THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY CHATA MALE & QUENTIN WODON DECEMBER 2017 THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY CHATA MALE & QUENTIN WODON BACKGROUND TO THIS SERIES More than one billion people may experience some form Development Goals are related to disabilities, including of disability. Of those, up to one in five may experience with respect to inclusive education. However, at the significant disabilities. Individuals with disabilities have country level, and especially in low and middle income on average worse socioeconomic outcomes than those countries, resources are often lacking to effectively without disabilities. They often have poorer health, lower promote inclusive education. levels of employment and earnings, and higher poverty rates. In developing countries, the prevalence of disability This note is part of a series on The Price of Exclusion: and its impacts on a wide range of development outcomes Disability and Education prepared as part of broader work are typically larger. program on out-of-school children with funding from the Global Partnership for Education. The series documents Children with disabilities are especially at a disadvantage gaps in education outcomes between children with in terms of school enrollment, educational attainment, and without disabilities. It also showcases examples of and learning. Many children with disabilities never enroll programs and policies and lessons from the literature on in school, or drop out prematurely. They also often learn how to improve inclusion in education systems. less while in school. Ensuring that these children have the same opportunities as other children is a challenge, but also an opportunity, in that inclusive education may bring benefits to all children, and not only those with disabilities. Awareness of the need for inclusive education systems is increasing. The United Nations Convention on the Rights of Persons with Disabilities has been ratified by 175 countries. It calls for full integration of persons with disabilities in societies. Several targets in the Sustainable 1 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | DECEMBER 2017 KEY MESSAGES the disability gaps for literacy also grew over time, reaching 16.2 points for boys and 15.5 points for girls. This note provides an analysis of gaps in educational opportunities for children with disabilities. It also measures • When looking through regression analysis at the the impact at the margin of exclusion related to various types marginal effects of exclusion associated with disabilities, of disabilities on education outcomes for children. Four main findings are similar to the results from simple statistical outcomes are considered: whether children ever enroll in comparisons. Across the 19 countries, the average school, whether they complete their primary education, reductions at the margin for children with disabilities in whether they complete their secondary education, and the probabilities of ever enrolling in school, completing whether they are literate. The analysis is implemented primary schooling, completing secondary schooling, using the most recent census data available for a total of 19 and being literate are estimated at 11.9 points, 16.8 countries. Key findings are as follows: points, 13.9 points, and 16.4 points respectively. • The Sustainable Development Goals (SDGs) call for • The marginal effects on education outcomes of exclusion ensuring inclusive and quality education for all and related to disabilities are often larger than the effects of promoting lifelong learning (Goal 4). The SDGs ex- other child or household characteristics. For example, plicitly mention equal access to all levels of education the marginal effect of a disability is often larger than and vocational training for the vulnerable, including that of the quintile of wealth of the child’s household. persons with disabilities. Unfortunately, gaps in ed- ucation outcomes between children with and with- • Overall, the analysis demonstrates that children with out disabilities have been increasing over time. disabilities are being left behind by global efforts to improve education opportunities for all. The rising • For children without disabilities, completion rates at gaps between children with and without disabilities the primary level have increased substantially over the in developing countries call for stronger policies and last few decades. Smaller gains have been observed for interventions to achieve the target of inclusive education children with disabilities. As a result, the gap in primary adopted under the Sustainable Development Goals. completion rates between children with and without disabilities has increased over time from a few percentage points a few decades ago to 17.6 points for boys and PRIMARY EDUCATION 15.4 points for girls in the latest available census data. COMPLETION • Many children with disabilities are never enrolled in school. Among children aged 11, the likelihood of This note measures gaps in educational outcomes between having ever enrolled in school was 13 percentage points children with and without disabilities using census data for lower for children with disabilities versus children 19 countries. Separate notes are being prepared with more without disabilities at the time of the latest available detailed results for each country. While census data are census data. As for primary education, the disability known to underestimate the extent of disabilities and may gap in ever enrolling has increased over time. thereby capture for the most part severe disabilities, they are nevertheless useful – given their large size – to look at trends • Large gains in secondary completion rates have also over time in educational outcomes for children with and been achieved for boys and girls without disabilities, without disabilities. Census data are also useful to measure but gains are again smaller for children with disabilities, the impact at the margin of exclusion related to disabilities leading to disability gaps in completion rates of on education outcomes. Four main education outcomes 14.5 points for boys and 10.4 points for girls. are considered: (1) whether children ever enroll in school; (2) whether they complete their primary education; (3) • The last indicator considered for the analysis of trends whether they complete their secondary education; and (4) over time is literacy. In part, as a result of differentiated whether they are literate. This section considers first primary trends in educational attainment by disability status, education completion (see Box 1 on methodology). DECEMBER 2017 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | 