REDUCING EARLY GRADE DROP OUT AND LOW LEARNING ACHIEVEMENT IN LAO PDR ROOT-CAUSES AND POSSIBLE INTERVENTIONS May 30, 2016 STANDARD DISCLAIMER COPYRIGHT STATEMENT This volume is a product of the staff of the International The material in this publication is copyrighted. Copying Bank for Reconstruction and Development/ The World and/or transmitting portions or all of this work without Bank. The findings, interpretations, and conclusions permission may be a violation of applicable law. The expressed in this paper do not necessarily reflect the International Bank for Reconstruction and Development/ views of the Executive Directors of The World Bank or The World Bank encourages dissemination of its work the governments they represent. The World Bank does and will normally grant permission to reproduce portions not guarantee the accuracy of the data included in this of the work promptly. work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply For permission to photocopy or reprint any part of this any judgment on the part of The World Bank concerning work, please send a request with complete information the legal status of any territory or the endorsement or to the Copyright Clearance Center, Inc., 222 Rosewood acceptance of such boundaries. Drive, Danvers, MA 01923, USA, telephone 978-750- 8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax : 202-522-2422, e-mail : pubrights@worldbank.org. | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR the Education Sector Development Plan Coordination Unit (ECU) Director, MoES, Mr. Somkhanh Didaravong, ACKNOWLEDGEMENTS Education Statistics & Information Technology Center, MoES, and Mr. Sisouva Vimon, Director General of the Department of Finance, MoES, for facilitating access This study was conducted under the overall guidance of to data and other resources for the study. We would Mr. Harry Patrinos (Practice Manager, Education Global also like to thank the Japan International Cooperation Practice, East Asia and Pacific Region, the World Bank), Agency (JICA) for sharing the findings and final report and Mr. Plamen Danchev (Senior Education Specialist, of their qualitative research into early grade repeti- Education Global Practice, East Asia and Pacific Region, tion and drop outs, and the Lao Statistics Bureau the World Bank). The authors of the study included (LSB) for providing information from the Lao Expendi- Mr. Pedro Cerdan-Infantes, Mr. Jeffery H. Marshall ture and Consumption Surveys (LECS) and the Lao and Ms. Emiko Naka. Labor Force and Child Labor Survey (LFS). The team would like to thank Dubai Cares and Plan Inter- The team is grateful to the following peer reviewers national for their support and advice throughout the for their invaluable inputs to help enhance the quality course of this research study. In particular, we would like of the study: Ms. Melissa Ann Adelman, Economist, to thank Ms. Jodie Fonseca, (LEARN Project Director, Global Practice Education, The World Bank; Ms. Raja Plan International) for leading effective collaboration Bentaouet Kattan, Program Leader, The World Bank; between the research team and key staff at Plan Interna- Ms. Colleen Loomis, Associate Professor, Wilfrid tional and Dubai Cares. We would also like to thank Dr. Laurier University; Mr. Abdeljalil Akkari, Professor, Sally Brinkman, Senior Research Advisor on Early Child- Education studies, University of Geneva; and Mr. Roy hood Development, Ms. Natasha Graham, Advisor on Huijsmans, Senior Lecturer, Erasmus University.We Disability, and Ms. Myrna Machuca-Sierra, Education would also like to thank Ms. Boualamphan Phouthavi Specialist, for their contributions to the workshops with souk for her excellent administrative support. Plan International and Dubai Cares and their inputs into the study. The study benefitted from the support of various officials and staff from the Ministry of Educa- tion and Sports (MoES). We would like to extend our appreciation to Mrs. Khampaseuth Kitignavong, Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | LIST OF ACRONYMS ASLO Assessment of Student Learning Outcome LECS Lao Expenditure and Consumption Survey CCT Conditional Cash Transfers LFS Labor Force Survey CNREG Censored Normal Regression LSS Lower Secondary School DESB District Education & Sports Service Bureau MOES Ministry of Education and Sports ECE Early Childhood Education NER Net Enrolment Rate EGRA Early Grade Reading Assessment OLS Ordinary Least Squares EMIS Education Management Information ORF Oral Reading Fluency System PA Pedagogical Advisor EQS Education Quality Standards PESS Provincial Education and Sports Services ESDP Education Sector Development Program PPS Probability Proportional to Size ESWG Education Sector Working Group RIES Research Institute for Education Sciences FE Fixed Effects SBG School Block Grant GER Gross Enrolment Rate SDP School Development Plan HLM Hierarchical Linear Model SES Socio-Economic Status JICA Japan International Cooperation Agency UIS UNESCO Institute of Statistics LEARN Lao Education Access Research and USS Upper Secondary School Networking Project VEDC Village Education Development Council | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR CONTENTS Acknowledgements 3 B. Learning Outcomes 26 List of Acronyms 4 C. Conclusions and policy areas Causes for drop-out and low performance 34 Summary 6 Policy areas 35 Introduction 7 References 41 A. Enrollment and Drop-out 9 Annex A – Root Causes Research 1. How many primary school-age Database Summary 42 children are not enrolled in school? 9 Annex B – Summary of Workshops on 2. Who and where are out the Causes for Drop-out 43 of school children? 11 3. When do children enroll (if ever), and when do they abandon school? 14 4. Underlying reasons for dropout and never attending school 18 Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | But the main reasons for never attending or for dropping out early are mostly related to perceptions about the quality of education is low, or that it lacks relevance. When looking at the determinants of learning, the same factors that drive drop-out are also associated with low early grade reading ability and low performance in standardized tests (ASLO). This suggests that low progress in learning may be resulting in disinterest and SUMMARY eventual drop-out. Unfortunately, it is not possible to test this empirically using existing data. This warrants further research into the underlying causes of Despite the impressive progress of Lao PDR in non-participation. underlying causes of non-participation. expanding the access to education at all levels, the country faces persistent problems related to the significant The results point to three general categories of policy number of children remaining out of school or leaving responses: (i) focus on learning outcomes, especially primary school early. More importantly, many children early grade reading ability; (ii) increase the demand for who remain in schooling show very low levels of reading education for those who perceive education to have a ability and learning. Both of these groups (out of school low value; and (iii) continue to increase and improve the children and low performers) have similar backgrounds: supply of schooling. Convincing families to enroll their they tend to be poorer, live in rural and (especially) children in school on time, and to keep them enrolled, remote areas, and come from non-Lao Tai ethnic groups. may require more than just an increase in complete and Early leavers are more likely to be girls, who start leaving better-equipped schools, or even scholarships. The schooling earlier than boys, at age 10-11. Physical results from this review also suggest that innovative access to school is still a problem for some groups: interventions are needed to make education more between 30 and40 percent of those who leave schooling relevant to rural, ethnic populations, both in the flexibility early live in a village where schools do not offer G4 and G5. of delivery and the content of what is taught. 6 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR INTRODUCTION 1. Participation in basic education in Lao PDR has Also, the increase in the net enrollment rate, combined improved steadily in recent decades. The net enrollment with the recent decline in the gross enrollment rate rate for primary schooling has climbed from 65 percent (GER), shows that progress is being made in terms of in 1990 to 98 percent in 2014 (Figure 1). This increase efficiency at the primary level. For lower secondary, the is a direct result of government policies, with partner results also show steady increases in both gross and net support, to build new schools, hire new teachers and enrollment rates over the 2000-2014 time period, improve access to all levels of schooling. It is the result although much work remains to be done to reach full of prioritizing support for specific populations (such as participation and improved efficiency. girls and children) who live in rural and remote areas. FIGURE PRIMARY AND LOWER SECONDARY 1 GROSS AND NET ENROLLMENT 140 120 122.9 121.7 120.7 93.8 120.1 95.4 118.9 87.1 118.2 95.1 116.3 PERCENTAGE (%) 100 84.7 115.5 82.2 113.9 111.7 106.7 94.2 94 98.5 89.5 80 80.6 71.3 75.7 60 64.5 64.9 52.6 59.7 56.1 31.7 56.2 53.8 40 53.6 47 44.9 51.5 51.7 42.6 2010 50.6 39 37.3 20 2010 2011 0 2012 1990 2000 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Source: UNESCO (UIS), various years 2. Lao PDR has met most MDGs in education, particularly 3. There has been clear progress in gender equality, especially MDG 2 (Achieve Universal Primary Education) and in pre-primary and primary education. However, this MDG 3 (Gender Equality). However, survival to grade 5 is becomes more off track as the level of education still low. Although the primary school participation rate increases. The Gender Parity Index (GPI) for lower and in Lao PDR has steadily increased over recent decades, upper secondary education was 0.94 and 0.87 the survival rate to Grade 5 remains off target, largely respectively in 20152. due to high early grade dropout. The survival rate to Grade 5 was 78 percent in 2015, compared to the target of 95 percent. The findings of the Joint Education Sector 1 This is an analysis of the sector carried out by MoES and Review Mission (JSRM 2014)1 also indicate that key Development Partners periodically. challenges to achieving MDG 2 were the high dropout 2 Education Sector Development Plan 2016-2020, Ministry of and repetition rates for grades 1 and 2. Education and Sport, 2015 Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 7 4. There are worrying signs about low quality in the basic 5. The EGRA was carried out in 2012 in a sample of education system, which may in turn be related to drop- schools that did not include remote, small or ethnic out and low attendance rates. Official figures show high schools (defined as those with more than 50 percent of dropout rates in grade 1 and, to a lesser degree, grades the population being non Lao-Tai). That the assessment 2-4. Importantly, these rates have declined in recent showed that around 1/3 of third graders are not able to years, but in grade 1 they remain above 10 percent. In read one word signals that education quality may be an addition, the results of the Early Grade Reading Assessment, important driver of drop-out. Considering the sample, (EGRA, 2012), show very high proportions of grade 2 and this is a very worrying result. Had the sample included 3 children who were unable to read even one word on ethnic and remote schools, the results would have been the Fluency and Comprehension sections of the EGRA significantly worse. This signals a clear urgency in test (see right half of figure 2). addressing reading ability deficiencies in early grades. Without the ability to read, students cannot absorb the curriculum in later grades. FIGURE THE PROBLEM: EARLY GRADE DROP-OUT RATES 2 AND EARLY GRADE READING EARLY GRADE DROP-OUT RATES 13 PERCENT (0-100%) 12.1 11.7 9.75 11 6.5 6.6 6.3 6 5.9 5.7 5.3 5.4.7 3.25 2010 5 4.5 2011 0 2012 Grade 1 Grade 2 Grade 3 Grade 4 % OF STUDENTS NOT ABLE TO READ ONE WORD 0.6 % SCORING 0 IN READING 0.5 54% 0.4 0.3 30% 27% 0.2 0.1 Grade 2 12% 3% Grade 3 7% 0 Grade 4 Speaks No Lao at Home Speaks Lao at Home Source: EMIS 2010, 2011, 2012, EGRA 2012 8 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR A. ENROLLMENT AND DROP-OUT 1. HOW MANY PRIMARY SCHOOL-AGE CHILDREN ARE NOT ENROLLED IN SCHOOL? 