. MoGE (Ministry of General Education). 2008. Baseline Survey on Basic Education in the Northern States of Sudan. Final Report. Directorate General of Educational Planning. Khartoum. MoHA (Ministry of Humanitarian Affairs). 2009. National Policy on Internally Displaced Persons. January 2009. Mullis, I. V. S., M. O. Martin, E. G. Fierros, A. L. Goldberg and S. E. Stemler. 2000. Gender Differences in Achievement. IEA’s Third International Mathematics and Science Study (TIMSS). International Association for the Evaluation of Educational Achievement, Boston. OECD (Organisation for Economic Cooperation and Development). 2005. Teachers Matter: Attracting, Developing and Retaining Effective Teachers. Education and Training Policy series. Postlethwaite, T. N. and K. N. Ross. 1992. Effective Schools in Reading: Implications for Educational Planners - An Exploratory Study. The IEA Study of Reading Literacy II. International Association for the Evaluation of Educational Achievement, Hamburg. 101 Psacharopoulos, G. and H. A. Patrinos. 2004. “Returns to Investment in Education: A Further Update.” Education Economics, No. 12(2), pp. 111-134. Rivkin, S. G., E. A. Hanushek and J. F. Kain. 2005. “Teachers, Schools, and Academic Achievement.” Econometrica, No. 73(2), pp. 417-458. Rockoff, J. E. 2004. “The Impact of Individual Teachers on Student Achievement: Evidence from Panel Data.” The American Economic Review, No. 94(2), pp. 247-252. Sirin, S. R. 2005. “Socioeconomic Status and Academic Achievement: A Meta-Analytic Review of Research.” Review of Educational Research, No. 75(3), pp. 417-453. Taylor, N., T. Mabogoane, J. Schindler and B. Akoobhai. 2010. Seeds of Their Struggle: The Features of Under- and Overage Enrollment Among Grade 4 Learners in South Africa. CREATE Research Monograph No. 47. The Consortium for Educational Access, Transitions and Equity, Brighton, UK. Todd, P. E. and K. I. Wolpin. 2003. “On the Specification of the Production Function for Cognitive Achievement.” The Economic Journal, No. 113(485), pp. F3-F33. UNESCO (United Nations Educational, Scientific and Cultural Organisation). 2007. Education for All by 2015. Will We Make it? EFA Global Monitoring Report 2008. Oxford University Press. Oxford, UK. UNESCO (United Nations Educational, Scientific and Cultural Organisation). 2007. Education Counts: Benchmarking Progress in 19 WEI Countries: World Education Indicators. Paris. UNICEF. 2011. Unpublished briefing note on the Nomadic Children Education Project. Vegas, E. and I. Umansky. 2005. “Improving Teaching and Learning through Effective Incentives Lessons from Education Reforms in Latin America.” In Incentives to Improve Teaching. Lessons from Latin America. Ed. Vegas, Emiliana. The World Bank, Washington, DC. Watkins, K. 2000. The Oxfam Education Report. Oxfam, London. World Bank. 2010. “Concept Note. Study on Service Delivery in Basic Education in Four States in Northern Sudan: Blue Nile, North Darfur, Red Sea and South Kordofan.” Washington, DC. World Bank. 2011. Education in Ghana: Improving Equity, Efficiency and Accountability of Education Service Delivery. Report No. 59755-GH, AFTED. The World Bank, Washington, DC. World Bank. 2011. Northern Sudan Education Sector Status Report 2010. Washington, DC. World Bank. 2011a. The World Development Report 2011: Conflict, Security and Development. World Bank and Oxford University Press. Washington, DC. World Bank. 2012. The Status of the Education Sector in Sudan. Africa Human Development Series. Washington, DC. World Bank. 2012a. Out-of-School Children and Student Dropout in Basic Education in Sudan. Draft. The World Bank, Washington, DC. 102