9. The number of out of school children of basic education age is significant - between 30,000 and 85,000, depending on the data source. Official government education statistics (EMIS) show that about 30,000 primary school age children are not in school, and this figure has declined substantially in recent years. However, as shown 6. In this context, Plan International and Save the in Figure 3, household survey data sources—which Children, though their Dubai Cares-funded Lao Educational include the LFS from 2010-11 and the LECS from 2012- Access, Research and Networking (LEARN) project, 13-report higher numbers of out of school children partnered with the World Bank to undertake an analysis about 85,000).4 of the patterns of enrollment and dropout in Lao PDR, as well as the root causes of the observed patterns in FIGURE enrollment and low levels of learning. The report also NUMBER OF OUT OF discusses policy options to address these causes.3 7. The report is divided into two parts. In the first section 3 SCHOOL CHILDREN BY DATA SOURCE (A. Enrollment and drop-out), it presents a complete picture of attendance and dropout, addresing four main NUMBER questions: OF CHILDREN (i) How many school-aged children are not enrolled in school? 100,000 (ii) Who and where are they? 90,000 (iii) When do they enroll and drop-out? 80,000 88,894 (iv) What are the reasons behind observed 85,862 enrollment patterns? 70,000 60,000 8. In the second part (B. Learning Outcomes), the report presents the analysis of the factors associated with 50,000 learning outcomes. The report concludes with 40,000 42,024 recommendations on a set of interventions to address 30,000 the most frequently cited causes of dropout and 20,000 29,694 non-participation that come out of the analysis. 10,000 3 The main result of this analysis is this report. The team also produced a much more extensive, mainly descriptive report, which serves as 0 background analysis for this report. The report is available from authors upon request. EMIS LFS EMIS LECS* 4 2010-2011 2012-2013 See Annex A for a description of data sources used in this analysis. Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 9 10. Discrepancies between household survey and admin- data analysis in this report we rely on household survey istrative data are not uncommon, and can result from data sources for summaries of in school and out of definitional differences, problems with sampling frameworks, school children. These discrepancies also point to the or inaccurate data reporting. One possibility is that need for better quality control in the collection of village education committees are enrolling children in enrollment data in the EMIS. grade 1 in order to meet targets for universal coverage, but these children are not actually attending school. 11. When looking at the last grade completed(as This would also predict high dropout and repetition rates reported in household surveys), permanent drop-out in in grade 1 on the basis of administrative data, since first or second grade is much lower than in later grades. many of these grade 1 children will not continue on to Significant numbers of children are dropping out of grade 2. These discrepancies between data sources do primary school before completing grade 5. About 40 not fundamentally alter the main results in this report. percent of those children live in villages where the school However, it is possible that official data sources are over does not offer a complete range of primary grades. How- stating early grade enrollment, so for the bulk of the ever, early grade dropout—according to the LFS 2010- analysis in this report we rely on household survey 11-is not taking place in grade 1, but rather in grade 2-4 (see figure 3). FIGURE TOTAL NUMBER OF SCHOOL DROPOUTS (BY GRADE) 4 AND EXCLUDED CHILDREN AGED 11-17 70,000 NUMBER OF CHILDREN 68,633 52,500 18,880 46,003 35,000 17,474 10,905 4,673 17,500 0 Dropout G1 Dropout G2 Dropout G3 Dropout G4 Dropout G5 Never Attended Source: LECS V, 2011/2012 12. In addition, a large number of out of school children 13. A significant number of children of primary school have completed the primary sequence (up to grade 5), age never enter school. This in turn touches on two but are no longer in school. While this report focuses elements. The first is late entry, since a 7 or 8 year old on primary schooling, this finding highlights the importance child who is not in school at the time of the survey may of focusing on the transition to lower secondary still enter school eventually. But the large number for education, and of continuing to increase lower secondary Never Attended also highlights the issue of exclusion, school availability, as well as the need to design policies which comes from both supply-side constraints (i.e. no to facilitate the transition between levels of education. primary school available) and household demand factors (i.e. the family sees no need to enroll the child). Understanding these underlying causes is one of the primary focuses of this study. The next section looks in more detail at the characteristics of those children who never enter school. 10 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR 2. WHO AND WHERE ARE OUT OF SCHOOL CHILDREN? 14. The typical out of school child lives in a rural area The populations with high rates of never enrolling are without access to a road, is poor and is more likely to be the same as those with early grade drop-out. Out of non-Lao Tai. Gender differences in enrollment start to school children are concentrated in rural areas, and be significant after age 14, with early drop-out being among the poorest households (figure 5). more common for girls than for boys. FIGURE PERCENTAGE OF CHILDREN WHO HAVE NEVER 5 ATTENDED SCHOOL, BY VILLAGE TYPE AND WEALTH QUINTILE BY VILLAGE TYPE 30 PERCENT (0-100%) 22.5 25.4 15 15.5 4.8 11.5 7.5 4.7 Age 6-9 1.5 0 Age 10-14 Urban Rural with road Rural no road BY WEALTH QUINTILE 30 PERCENT (0-100%) 22.5 23.2 17.8 15 11.2 6.9 7.5 9.8 4.4 8.8 2.7 Age 6-9 0.9 2 0 Age 10-14 Q1 (Poorest) Q2 Q3 (Average) Q4 Q5 (Wealthiest) Family Wealth Quintile Source: LFS, 2010 Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 11 15. Ethnic minority groups have the highest probabilities of never attending school (Figure 6), especially the FIGURE Chinese-Tibet. In terms of raw numbers (Figure 7), RAW TOTALS OF PERSONS Mon-Khmer children make up the largest group among those ever enrolled, with relatively small numbers of 7 WHO HAVE NEVER ATTENDED SCHOOL, Chinese-Tibet and Hmong-TuMien. BY ETHNICITY FIGURE PERCENTAGE OF CHILDREN 6 30,000 WHO HAVE NEVER TOTAL NUMBER 22,500 ATTENDED SCHOOL, BY AGE 24,653 12,888 GROUP AND ETHNICITY 15,000 7,901 7,487 3,989 5,662 13,318 2,893 7,500 PERCENTAGE (0-100%) 30 0 28.5 22.5 Lao-Thai Mon- Khmer Chinese- Tibet Hmong- Tu-Mien 22 15 17.6 17.3 7.3 7.5 4.8 Age 10-14 Age 15-20 9.1 1.9 0 Source: LECS 2012, 2013 Lao-Thai Mon- Khmer Chinese- Tibet Hmong- Tu-Mien 16. Children who have never enrolled in school are concentrated in a few provinces, namely Savannakhet, Champasack, Saravan and Oudomxay. Among 10-14 Age 6-9 Age 10-14 year olds, the numbers are considerably smaller, which does reflect the fact that most children are eventually Source: LECS 2012, 2013 entering school (figure 8). The gap between the two bars in Figure 8 is a reflection of late age entry, since many of the 6-9 year olds who have never attended will eventually do so. 17. Finally, there is evidence of a “triple condition”, where the interaction between being poor, rural and female results in high exclusion rates for certain groups. For example, among the rural poor, nearly 27 percent of non Lao-Tai girls aged 6-9 have never attended school, compared with only 10 percent of Lao-Tai males in rural areas who are not poor (figure 9). 12 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR FIGURE RAW TOTALS OF CHILDREN WHO HAVE NEVER 8 ATTENDED SCHOOL, BY AGE GROUP AND PROVINCE 14886 6612 16,000 14,000 12108 4534 12,000 10,000 4305 7391 8,000 2349 5353 4736 1707 6,000 4140 1768 3700 3420 2116 1160 3352 1722 4,000 2904 1495 3159 2162 2438 516 446 1715 1620 566 1110 1109 278 403 2,000 678 0 Savan nakhet Cham pasack Saravan Oudomxay Luang namtha Phonsaly Kham muane Bokeo Hua phanh Sekong Attapeu Luang prabang Xieng khuang Xaya bury Vientiane Borikha mxay Vientiane Age 6-9 Age 10-14 Source: LECS 2012, 2013 FIGURE PERCENTAGE OF RURAL CHILDREN WHO HAVE 9 NEVER ATTENDED SCHOOL, BY AGE GROUP, POVERTY, GENDER AND ETHNICITY PERCENTAGE (0-100%) 30 26.8 22.5 23.1 15 15.5 13.2 6.2 11.9 11.9 6.1 7.5 3.4 2.5 10.1 2.5 8.4 9.5 3 1 0 Lao-Tai Non Lao-Tai Non Lao-Tai Non Lao-Tai Non Lao-Tai Lao-Tai Lao-Tai Lao-Tai Female Male Female Male Rural Poor Rural Non-Poor Age 6-9 Age 10-14 Source: LECS 2012, 2013 Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 13 3. WHEN DO CHILDREN ENROLL (IF EVER), AND WHEN DO THEY ABANDON SCHOOL? 18. One of the main findings of this report is that statistics 19. The official age for starting school is 6, but on out of school children are related to age at entry. enrollment profiles by age again confirm the existence Most children will eventually enter school in Lao of late-entry. Fewer 6 and 7 year olds are in school PDR(although there are still a significant number of compared with eight and nine year olds, especially in excluded children who never set foot in a school). But at rural villages (see Figure 10). There is also evidence of any given point in time, a significant proportion of the exclusion, as attendance rates never reach 100 percent out of school population is made up of children who at any age, although as demonstrated in the previous have not yet entered school. section of this report, the percentage of children who never set foot in a school is clearly in decline. For gender, the results again show relative parity among the youngest cohorts, but after age 11 there is clear divergence, as more girls are leaving school (or never entering). FIGURE ENROLLMENT PROFILES 10 BY AGE AND GENDER 95.7 86.5 92.6 93.2 90.9 92.9 93 94 PERCENTAGE (0-100%) 100 89.8 89.1 85.4 85.7 82.3 87.5 85 70.4 74.9 68.3 78.5 66.3 71.6 62.4 66.5 70 55 Boy 40 Girls 6 7 8 9 10 11 12 13 14 15 16 Source: LECS, 2012, 2013 14 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR 20. A substantial percentage of young people in Lao PDR people from the wealthiest households (quintile 5) entered are beginning school at the ages of 7 and 8 and, in the primary school at age 5 or 6, compared with just 28 percent case of the poorest children, at age 9 or higher. Figure 11 of children from the poorest households (quintile 1). The summarizes age at school entry (left hand side), and shows consequences of late entry are shown clearly in Figure 11, the relationship between age at initial entry and grade as overall attainment (and duration) are higher among attainment and duration of attendance (right hand side). children who entered school at age 6 in comparison with The results show that age at initial entry varies little by those who entered school later. This is consistent with a gender, but it does vary considerably by socioeconomic “late in, early out” dynamic. profile of the household: nearly 80 percent of young FIGURE LATE ENROLLMENT 11 AND ATTAINMENT AGE OF SCHOOL ENTRY, BY GENDER AND SES QUINTILE GENDER Girls 3% 49% 29% 13% 6% Boys 2% 48% 30% 14% 7% 1% 28% 33% 22% 16% SES QUINTILE 1% 43% 32% 16% 8% 3% 52% 31% 11% 3% 4% 61% 26% 9% 0.5% 6% 72% 17% 3% 2% 0% 25% 50% 75% 100% 5 6 7 8 9+ ATTAINMENT AND SCHOOL DURATION BY AGE AT ENTRY Source: LECS, 2012, 2013 5.3 9 6 4.7 10 4.9 AGE AT ENTRY 4.4 11+ 3.1 5.3 9 6 4.7 10 4.9 Attainment 11+ 4.4 Duration 3.1 Years 0 1.5 3 4.5 6 Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 15 21. Differences by village type are most evident when One possibility is that as children complete primary comparing enrollment profiles: children in rural areas are school—and have limited options, or interest, in continuing much more likely to enroll late and drop out early (Figure to post-primary—they leave the school system. In a 12). In rural areas, the ‘inflection point’ for school separate analysis (not presented), it was found that attendance—where attendance rates begin to decline- among 12-16 year olds, about 97 percent of urban occurs at around age 10, whereas in the urban children who completed grade five continued on to complete sample it occurs a little later. This early inflection point grade 6, compared with 87 percent in rural areas (with in rural areas, combined with late entry, in effect road), and 80 percent in rural areas with no road. So in shortens the available time for many children to attend rural areas the attendance decline is explained in part school. For the purposes of understanding early grade by a termination of studies after completing primary dropout, it is important to understand the underlying school, although the cause for not continuing is not yet explanation(s) for school attendance beginning to fully clear. drop-off around age 10-11. FIGURE ENROLLMENT PROFILE 12 BY LOCATION 100 93.9 96.9 88.1 97.9 97.6 97 84.8 94.6 98 90 88.1 94 84.3 93.5 92 91.4 79 89.2 88.7 80 87.9 80.4 81.8 81.8 80 PERCENTAGE (%) 77.2 70 73.4 69.3 60 61.3 59.4 50 56.3 52.2 47.4 40 38.6 30 Urban Rural 20 Rural (no road) 6 7 8 9 10 11 12 13 14 15 16 Source: LECS 2012, 2013 16 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR 22. Almost 1/5 of poor non-Lao Tai girls living in rural areas never enter school, and a further 16 percent drop-out before completing primary. The corresponding rate for poor boys in rural areas is about 10 percent, while among Lao-Tai the rates are even lower. Another important result in Figure 13 is overage primary enrollment. Among poor non Lao-Tai minority groups, about 20 percent of children aged 14-17 are still enrolled in primary school. It is, of course, possible that these children will go on to finish primary school, but by the time they do so they may no longer have time to continue their studies at secondary level. If the family’s—or the child’s—goal was simply to complete primary school, then this issue of late entry may not be particularly consequential. But if these children hope to obtain higher levels of education, their late entry is going to be a serious constraint on future schooling for the simple reason that they are entering a period of their life when they can help their families more with work (“opportunity costs”), or even start their own families. FIGURE SCHOOLING HISTORY PROFILE FOR CHILDREN 13 AGED 14-17 BY SESAND ETHNIC GROUP 2% 2% 1% WEALTHY Girls 13% 83% 1% 3% 1% LAO-TAI Boys 15% 80% POOR NON Girls 19% 17% 20% 17% 28% LAO-TAI Boys 10% 13% 21% 18% 37% POOR Girls 6% 14% 4% 26% 50% LAO-TAI Boys 5% 9% 10% 31% 46% 0% 25% 50% 75% 100% Never entered Dropped Enrolled Dropped out Enrolled in Source: LECS 2012, 2013 school out before G5 in primary after primary post-primary Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 17 23. A summary of what this report has so far examined highlights a number of important findings related to school attendance and dropout in Lao PDR. These include: i. Never enrolling is more common than early grade drop-out, although it is possible that children are dropping out of grade one repeatedly, which is a version of grade repetition; ii. Late enrollment in primary school is common, especially among the poorest households, and among non Lao-Tai families; iii. Late entry in turn appears to reduce the total time that children have to study, although the ultimate impact of late entry does depend on how far the family intended the child to study; iv. Leaving school before completing grade 5 is prevalent among certain groups, especially girls from ethnic groups with disadvantaged backgrounds, who live in rural and remote areas. FIGURE 4. UNDERLYING REASONS FOR A SIMPLE FRAMEWORK TO DROPOUT AND NEVER 14 EXPLORE DROP-OUT: “PUSH” VS “PULL” FACTORS ATTENDING SCHOOL 24. When school attendance laws are not effectively Factors thats “push” Factors thats make enforced, families are largely free to decide whether the children out of the parents “pull”children child will attend school or not. Some individual children system out of the system may also make this decision for themselves. A decision to Insufficient “supply” Poverty attend school is a form of investment, where the skills and Schools credentials that are obtained through education will Classrooms Costly attendance generate a long-term payoff in the form of higher future Teachers Schools fees earnings for the individual child. In addition to this narrow Other costs (uniforms, focus on future earnings, school attendance can also be transportation, meals,..) justified by other (non-monetary) factors, such as providing Inadequate school environment (dangerous, children with an enjoyable activity (learning, playing, being not inclusive, tec) Opportunity cost (i.e. have to contribute to household with friends), or conforming with societal expectations economy) about education. Differences in the underlying reasons Inadequate support for disabled students for sending children to school can have consequences for Low perceived value of the effectiveness of certain kinds of interventions, but the Low quality of education education most important point is that school attendance is justified Own elaboration by some kind of expected benefit.5 5 The discussion in this section is based on fairly standard economic models of school attendance decision-making, which can be traced to the earliest versions of human capital theory: see Schultz, T.W. (1963). Human capital. Chicago: University of Chicago Press; and Becker, G.S. (1967). Human capital and the personal distribution of income: An analytical approach. Woytinsky Lecture no. 1, Ann Arbor: University of Michigan. There are also very large literatures in the economics andsociology of education covering topics related to the determinants of schoolattendance, as well as student outcomes like achievement. See Glewwe, P., E.A. Hanushek, S. Humpage and R. Ravina (2015), “School resources and educational outcomes in developing countries: A review of the literature from 1990 to 2010”, in Glewwe, P. (2014), Education policy in developing countries. Chicago: The University of Chicago Press; Fuller, B. and P. Clarke (1994), “Raising school effects while ignoring culture? Local conditions and the influence of classroom tools, rules and pedagogy.” Review of Educational Research, 64, 119-157; and Marshall, J.H. (2011). School quality signals and attendance in rural Guatemala. Economics of Education Review, 30(6), 1445-1455. 18 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR RETURNS TO EDUCATION 27. These expected benefits must be compared to the costs of attending schooling. In a narrow economic view, 25. The clearest benefit of attending and completing school investment is only justified when the expected additional years of schooling are the economic returns to benefits outweigh the costs. But non-economic discussions education – that is, the increase in salary associated with of school attendance also refer to costs, for the simple an additional year of schooling. In Lao PDR, these returns reason that the costs are a potentially serious constraint are positive, although smaller than those seen in other for many poor families, regardless of the underlying countries6. Over the last five years, wage workers in Lao ‘calculation’ that goes into deciding whether or not PDR earned an estimated average of three to four percent children go to school. Compared with benefits, costs are more for each additional year of schooling. Higher levels also relatively easy to calculate. They include direct costs of education not only command higher wages, but also in the form of fees (official and unofficial), transportation yield positive returns on investment. In short, education costs, uniforms and school materials. There are also is a good investment in Lao PDR. indirect costs—usually referred to as “opportunity costs”—that refer to the lost time associated with school 26. The labor market rewards men and women for their attendance where the child could be helping the educational investments differently, with women achieving household with some kind of productive work. greater returns to education than men. Regardless of the data source, average returns to education (when all workers 28. What kinds of factors therefore influence school are included) remain positive, and above three percent. attendance and dropout? Poverty and other domestic Using the STEP data set, returns to education for women factors such as family size and work calendars help determine in 2011 are much higher than with any other data source, the household’s ability to pay for schooling, and the time at around 6.9 percent (compared to 5.2 and 2.2 with other that is available for a child to go to school. When a child sources). For males, estimates also differ by data source. is needed to help around the house, or in the field (or Returns for males are around 2.5 percent (using the STEP shop), then the family may be forced to pull them out of data set), compared to 3.4 percent and two percent when school. The same is true when the family simply does using other data sources. not have the resources to afford the direct costs of 6 schooling. See Montenegro and Patrinos (2014) Returns to Schooling Around the World, World Bank. 29. School characteristics can also affect how long children remain in school. School environment variables FIGURE (or school climate) make up one set of influences, since children may suffer from problems related to bullying, or ANNUALIZED RETURNS 15 TO EDUCATION may simply find school to be very boring. Quality is another potentially important characteristic. When teachers are frequently absent, or learning materials and basic infrastructure are missing, then there may be concerns about whether or not children are likely to learn. Frequent repetition, presumably caused by not learning, 7.0% can also reduce the perceived benefits of staying in school. PERCENT (0-100%) 6.9% 5.3% 5.2% 3.5% 4.2% 3.4% 3.1% 3.0% 2.5% 1.8% 2.2% 1.9% 0.0% LECS IV LFS STEP 2007/08 2010 2011 Male Female All Source: World Bank (2013), Skills for Quality Jobs and Development in Lao PDR Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 19 30. School access refers to a specific sub-set of school characteristics. For young children, the most basic indicator of access is whether or not there is a primary school in the child’s home village. But access must also be viewed in terms of the availability of post-primary levels of schooling, since the importance of completing primary school may be reduced when post-primary schooling options are far away. Class size and crowding can also be thought of as access characteristics, since they potentially impact the student’s access to the teacher, and to learning materials. 31. Finally, culture is a commonly cited influence on school attendance decisions. In societies with strong patriarchal norms, girl’s schooling may be valued less, or even viewed as threatening to the established order. There are also general cultural beliefs about the importance of education that can, in turn, impact a household’s prioritization of school attendance and the help that is provided children outside of school. This is also related to child-rearing practices and the degree to which parents provide a nurturing environment in the home that facilitates learning and going to school. 33. The push and pull dichotomy helps with the categori- 32. Taken together, these different groups of factors can zation of potential interventions to reduce early grade be organized into a fairly simple framework. Early grade dropout. On the one hand, interventions designed to dropout is a result of household socioeconomic and address a household’s ability to afford schooling (such as cultural factors that result in the child being pulled out Conditional Cash Transfers, CCTs) have the advantage of of school. Alternatively, children may in effect be pushed directly targeting the dropout problem by helping poor out of school due to specific environmental characteristics families. But they can be costly and difficult to implement, (like school climate), or due to larger concerns about the especially if factors like agricultural work calendars are in schooling system (access, overall quality, etc.). Other, play. On the other hand, investments that make schools related ways of organizing these factors is to think of better places for children to spend time have the potential long term versus short term constraints, and supply to reduce dropout and improve other outcomes, like versus demand factors7). student learning. But these interventions require changes in school management and the teaching and learning environment, which are challenging areas to address, and are likely to take time to impact. These issues are returned to in later sections of this report when different interventions are considered. 7 Handa, S. (2002). Raising primary school enrolment in developing countries: The relative importance of supply and demand. Journal of Development Economics, 69(1), 103-128. 20 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR SUPPLY OF SCHOOLING 34. Supply constraints are an issue, but they do not Regardless, when combining data on access to a school explain all drop-out. In rural areas a significant proportion with data on attendance, it can be seen that about 40 of schools only report having enrollments until grades percent of children aged 12-16 who left school after grade three or four, and do not appear to offer the full complement three are from villages where the school(s) did not report of primary grades (Figure 16). In remote rural areas (no enrollments in grades four or five. However, about 70 road), this problem is especially acute: 27 percent only percent of grade four dropouts were enrolled in village report enrollments until grade three. It is not exactly clear schools that did offer the full five grades. So we cannot from Figure 16 whether this supply cutoff reflects a lack conclude that the problem of early grade dropout is solely of teachers and facilities, or if families in the village are related to supply constraints, though it does seem to be simply not keeping their children in school past these a problem in remote rural areas. grades. FIGURE PERCENTAGE OF SCHOOLS THAT OFFER 16 A PARTICULAR GRADE, BY LOCATION NATIONAL 5% 9% 3% 84% 3% RURAL 27% 8% 63% (NO ROAD) 3% RURAL 7% 9% 82% 2% 3% URBAN 95% 0% 25% 50% 75% 100% Grade 1-2 Grade 3 Grade 4 Grade 5 Source: EMIS (2015) DEMAND FOR SCHOOLING 35. For demand, the potential range of influences is much broader. Most discussions focus on direct and indirect costs associated with schooling, which can present substantial barriers to participation for the poorest fami- lies. This is because they cannot afford even the basic school materials that are required (uniform, pencils, etc.), or because their children are needed to do other things around the house, field, or market. 36. However, low demand for education can be a result of other factors, although these are harder to establish Alternatively, families may have concerns about the quality empirically using household survey data sources, and of the education system, or question the ability of schools usually require more qualitative data. For example, families to address their child’s specific needs; this is especially may not see much utility in sending their children to relevant in ethnic minority contexts where issues related school, perhaps because they expect them to work in the to language are closely tied to relevance. Perceptions fields, or become mothers, and do not see aneed for even of low quality schooling can also help explain dropout, basic literacy and numeracy skills. especially when the child is not learning. Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 21 37. Non-attendance in Lao PDR seems to be driven mainly For 6-9 year olds, the main reason is that the child is still by low demand for education among certain groups of considered to be too young, which again highlights the the population.This is perhaps linked to perceptions of low important issue of late entry. Among older children quality or lack of relevance, as well as potentially high (10-14) there are some concerns about supply, but the costs, especially opportunity cost. Figure 17 begins with most commonly cited reasons are related to demand. a summary of the reasons why children have never About one third of children aged 10-14 who are not in attended school. Among older Lao people, the proportion school cite reasons related to direct and indirect costs of who did not have a school to attend is quite high, espe- schooling. But an even larger group (41 percent) cite “not cially in comparison with younger cohorts. This result is interested” as the reason why they are not in school. simply a reflection of the steady improvement that has been made in school access. FIGURE REASONS FOR NEVER ATTENDING SCHOOL 17 BY AGE GROUP 4% 3% 3% 1% 1% 6-9 61% 4% 5% 6% 14 2% 3% 10-14 7% 5% 10% 8% 8% 5% 12% 41% AGE 5% 3% 3% 20 15% 9% 8% 16% 41% 2% 2% 2% 40 42% 6% 6% 7% 10% 20% 0% 25% 50% 75% 100% No school/teacher Too young Disabled Cannot afford Have to work (paid) Have to work (chores) Family won't allow Not interested Not worth it School not safe Source: LFS 2010 22 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR 38. Ethnic groups —especially Mon-Khmer and Chinese-Tibet—are much more likely to cite “Not Interested” as the reason for never attending school: roughly 70 percent of these households chose this reason. Figure 18 compares reasons for never attending school (for 10-14 year olds) by ethnic group and gender. Several results stand out. Firstly, the explanations are similar to those given in the 2010 LFS data, with “Not interested” being mentioned most frequently. However, the prevalence of “Not Interested” among ethnic groups does not appear to be explained by a cultural “mismatch,” or concern about the language or adequacy of the teaching force. FIGURE REASONS FOR NEVER ATTENDING SCHOOL 18 BY ETHNICITY 5.3 LAO- Female 9.4 48.9 5.6 3 27.8 2.6 3.2 TAI Male 10.2 5.5 48 8.5 22.1 2.6 2.6 1.4 1.7 3.2 MON- Female 4.2 79.6 4.6 2 2.4 1.2 KHMER Male 4.6 67 4.4 13.3 5 2.4 CHINESE- Female 61.7 35.9 TIBET Male 4 88.2 7.8 HMONG- Female 46.7 31.6 18.1 3.7 IU-MIEN Male 25.9 32.8 32.8 8.6 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Too young Too expensive Not interested Have to work School too far No Teacher Illness Language Other Source: LECS V Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 23 39. When asked to explain dropout, very few families cited factors when parents consider enrolling the children in school supply, with most citing reasons related to demand; school. However, in line with those who never attend the results for ‘never attended’ are very different when school, a substantial proportion of these families compared with those in Figure 17 . “School is too also claimed that the child was “Not interested”. This expensive” and “Have to work” are more commonly cited explanation was more commonly cited for females than for drop-out than for never attending, indicating that cost males (figure 19). (both direct and opportunity cost) are important FIGURE REASONS FOR DROPPING OUT OF SCHOOL 19 BY LAST GRADE ATTENDED AND GENDER GRADE Female 38 23 33 6 2 1-2 Male 62 15 5 10 4 3 1 Female 50 25 20 5 GRADE 3-4 Male 52 19 22 8 3 GRADE Female 28 36 23 5 6 5 Male 52 21 20 3 3 2 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Too expensive Not interested Had to work No teacher Illness Language Other Source: LFS (2010) 40. Taken together, the results in Figures 17 and 19 help to better understand the underlying dynamics of school attendance in Lao PDR, which in turn will help identify policy options and new interventions (see below). In terms of supply-side influences, relatively few households cite a lack of schools (or teachers) as the reason for not entering school, or dropping out. However, these responses may be somewhat misleading, since roughly 40 percent of children who left school after completing grade 3 were studying in a school that did not offer grade 4, while 30 percent of children who left after grade 4 had no grade 5 to go on to. These results again point to the need to make all primary schools in Lao PDR complete. However, it is important to note that not all children who drop out before completing primary school are doing so because of supply-side constraints. 41. Rather than supply constraints, the most commonly cited reasons for never attending or dropping out indicate a combination of low perceived value of education ("no a large proportion of children contribute to household interest", "no value", etc.) and high perceived cost. The chores and economic activity.As they get older, the cost of attending school can be relatively high for a poor potential value of children to their households only family, which mainly refers to the cost of uniforms, food increases(Figure 21). This appears to be especially true and transport (Figure 20). In terms of opportunity costs, for girls. 24 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR FIGURE AVERAGE YEARLY COST OF ATTENDING SCHOOL 20 (IN THOUSAND LAK) ANNUAL SCHOOL NATIONAL SAMPLE : RURAL ONLY: SPENDING ALL PRIMARY LSS AND PRIVATE (IN 1,000 KIP): LEVELS ONLY USS SCHOOLS ALL GIRLS BOYS Tuition and fees 97.4 69.0 119.2 1,453.4 29.2 28.3 30.1 Parent Association 13.3 8.6 21.0 9.2 11.6 10.6 12.4 Uniforms clothing 175.1 124.4 256.4 249.9 154.6 152.3 156.7 Textbooks 26.6 15.7 43.0 72.2 19.3 19.4 19.2 Other materials 68.3 50.3 97.0 112.2 59.7 59.3 60.1 Meals transportation 310.1 151.7 588.1 939.3 201.7 203.0 200.6 Other 138.1 79.9 230.4 363.5 102.8 98.8 106.3 Total Spending 828.7 500.0 558.1 939.3 201.7 203.0 200.6 Sample Size 7,622 4,719 2,779 201 5,922 2,777 3,145 FIGURE PERCENTAGE OF CHILDREN WHO REPORT 21 SUPPORTING ECONOMIC ACTIVITIES FOR THE FAMILY AND HOURS PER WEEK SPENT ON ECONOMIC ACTIVITIES 100 92.2 90.9 90 89.9 90.7 86.3 80 79.1 74 70 67.3 86.6 89.4 90.8 86 60 62.8 83.1 78.6 PERCENTAGE (0-100%) 73.2 55.8 50 65.6 58 50.5 39.9 40 41.3 41 28.2 39.8 39.7 38.5 36 36 37.8 35 30 32.1 33 20 35.3 35.3 40.4 34.4 38 39 34.7 32.3 31.5 Female (Economic) 30.5 10 Male (Economic) 25.9 Female (Home) 0 Male (Home) 6 7 8 9 10 11 12 13 14 15 16 AGE Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 25 30 26.4 HOURS PER WEEK 3.5 5.6 6.3 6.5 20.8 1.8 2.6 5.2 5.3 15 1.9 3.3 4.5 4.8 20.1 12 13.1 16.2 1.7 2.1 3.5 3.6 1.5 2 3.9 4 1.2 1.5 3 3.2 1.2 1.5 1.7 2.1 10 7.3 11 8.1 9.1 Female (Economic) 7.6 5 Male (Economic) 7.4 7.4 6.4 Female (Home) 6 0 Male (Home) 6 7 8 9 10 11 12 13 14 15 16 AGE Source: LECS 2012, 2013 B. LEARNING OUTCOMES 42. The low levels of learning outcomes among those 43. In order to analyze fluency and comprehension children who do attend school starts with low reading together, students in the EGRA sample were divided into ability in the early grades. The EGRA 2012 Survey Report four groups: (2014) had mixed findings in relation to reading ability - some were concerning and some were positive. Many i. Non-readers. Students unable to read any words students are only able to fluently read and comprehend correctly in the first (and easiest) row of the text. a simple 60-word text after Grade 4 –that is, it takes ii. Readers without comprehension. Students able to about four years of instruction to achieve this basic level read some words correctly, but unable to answer of literacy. On the other hand, the survey results also any reading comprehension questions correctly. show that there is grade progression. In each grade, iii. Readers with some comprehension. Students able students gain new knowledge and skills in all sub-tests. to read some words correctly, but able to answer However, although students are progressively learning three or fewer (out of five) reading comprehension to read more letters and words (and with increasing questions correctly. accuracy), fluent reading with comprehension is the iv. Readers with high levels of comprehension. ultimate goal of reading instruction. Students able to read most of the words correctly, and able to answer four or five (out of five) reading comprehension questions correctly. 26 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR 44. Figure 22 shows the distribution of students in each FIGURE of these groups. At the beginning of Grade 3, only 10 DISTRIBUTION OF percent of students have a high degree of reading com- prehension and are therefore in the top band. By the 22 STUDENTS BY READING SKILL AND GRADE start of Grade 4, 25 percent of students are able to read a simple story and extract most of the meaning, whereas in Grade 5, almost half of the students are able to show 100 the same degree of comprehension. These results 8% suggest that, on average, students make significant 90 25% gains in fluency and comprehension in Grades 3 and 4 but it is only in Grade 5 that the majority (86 percent) 80 of students are able to read with fluency and compre- 34% 47% % OF STUDENTS hension. This finding is relevant in terms of curriculum 70 expectations, since Grade 5 coincides with the last grade of primary school in Lao PDR. 60 44% 45. Figure 22 also shows the average composition of 50 classrooms in Grades 3–5, in terms of the ability of 25% students to read and comprehend a simple text. Teach- 40 ers around the world are no strangers to variation in student performance in the classroom, and the challenge 30 39% of adjusting their teaching strategies to student ability. Yet it seems that on average, Grade 3 teachers in the 20 17% 32% sample have the most challenging teaching environment, 10 as Grade 3 shows almost equal proportions of students 10% 13% at three out of four levels of performance. By the start 0 4% of Grade 4, both the proportion of students without com- prehension and the non-readers have decreased Grade Grade Grade 3 4 5 substantially (17 percent and 13 percent, respectively). In Grade 5, although most of the students are able to Readers with high levels Readers without read (with varying levels of comprehension), there are of comprehension comprehension still a few students whose poor reading and comprehen- Readers with some Non readers comprehension sion skills compromise their ability to access content material. Source: EGRA 2012 Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 27 46. If learning to read involves a second language, one that the beginner reader is still acquiring, the task of learning to map sounds to symbols is the same in general, but is severely complicated by the need to learn the second language simultaneously and, in some cases, Across all grades tested, learning outcomes were weakest in a second writing system or orthography8. The Lao amongst students without adequate Lao language skills. EGRA noted that many students had difficulty Figure 23 presents the distribution of student proficiency communicating in Lao language, but this also improves in reading fluency and reading comprehension, depend- over time. It is important to note that the survey did not ing on their ability to communicate in Lao language. follow the same children over time; it collected data through a random sample of students in each grade 8 Research has already established that there is an advantage in tested. Therefore the ‘improvements’ observed could be learning to read in the reader’s first language (L1) since the learner can rely on her existing language knowledge as a vehicle to map partly the result of actual improvements in Lao language sounds and symbols. Learning to read in a second language (L2) or a proficiency, but could also be due to a more homoge- second writing system would be facilitated to some extent by the fact neous composition of classrooms in Grades 4 and 5, as that reading is a process that firmly attached to language and writing systems. However, there are specific variations in languages, students with poor Lao language skills would be at high orthographies, and writing systems that will challenge the learner risk of dropping out before finishing primary school. differently. FIGURE STUDENT DISTRIBUTION IN READING FLUENCY 23 AND COMPREHENSION BY ASSESSED LAO LANGUAGE PROFICIENCY 100 1% 12% 90 14% 80 46% % OF STUDENTS 70 45% 41% 60 50 40 30 42% 22% Readers with high levels 44% of comprehension 20 Readers with some comprehension 10 21% Readers without 7% comprehension 0 5% Non readers Able Able Able to communicate to communicate to communicate proficiently in a limited way in only a few words 28 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR 47. Students identified to be able to communicate only a few words in Lao language9 were most likely to fall into the non-reader and reader without comprehension categories. By contrast, the majority of students considered to be able to communicate proficiently in Lao 48. When looking at the factors associated with learning language showed at least some reading comprehension outcomes, in both EGRA and ASLO, socioeconomic skill, with nearly half of these students demonstrating a characteristics of students are very strong determinants high degree of reading comprehension skill. In order to of learning. Ethnicity and socioeconomic status (SES) are make gains in student outcomes, this evidence suggests the main drivers of performance.Relatively few process that the complex interplay between language and variables are significant predictors of better language literacy must be considered and supported. scores, and the effect sizes are smaller than those associated with the main background variables (such as ethnicity and SES). Student responses for the frequency 9 they receive homework, and the frequency they use the Although the sampling strategy for EGRA did not allow for a contrast group by language proficiency, the survey team defined a simple yet Lao textbook in class, were averaged at the classroom consistent criteria to broadly classify students by their ability to level.This should make them more robust indicators of communicate in Lao language during the interview, the student questionnaire questions and three additional questions where teaching processes, and not just capture the student’s students could talk about freely about their family or hobbies at level of commitment (Figure 24). school. FIGURE SUMMARY OF EFFECT SIZES FOR SIGNIFICANT 24 PREDICTORS OF TOTAL LAO SCORE IN ASLO EFFECT SIZE IN STANDARD DEVIATIONS 0.4 0.3 0.2 0.3 0.06 0.05 0.1 0.05 0.05 0.03 -0.07 0.11 0.11 0.09 0.08 0 -0.1 Lao-Tai SES Teacher female Student female Family attention Freq. Lao homework Freq. Lao text use Freq. of individual study Freq. of conversation Freq. set tests -0.16 -0.1 Hours of class in pm Extra hours marking -0.2 Has preschool Baseline Variables Student- Teacher-reported reported Source: ASLO, 2013 Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 29 49. In EGRA, the results are similar, with socioeconomic characteristics having the largest effects on learning outcomes, although teaching practices are shown to have larger effects than in the ASLO results. Two teaching and learning process indicators have positive effect sizes of the same magnitude as the most significant back- ground variables (e.g., gender, SES). Many variables from the EGRA questionnaires are not included in the summary due to a lack of statistical significance. These include the frequency of parental meetings, the frequency of pedagogical advisor visits, reading hours, coverage of pedagogical advisor topics, the existence of a reading corner at the school and the number of hours of its use. The preponderance of insignificant classroom 50. Several factors could be driving this strong environment variables, and general teaching variables, association between socioeconomic characteristics and is not an unusual finding in quantitative research, and performance – starting with differences in child development highlights the built-in challenges of identifying root driven by malnutrition. In Lao PDR, 44 percent of causes for low reading scores. children under five are estimated to be stunted, 27 percent are underweight, and 6 percent are wasted (Figure 25). Malnutrition in Lao PDR remains persistently high, and the factors associated are largely related to behavior. Stunting rates differ markedly among different groups and across provinces. 51. The aggregate number hides substantial spatial variations, with far worse outcomes in some provinces, and the absolute number of stunted children may actually have increased. Stunting rates are higher than 60 percent in Sekong and Phongsaly for example, far higher than the national average of 44 percent (see Figure 26). Child development has a very strong association with a number of outcomes, including learning. While a more comprehensive picture of different dimensions of child development outcomes (being supported by the LEARN project in collaboration with the World Bank) sheds more light on the particular needs of young children in Lao PDR, the differences in stunting rates are a clear sign that this is a key priority. 30 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR FIGURE NUTRITION INDICATORS 25 AND STUNTING RATES BY PROVINCE POOR PERFORMANCE ON KEY NUTRITIONAL INDICATORS Comparison of Lao PDR provinces to LIC/LMIC Countries, 2012 Stunting Underweight Wasting 60 SEK PISE Timor-Leste LKK Lao PDR % OF CHILDREN <5 LPB SEK Timor-Leste Cambodia Indonesia LNT 40 Myanmar PSL Lao PDR Cambodia Vietnam HP LNT Myanmar 20 VTEP LRB Timor-Leste Indonesia VTEP Vietnam Indonesia Cambodia VTEP Lao PDR PSL Vietnam LRB 0 RR Laos Other Laos Other Laos Other Sources: LSIS (Laos) & WDI (Other) HP = Huaphanh; LNT = Luangnamtha; LPB = Luangprabang; PSL = Phongsaly; XEK = Sekong; XK = Xiengkhuang; VTE = Vientiane Capital STUNTING RATES BY PROVINCE: 2011/12 Phong sary Luang nam stunting rate (%) tha Oudo Bokeo mxay Huaphanh (60,70) Luang prabang (50,60) Xieng (40,50) khuang Xayabury (30,40) Vientiane. P Borikhamxay (20,10) (10,20) Kham Viantiane. C muane Savannakhet Saravane Sekong Champasak At tapeu Source: LSIS, 2011, 2012 Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 31 predicts better ORF scores is perhaps not surprising, and it should be noted that student ORF scores are lower when teachers report more frequent use of oral evaluations. Nevertheless, at the very least these results are a reminder of the importance of having opportunities to read out loud, with supervision, and (ideally) with feed- back provided to correct mistakes. The second set of variables are for homework, which was one of the few significant predictors of ASLO achievement. As is the 52. Socioeconomic characteristics are not the only factor case for reading opportunities, the homework variables explaining outcomes: classroom practices matter. Firstly, from the student interviews are significant individually, several indicators related to oral reading activities (reading as well as when measured as a classroom average(which aloud in class, retelling a story, etc.) are associated with is likely to be a better indicator of the use of homework). higher oral reading fluency (ORF) scores. This association is significant, based on the answers to questions posed This is another result that is hardly ground breaking, to students, as well as on the answers to some of the but it again points to the importance of having questions in the teacher questionnaire. That a opportunities to improve reading in different formats higher number of oral reading opportunities (Figure 26). FIGURE SUMMARY OF EFFECT SIZES FOR SIGNIFICANT 26 PREDICTORS OF TOTAL LAO SCORE IN ASLO 0.4 EFFECT SIZE IN STANDARD DEVIATION 0.3 0.2 0.33 0.29 0.31 0.1 0.06 0.18 0.17 0.16 0.16 0.2 0.13 0.13 0.12 0 -0.1 SES Female Textbook Meals Read aloud (class average) Homework explained frequency (class avg) Homework ever (class avg.) Homework frequency (class avg.) Read aloud (individual) Review of homework Students read aloud Students retold story -0.16 -0.1 -0.29 -0.32 -0.2 -0.3 Repeated Oral evaluations -0.4 Baseline Variables Student-reported Teaching strategies Source: EGRA, 2013 32 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR 53. The importance of classroom and school practices of students do not understand Lao well, in this study needs to be investigated further, but there is clear indicative they have not been observed speaking other languages evidence of its importance. In 2012, alongside the in the classroom. A consistent concern voiced by teachers surveys for the EGRA report, a tool to assess school and was the difficulty in catering to students of ethnic groups classroom practices was used in a sample of schools, to and including them in teaching and learning activities. explore if there were differences in the school and Achieving greater inclusiveness is an area worthy of community environment that made some schools and further research. classrooms more conducive to reading acquisition. A tool was selected for the purposes of this small exploratory 56. Classroom activities with varied forms of reading, study, which included Snapshot for School Management including students reading aloud, teacher assisted reading, Effectiveness (SSME) questionnaires for teachers, buddy reading, group reading and comprehension questions students and principals, plus an inventory of school were more likely to be observed in the top performing infrastructure and a log for classroom observations. All cohort than in the low performing cohort. Classroom were translated and adapted to the Lao context. resources more likely to be observed in the top performing Additional questionnaires for pedagogical advisors and cohort included a Lao alphabet, Lao spelling and vocabulary the village education development committee (VEDC) words, displays of student work, a good supply of members were developed, reflecting their particular roles printed materials, sufficient classroom space, a reading in supporting teaching and learning in their districts. corner and marked exercise books. 57. Principals in the low-performing cohort reported teaching more hours in a week than their colleagues in other cohorts. They also reported more multigrade classrooms and irregular school closures. In the low and middle-performing cohorts, volunteer teachers were reported and teachers were more likely to leave the school for administrative duties. Principals reported higher levels of teacher absenteeism among these cohorts. 58. Principals in the top performing cohort were more likely to report a higher ratio of textbooks to students, and that it was easier to provide assistance to teachers. 54.The SSME exploratory study took as its basis nine In the middle and top-performing cohorts, principals case study schools, selected from those schools who were also more likely to report regular assessment of participated in the 2012 Lao EGRA survey. Three schools students than in the low performing cohort. from the top performing 25% of schools, three from the middle 50% and three from the lower 25% were selected 59. Schools in the top performing cohort also had better at random.Teams visited these schools to interview facilities. They were more likely to have a water supply, principals, teachers, students, pedagogical advisers, a girls’ washroom and library. Electricity was not supplied VEDC members and to observe classrooms. In total, 160 to any of the schools in the low-performing cohort. students, 27 teachers, nine principals, 25 pedagogical However, it was present in at least some of the advisers and 18 VEDC members were interviewed, and schools in the other cohorts. 26 classrooms/lessons were observed, across six prov- inces and nine villages. The design of the study was largely qualitative rather than quantitative. 55. Lao language was the only language that enumera- tors observed in use by by both students and teachers. The study’s findings are in contrast to reports of students for whom Lao language is clearly not the language they are most proficient in. While a proportion Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 33 60. The role of the pedagogical advisors also seems crucial, as they provide support to teachers and to VEDC members. Both pedagogical advisors and VEDC members were inter- viewed. The typical pedagogical advisor interviewed had worked in that capacity for 12 years, and most had gradu- ated from a training college. They were mostly male. In the middle and top-performing cohorts, all pedagogical advisors had graduated from training college. This was less common in the low-performing cohort. Those in the top-performing cohort reported spending less time on administrative work, more time observing teachers, and were more likely to report visiting schools. These pedagogical advisors in the top-performing cohort were also more likely to share their evaluations with the school. 62. Overall, the differences observed between schools with higher and lower performance in the EGRA survey included 61. Most VEDC members interviewed were local to their differences in resourcing, teacher/principal experience, village, and those who were not had been in the village for teacher support, education level of pedagogical advisors and an extended number of years. The committee members from classroom activities. Further, there are observed differences the top-performing cohort reported having more experience between both top and low performing cohorts in regards to than those in other cohorts. These community members community engagement and support. This evidence were more likely to meet more often, more likely to monitor suggests that early grade reading is enhanced by participa- student absence and more likely to encourage students to tion by the community beyond the school boundaries. attend school in the top performing cohort. C. CONCLUSIONS AND POLICY AREAS CAUSES FOR DROP-OUT AND LOW PERFORMANCE 63. It is important to stress the great progress Lao PDR has made in enrolling children in primary school. The expansion in the number of schools and enrollment has been truly impressive. Problems with children never enrolling or drop- ping out are now concentrated in remote rural areas, and are particularly prevalent for the poor and some ethnic groups. This facilitates the targeting of programs to ensure that every child in Lao PDR is enrolled in school. 64. Despite this progress, Lao PDR faces persistent problems of children remaining out of school or leaving primary school early in some populations. They tend to be poorer, live in rural and (especially) remote areas, and come from non-Lao Tai ethnic groups. Early leavers are more likely to be girls, who start leaving school earlier than boys, at ages 10-11. The number of out of school children is significant, but it is different according to different data sources. There are indica- tions that the number may be overstated in EMIS. Thus, a first recommendation is to improve the quality of the underlying enrollment data in EMIS. 34 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR POLICY AREAS 67. The analysis presented here points to three priority areas for action: (i) focus on learning outcomes, especially early grade reading ability; (ii) increase the demand for education for those who perceive education to have a low value; and (iii) continue to increase and improve the supply of schooling. These priority areas and policy interventions are presented 65. More importantly, too many children who remain in in Figure 27. school show very low levels of reading ability and learning. This may be partly due to the rapid expansion in the supply 68. Firstly, there is there is the need to focus on learning of schooling, which can sometimes come at the cost of lower outcomes, especially early grade reading ability. Simply quality of education. There is overwhelming evidence about attending schooling is not a sufficient condition for learning. the importance of learning outcomes beyond schooling The generally low results on early grade reading ability signal attainment. Ensuring that children are not only in school, a dismal performance of students in remote schools and but that they develop both cognitive and non-cognitive skills those with more nonLao-Tai students. This may explain why while they are there, is of fundamental importance. In 2012, these households decide to pull their children out of the in a sample of non-remote schools(of mixed ethnicities but school system - because they do not see the value of attend- never with a non Lao-Tai majority), about 1/3 of 3rd graders ing. This may be due to low quality teaching, or a lack of in 2012 could not read a single word in a text and about 70 materials, in addition to the language barriers for students percent could not understand what they read. It is safe to with poor Lao language skills. School readiness through assume that the result in remote and ethnic schools is worse. improved nutrition and early childhood education is crucial. Improving learning in general, and reading ability in particular, should be top priorities for the education sector. 69. The Early Grade Reading Assessment report emphasizes the importance of improving teaching and learning methods 66. Aggravating the problems with low quality of education for reading in the early grades. The ongoing effort to revamp is the fact that the lowest performers in reading and learning the curriculum would have the biggest impact if it focuses assessments are also those with the highest rates of drop-out, strongly on early grade reading. Other factors that were which suggests that the two phenomena are connected. The mentioned in the workshops10 include: (i) concerns regarding main reasons for never attending or dropping out are related school safety; (ii) a school environment that does not facilitate mostly to low perceived quality or relevance of education. teaching-learning; (iii) limited capacity to accommodate When looking at the determinants of learning, the same children with disabilities; (iv) lengthy and poor quality primary factors that drive drop-out are also associated with low early curriculum; (v) low student learning achievement compared grade reading ability and low performance in standardized to the expected grade level competency; (vi) inexperienced tests. This suggests that low progress in learning may be teachers, and low quality of teacher training and professional resulting in disinterest and eventual dropout. Unfortunately, development; and (vii) lack of pedagogical support. this is not possible to test empirically with existing data, since we do not have access to a panel of students with socioeco- 10 nomic information, learning outcomes and school enrollment. A series of regional consultations on early grade were held in Vientiane Province, Savannakhet, Oudomxay and Xieng Khuang. Even if such detailed data is gathered only in a sample of Participants included 160 officials from 80 Districts Education and districts, it would allow a much more thorough analysis of Sports Bureau (DESB) and representatives from all Provincial Education the relationship between low academic performance and and Sports Services (PESS) responsible for Primary Education and Monitoring and Evaluation, March 2015 drop-out. Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 35 70. Secondly, the evidence in this report shows that the main 72. There are also relatively high costs associated with educa- factors driving never enrolling and drop-out are related to tion, especially opportunity costs, and again this influence is the need to increase demand for education. This starts from more prevalent for certain communities and household types. preventing late enrollment by addressing parents’ concern The findings from workshops suggest that demand for child that children are too young to enroll (at the right age) labor (for household chores and economic work) and poverty through providing information and through increasing school level contribute to early grade dropout and repetition. When safety. Late enrollment is associated with lower attainment, poor families face high direct and indirect costs for sending so there is a need to inform parents of the importance of their children to school, they may be especially sensitive to education. Many families of non-enrolled children aged the perceived relevance and quality of schooling. In other between 6-9 years cite “too young” as the reason, and a words, it is a mistake to attribute out of school children to significant share of households report “completed school” “irrational” or shortsighted parents who are compromising even if their child has left after grade 3 or 4 (see full report). their children’s futures for the household’s short term gain. The issue of late entry needs to be followed up with more These households may have very real concerns that need to qualitative research. be addressed, which in turn highlights both the need for more research, and for policy solutions that address these 71. In addition, low perceived value and the relatively high underlying factors. indirect cost of education in certain contexts and for some population groups is one of the main reasons for enrolling 73. Finally, there is still a need to increase and improve the late or dropping out before completing primary education. supply of schooling, both in terms of providing complete The prevalence of “not interested” as an explanation suggests schools with qualified teachers in remote rural areas, and the need to improve the relevance of education in some ensuring that school materials (textbooks and learning mate- contexts, particularly in remote areas and among ethnic rials) are available – and being used. About 40 percent of students. Language is an important factor in determining children who drop out live in a village that does not offer the the relevance of education. If the language of instruction is grade in which they dropped out. This is especially significant Lao, it is essential that students arrive in grade 1 with good in the transition between grade 5 and lower secondary Lao language skills. It was cited in the consultation work- education. Safety also seems to be a concern for about 3 shops that language difference is one of the barriers for percent of parents who choose not to enroll their children. learning. This needs to be addressed in early childhood While this is not a high percentage, it is still important to education and pre-primary education. Non-formal education address. Unfortunately there is no specific information about (such as that provided by Community Child Development what the perceived lack of safety relates to (building safety Groups) can provide an avenue to introduce children to Lao versus bullying, for example), so this warrants further language. However, the effectiveness of this approach has exploration. not been yet tested. Mother tongue instruction has been shown in other countries to be an effective way to improve learning outcomes, but it is not an easy policy to implement nationally when you have many languages, and have class- rooms with a significant mix of languages (and low availability of teachers). In the Lao context, experimenting with different approaches and rigorously evaluating their impact is likely to be the best approach before scaling up any of these language policies. 36 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR 74. This diagnostic points to three potential policy areas Second, there is a need to increase demand for schooling, which, based on existing evidence in similar settings, address reducing the indirect costs of attending school through the root causes identified in the diagnostic. The areas are stipends. For certain populations, reducing this cost may prioritized in accordance with the analysis presented in this mean compensating for the opportunity cost of attending report. Improving the quality and relevance of education, with school (i.e., what the student would earn if she were attending a focus on early grade reading and especially in rural areas and contributing to household activities, whether economic and for ethnic populations is considered the number one or household chores). In this case, targeted scholarships can priority. In addition to the overwhelming evidence about provide incentives for attendance and even learning (if they the importance of skills on labor market outcomes11 , the are also based on merit and good progress). Lastly, there is perceived low quality and relevance of education seems to still a need to continue to improve supply (completing schools be one of the most important drivers of children dropping and ensuring teachers are available to provide all five grades out. Improving quality is thus also a strategy for improving and pre-primary in all primary schools). enrollment and reducing drop-out. 11 Lao Development Report (2014) and Lao Skills Survey (2013) document the importance of skills in the adult population for labor market outcomes in Lao PDR. FIGURE PRIORITY AREA 27 AND POSSIBLE INTERVENTIONS TIMELY AND UNIVERSAL ENROLLMENT PREVENTING DROPOUT ENSURING LEARNING 1. 2. 3. FOCUS ON IMPROVING INCREASE DEMAND CONTINUE TO INCREASE LEARNING OUTCOMES FOR SCHOOLING AND IMPROVE SUPPLY i. School readiness, i. Stipends and targeted i. Complete incomplete nutrition and language scholarships, School schools Lunches ii. Improve early grade ii. Learning materials reading teaching ii. Information & textbooks methodology Campaigns iii. Scholarships for rural iii. Focus on classroom iii. Early childhood students to become practices education teachers insame areas iV. Introduce flexible and iV. Introduce flexible iV. School safety and multi grade curriculum relevant to inclusion methodologies remote areas (i.e. Escuela Nueva) Possible Interventions to Address Root-causes Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 37 b. Improve early grade reading teaching methodology. Improving reading outcomes is perhaps one of the most important priorities in the sector – and the right methodology of instruction can have very large impacts. The EGRA report points to reading instruction as one of the key priorities for the country. The Global Partnership for Education is supporting a pilot of this methodology. It is important that this pilot is evaluated before being incorporated into the national curricu- lum. Getting the methodology right is fundamental to ensuring thatearly grade reading ability improves in Lao PDR. c. Focus on classroom practices. There is clear evidence in the EGRA report that school and classroom practices matter. However, the sample in that study was relatively small, so a more comprehensive look at classroom practices would be of great value. In the short term, strengthening the role of peda gogical advisors seems a clear priority. d. Introduce multi grade methodologies proven to be relevant and effective for rural areas. In other countries, delivering the curriculum in short modules has proven to provide the flexibility needed for rural areas and in multi-grade 75. Improving learning, especially early grade reading, is schools. Delivering good quality education that is relevant for perhaps the most important policy priority for Lao PDR, and rural contexts is a challenge that many countries have faced. not only to prevent children dropping out. If children do not Rural education has certain characteristics which make learn how to read, they cannot absorb the rest of the traditional education difficult to deliver: difficulty in getting curriculum, which puts them at risk of dropping out. The EGRA teachers to these areas, content that is not relevant to the rural report identifies the current methodology of instruction in context, and particularities in the schedules because of reading and writing in early grades as a key constraint for agricultural work. This demands flexibility, the ability to deliver effective teaching and learning. With support from the Global content effectively in multi-grade settings and a new approach Partnership for Education, the Ministry of Education and to the relevance of education. Educational models like Escuela Sports is revamping this methodology. Continuing this work Nueva13 in Colombia have proven to do this well by developing to ensure that the methodology is effective is crucial to modular content that can be delivered simultaneously to prevent drop-out, among other things. students in different levels and with flexibility in schedules. Vietnam is implementing the model with World Bank support; a. School Readiness, including nutrition and language. Lao PDR also has the right conditions to implement these Improving child nutrition is a clear country priority and this modalities. However, because the context of Lao PDR is different would contribute to improving child development, school (and difficult), any of these initiatives should be piloted and readiness and contribute to subsequent learning12. Similarly, rigorously evaluated before being scaled up. guaranteeing that students can learn in Lao language by the 13 time they reach primary school is necessary for them to learn See McEwan (1998) and Forero (2006) for a description and how to read and absorb the curriculum, but this is difficult to evaluation of the Escuela Nueva methodology. Filmer and Schady (2009), Barrera-Osorio & Filmer (2014). do. Community based early childhood education (such as that provided by the World Bank-supported Community Child Development Groups) can provide a good avenue to test interventions for language development including using mother tongue. Because of their complexity, these strategies should be piloted and evaluated before being scaled up. 12 For a full set of recommendations on nutrition see Pimhidzai (2016). 38 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR This is especially true for girls in poverty living in rural and remote areas. They are also more likely to belong to certain ethnic groups. Targeted scholarships based on need and merit have proven to have large impacts in school attendance and retention.17 In the case of Lao PDR, the analysis shows that a scholarship targeting ethnic girls in remote areas, especially in the transition to lower secondary education, would help reduce the cost of attending school and would likely lead to improvements in completion and transition rates. School lunches can also be an incentive for parents to enroll their children in school. 18 b. Information campaigns. Information about the value of schooling has been shown to have significant impacts on school attendance.19 When information about returns to 76. Increasing demand for schooling. The second pillar education is not available, and is only observed through peers includes interventions targeted to increase demand for school- in the community, parents may perceive those returns as low. ing. Firstly, addressing the relatively high indirect costs that Children who continue their studies very often move away some households face through stipends can be an effective from the community, especially in rural areas, so the returns way to increase school enrollment. These sorts of stipends are are not observed by members of the community. The fact intended to address the cost of attending. In addition, targeted that children leave the community may be perceived as a scholarships (where need is combined with merit) or conditional negative return for parents who need the assistance of their cash transfers (where the transfer is conditional on the recipient children in household chores, or because of certain cultural meeting certain conditions, including attending school) have perceptions. This is likely to be more important for girls than been proven to to increase enrollment, retention and facilitate for boys, as reflected in the analysis presented in this report. transitions between levels in many different countries and However, information about the importance of and returns settings. The evidence from the impact evaluation of a scholar- to education only seems to be effective when education ship program for girls in Cambodia (based on a combination quality is at acceptable levels. Otherwise, similar programs of need and merit) is clear: these types of programs can have have shown no impact on drop-out.20 In Lao PDR, community very large impacts on reducing drop-out and improving campaigns through the Village Education Development completion.14 Providing information to students and families Councils are under way for Early Childhood Education, with has been proven to result in reduced drop-out and higher prob- support from the World Bank, which could provide an ability of completion.15 Secondly, considering the association opportunity for LEARN to generate synergies with the between late enrollment and grade completion, preventing late World Bank-supported government programs. enrollment should be a priority. Thirdly, since dropping out is partly driven by the high cost of attendance (both in terms of 14 Filmer and Schady (2009), Barrera-Osorio & Filmer (2014). the direct cost and the opportunity cost of attending school) 15 The Heterogeneous effect of information on student performance : reducing the cost by providing targeted scholarships may be evidence from a randomized control trial in Mexico WPS7422 Fecha del effective in keeping children in school. Lastly, improving the documento: SEP 28, 2015Avitabile,Ciro; De Hoyos Navarro, Rafael E. de Hoyos, Rogers. See also Jensen (2014). relevance and form of delivery of education in rural areas has 16 Mckewan (1998), Forero (2006). also shown significant effects on both cognitive and 17 Fimer and Schady (2006) non-cognitive outcomes16 in similar settings. 18 See Lawson (2012) for a systematic review of the evidence on school feeding programs. Alderman and Bundy (2011) for a broader analysis a. Stipends and targeted scholarships. The evidence presented of the impact of these programs. See also Impact evaluation of school feeding programs in Lao PDR. AM Buttenheim, H Alderman, J Friedman. in the report suggests that targeted scholarships may be an World Bank Policy Research Working Paper, 2011. effective tool to address some forms of drop-out in the Lao 19 In a randomized controlled trial in the Dominican Republic, Jensen context, especially for girls. There are clear signs that the (2010) shows that providing information about the returns to opportunity cost of attending school in some contexts is education to secondary students increases their chance of graduating and continuing their studies. considered high by some parents, and this is confirmed by 20 See Glewwe and Muldharidaran (2015) for a comprehensive review analysis of household economic activity undertaken by of evidence of interventions to improve quality. Loyalka et al. (2013) for children in these populations. the China program that failed to show impact. Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 39 c. Early childhood education and care. Access to early child- hood education can facilitate on-time transitions to basic education. In addition, when implemented well, early childhood education programs improve school readiness for children across various dimensions (not only cognitive), facilitating the successful transition to basic education. However, the effectiveness of such programs depends on parents actually wanting to take their children to early child- hood centers.21 To the extent that cultural perceptions about the quality or relevance of education are a limitation on parents enrolling their children in primary school, they may a. Continuing with the use of community based construction also be a limitation for early childhood education. However, to complete schools, including pre-primary. Community the importance of starting education early cannot be based construction (CBC) has been used extensively in overstated. Increasing access to early childhood education projects supported by the World Bank (and other develop- is crucial to prevent late enrollment and drop-out before ment partners) and has been shown to be a cost-effective completing primary education. The quality early childhood method of building which, with the right supervision of imple- education is also extremely important;monitoring and to mentation, can lead to the same or better quality of construc- ensure the quality of service provision is vital. tion. The lessons learned from the implementation of commu- nity based construction in Lao PDR highlight the fundamental d. Introduce, pilot and evaluate a flexible curriculum and role of having well qualified district engineers who support methodologies of instruction. In addition to improving and supervise the construction, and who have adequate quality, introducing a flexible curriculum that can be resources to perform this role. Without these structures in delivered in short modules can help improve the relevance place, CBC runs the risk of resulting in low quality or unsafe (andfacilitate instruction) for children who may have to miss construction, so CBC may not be appropriate in these school or and therefore fall behind frequently. circumstances. 77. Improving Supply. The availability of schools, while not b. Ensure availability of adequate learning materials and sufficient by itself, is a necessary condition for children to textbooks. However, more important than their availability attend school. About 30 to 40 percent of children who in the school is their effective use. Pedagogical advisors in drop-out early live in a village that does not offer the grade districts should support principals and teachers in the use of they were supposed to attend. Thus, increasing availability materials and classroom practices. At the moment, pedagogi- of preschools, completing schools and facilitating access to cal advisors lack the resources and sometimes the skills to lower secondary school continue to be clear priorities in Lao carry out this support. PDR. Inclusion policies for disabled children and other groups, and school safety, also come out clearly in 21 the analyses as important for certain groups of the Low demand from parents was one of the constraints in the program population. implemented in Cambodia (Filmer et al). . 40 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR c. Scholarships for teaching in rural areas, in which secondary Filmer, Deon & Schady, Norbert, 2009. "Are there diminishing returns to transfer size in conditional cash transfers?" Policy Research Working school students are provided a scholarship to become teach- Paper Series 4999, The World Bank. ers if they agree to return to their village to teach, have proven Fuller, B. and P. Clarke (1994), “Raising school effects while ignoring culture? effective in getting teachers to remote areas. While an evalua- Local conditions and the influence of classroom tools, rules and tion of the program should be conducted before scaling it up pedagogy.” Review of Educational Research, 64, 119-157; significantly, there are indications from qualitative evaluations Glewwe, P., E.A. Hanushek, S. Humpage and R. Ravina (2015), “School that this program is more effective than other alternatives resources and educational outcomes in developing countries: A review of the literature from 1990 to 2010”, (like bonuses) in bringing and keeping teachers in remote areas. Glewwe, P. (2014), Education policy in developing countries. Chicago: The University of Chicago Press; d. Safety. While not very prevalent in household responses Handa, S. (2002). Raising primary school enrolment in developing countries: The relative importance of supply and demand. Journal of Development as a reason for dropping out, there seems to be concerns Economics, 69(1), 103-128. about safety in school in some contexts. The importance of Jensen, Robert. "The (perceived) returns to education and the demand for this for reasons beyond educational outcomes, and the lack schooling." Quarterly Journal of Economics 125.2 (2010). of specific information in household surveys, means this is an Lawson, Ty M. Impact of School Feeding Programs on Educational, issue to be evaluated farther. Nutritional, and Agricultural Development Goals: A Systematic Review of Literature. Diss. Michigan State University, 2012. Loyalka, Prashant, et al. "Can information and counseling help students REFERENCES from poor rural areas go to high school? Evidence from China." Journal of Comparative Economics 41.4 (2013): 1012-1025. Marshall, J.H. (2011). School quality signals and attendance in rural Guate Alderman, Harold, and Donald Bundy. "School feeding programs and mala. Economics of Education Review, 30(6), 1445-1455. development: are we framing the question correctly? " The World Bank McEwan, P. J. "Evaluating rural education reform: The case of Colombia's Research Observer (2011): lkr005. Escuela Nueva program." La Educación 132.133 (1999): 35. Barrera-Osorio, Felipe & Filmer, Deon, 2013. "Incentivizing schooling for Montenegro, Claudio E., and Harry Anthony Patrinos. "Returns to Schooling learning: evidence on the impact of alternative targeting approaches," Policy around the World." Background Paper for the World Development Research Working Paper Series 6541, The World Bank. Report(2013): 8258024-132095074719. Buttenheim, Alison M., Harold Alderman, and Jed Friedman. "Impact National Assessment of Student Learning Outcome (ASLO III) Grade 3, evaluation of school feeding programs in Lao PDR." World Bank Policy Research Institute for Educational Sciences (RIES), Ministry of Education and Sports, June 2014 Research Working Paper 5518 (2011). Schultz, T.W. (1963). Human capital. Chicago: University of Chicago Press; De Hoyos Navarro, Rafael E. “The Heterogeneous effect of information on and Becker, G.S. (1967). Human capital and the personal distribution of student performance: evidence from a randomized control trial in Mexico” income: An analytical approach. WPS7422, September 28, 2015 Woytinsky Lecture no. 1, Ann Arbor: University of Michigan. Early Grade Reading Assessment, Research Institute for Educational Sciences (RIES), Ministry of Education and Sports, 2013 World Bank (2013) Skills for Quality Jobs and Development in Lao PDR : A Technical Assessment of the Current Context Filmer, Deon & Schady, Norbert, 2006. "Getting girls into school: evidence from a scholarship program in Cambodia," Policy Research Working World Bank 2014. Lao PDR Development Report: Expanding Productive Employment for Broad-Based Growth, World Bank. Paper Series 3910, The World Bank. Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 41 ANNEX A – ROOT CAUSES RESEARCH DATABASE SUMMARY DATABASE COVERAGE COMMENTS National Assessment of 5,860 students in Grade 3, 444 Nationally representative sample accompanied by detailed questionnaires for Student Learning schools, Lao and Mathematics students, teachers, school director; summary of multivariate statistical analysis (ASLO), 2012 included in main report, our analysis builds on this previous work with a more detailed summary of variables that are associated with Lao language in G3. The main advantages of ASLO (relative to EGRA) is that it is nationally representative. However, it is only for one grade. Early Grade Reading 3,409 students in G3-G4-G5 Schools were drawn from 6 provinces and 3 regions, so not nationally representative Assessment (EGRA), (beginning of school year), although there are language minorities included; includes detailed student and 2012 72 schools teacher questionnaire; higher grades were selected since the test was conducted at the beginning of the school year, so these samples correspond (in theory) to end of year G2-G3-G4. The main advantage with EGRA is the focus on early age literacy (reading, dictation, oral fluency). However, the sample is not strictly representative at the national level. Education Manage- Various years EMIS data were obtained to get totals for enrollment and grade repetition; dropout ment Information rates can be calculated using a cohort change method that compares enrollments System (EMIS) over time based on repeaters and new entrants; also, some of the EMIS figures reported in this study come from the UNESCO education indicator website that official EMIS data feed into. Lao PDR Expenditure 43,000 persons, 8,200 Multi-purpose survey with focus on consumption and expenditure, less information and Consumption households on child’s schooling and work activities (compared with LFS); surveys conducted Survey (LECS),2012-13 between April-December 2012, and January-March 2013. The main advantage with LECS is that the data are more recent. Labor Force Survey 58,000 persons, 10,800 Detailed household survey that includes more information on children than in LECS, (LFS), 2010 house-holds, 540 villages including child labor (inside the household and outside) as well as variables for age at school entry and age at school dropout; data collected during a short time period, mainly October-December 2010. The tradeoff with the LFS data is between variable coverage (more detail) and timing (older data). DATA SOURCE ADVANTAGES CAVEATS EMIS Census of schools School level, not individual Collected directly from principals No socioeconomic information Well developed system but still insufficient quality control at school level LECS V Detailed information about household Sample based Includes education history of the household Spanning over two school years Answered directly by parents Parents perspective, not child perspective LFS Includes employment consumption, etc Relatively old data EGRA Detailed reading outcomes diagnostic Sample based Only focused on reading Not nationally representative ASLO Detailed learning outcomes diagnostic Relatively old data (new one planed) Three subjects Sample based 42 | Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR v. Primary School Curriculum: Primary curriculum is long and of poor quality, and does not reflect real situation in local areas. There is also lack of qualified/quality teachers; ANNEX B – SUMMARY OF and students often do not achieve the expected learning WORKSHOPS ON THE CAUSES competency at the end of each academic level. FOR DROP-OUT vi. Teaching-Learning: The problems raised in this area are poor teaching and learning quality and environment; low This annex provides a summary from MoES workshops students’ reading and analytical skills. conducted on the topic of Grade 1 Dropout and Repetition. vii. Supervision: There is lack of pedagogical support to A series of four regional consultations were held in Vientiane teachers due to limited budget and capacity at local level. Province, Savannakhet, Oudomxay and Xieng Khuang in March 2015. Participants included 160 officials from 80 viii. Management: School principals do not receive sufficient Districts Education and Sports Bureau (DESB) and representa- trainings and do not work effectively with VEDC and local tives from all Provincial Education and Sports Services (PESS) authority in school management. Some principals who responsible for Primary Education and Monitoring and Evalua- also teach in classes do not have enough time to perform tion. The findings were expected to serve as inputs for the management tasks. development of the Education Sector Development Plan (ESDP), particularly for the design of sub-sector policy Taken together, the results from the workshops provide a and interventions. The workshop results describe eight key very useful complement to the analysis in previous sections, areas contributing to drop out and grade repetition, as follows: including the stated reasons for dropout in the previous sub-section. Regarding root causes, the responses in the i. Parents/Guardian and Community: Key challenges in this workshops are generally consistent with what households area include parents/guardians’ low perceived value of were stating in the LFS and LECS, and provide some more education, lack of involvement of VEDC in school detailed explanations. For example, a number of specific management, and lack of parents’ support to send school quality and management factors are cited in the work- disabled children to school. shops, although it is not clear if these are relevant to grade repetition only, or also to dropout. Regardless, the workshop ii. Children: Key constraints include the needs for children responses provide some more specific clues about the under- to help with agricultural work, poverty, traditional customs lying causes of low quality that may in turn result in higher and beliefs that prevent children to attend schools rates of grade repetition and dropout. particularly for girls, lack of school readiness and early childhood development programs, and remoteness. For family background factors, the workshops also provided iii. School Environment and Facilities: This largely relates to some more detail, with references to agricultural work calen- the quality of schools and the quality of teaching and dars that require children to move around, ethnicity and learning. Key issues include insufficient operating budget concerns about the value of education and, presumably, the (insufficient school block grant provided to schools), need to invest time and resources in the child’s schooling. ineffective teacher deployment and lack of school facilities to facilitate teaching and learning. Finally, the workshop responses are consistent with the LFS-LECS data sets in terms of some factors that do not iv. Curriculum of Teacher Education and Teacher Training: appear to be very important. Access is infrequently refer- This includes lack of relevance of teacher education enced, including rainy season inaccessibility.Between the curriculum and the actual teaching in schools, ineffective MOES workshops, and the LFS-LECS data, a more compre- teacher development program, and lack of resources in hensive picture about the root causes of early grade dropout Teacher Training Institutes. in Lao PDR has emerged. Reducing Early Grade Dropout and Low Learning Achievement in Lao PDR | 43 The World Bank Country Office Education Global Practice Xieng Ngeun Village, Chao Fa Ngum Road The World Bank P.O. Box. UN345, Chanthabouly District 1818 H Street, NW Vientiane, Lao PDR Washington D.C. 20433, USA Tel: (856-21) 266200 Tel: (202) 473 1000 Fax: (856-21) 266299 Fax: (202) 477 6391 www.worldbank.org/lao www.worldbank.org www.facebook.com/worldbanklaos www.facebook.com/worldbank