2 Disability gaps in educational the time of the last publicly available census data for each country (three-year moving averages are used to attainment for older individuals better capture underlying trends). For children without could be underestimated. This disabilities, completion rates at the primary level increased is because older individuals may substantially. The gain is at 20.8 percentage points for have suffered from a disability boys and 34.5 percentage points for girls over the 42 after leaving school. This would years separating the youngest and oldest groups. Girls have essentially caught up with boys in terms of primary tend to reduce the measures of completion. Smaller gains are observed for children with disability gaps observed for older disabilities, at 7.9 percentage points for boys and a much individuals. larger 23.8 points for girls. Girls with disabilities have caught up with boys with disabilities at that level, but for both boys and girls, there seems to be a plateau in terms of Figure 1 provides primary completion rates for individuals the completion rates for primary school. This is especially aged 16 to 58. Completion rates for children aged 12-15 visible with the trend for boys with disabilities, with smaller are not shown because they tend to be too low versus gains over time and with girls also potentially affected by expected lifetime completion rates. This is because in the same plateau according to data for the most recent many countries, some children enter primary school late years. or repeat grades, which leads them to complete primary school well beyond the normal completion age of 12 in most As a result of these trends, the absolute gap in primary countries. The census data for the various countries were completion rates between children with and without for the most part implemented between 2008 and 2011. disabilities has increased dramatically over time. This This implies that changes that may have taken place in the is visualized in Figure 2. Several decades ago, as fewer last few years are not accounted for, but it is unlikely that children had the opportunity to complete primary school, such changes would reverse the long-term trends observed differences in completion rates by disability status were at across countries in the census data. only four to six percentage points for boys and girls alike. At the time of the latest available census data for the countries Four groups are considered: boys with no disability, girls included in this analysis, recent gaps are at 17.6 points for with no disability, boys with one or more disabilities, boys and 15.4 points for girls. In short, completion rates and girls with one or more disabilities. The statistics in have improved for children with disabilities, but much Figure 1 are average completion rates for individuals more slowly than has been the case for children without of the corresponding age across the 19 countries at disabilities, so gaps have widened. For children without disabilities, completion rates at the primary level have increased substantially. Smaller gains have been observed for children with disabilities. As a result, the absolute gap in primary completion rates between children with and without disabilities has increased dramatically over time. 3 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | DECEMBER 2017 BOX 1: ANALYZING DISABILITY AND EDUCATION OUTCOMES WITH CENSUS DATA The use of census data for work on disability and education has both advantages and limits. The main advantage is the fact that the number of observations in the dataset is large even when only a subsample is available for the analysis. This means that the number of individuals with disabilities in the dataset is sufficient to measure trends over time in education outcomes and the impact at the margin (controlling for other factors affecting education outcomes) of exclusion related to disabilities. Marginal impacts can be estimated for children with disabilities, as well as by type of disability. For example, the impact of exclusion related to psychosocial or cognitive disabilities can be compared to that of visual impairment, or hearing impairment (while this is not done in this note, it is done in separate notes in this series). By contrast, in most cases, the sample size of household surveys with questions on disabilities is insufficient to consider trends over time or to disaggregate the analysis by type of disability. There are also limitations when using census data for the analysis of the relationships between disability and education outcomes. First, the data tend to underestimate the prevalence of disability. Second, despite efforts to improve questionnaire, the only information that is typically available relates to whether a child has a disability or not, and not whether the disability is severe or mild. These two limitations are related. To identify in a census individuals with disabilities, it is more useful to ask questions about the difficulties that individuals may have in performing various tasks and activities. The Washington Group on disability has proposed six such questions related to core functional domains: seeing, hearing, learning, walking, cognition (remembering or concentrating), self-care (washing all over or dressing), and finally communicating. For each question, four responses are suggested – the individual has no difficulty, some difficulty, a lot of difficulty, or no ability at all to perform the task. These questions help detect disabilities that may be less severe. By contrast, when a single question is asked in a census, the result is typically a substantially lower rate of disability identifying only those with the most severe disabilities in the household as having a disability. The limitations of census data impose limits to the analysis that can be performed. Readers should be aware that in practice, only severe disabilities tend to appear in census datasets. Another limitation of the analysis is that in most countries censuses are implemented only once every ten years. Therefore, the data may not always be fully representative of the latest developments in education systems. A third limit is that the timing of the disability is not observed. Especially for older individuals, the disability may have been observed after the individual has left school. In that case, the comparisons of educational attainment for individuals with and without disabilities (through statistics or regression analysis) may be less precise in identifying the role of disabilities in affecting educational attainment for older individuals. This is one of the reasons why the regression analysis is performed on comparatively younger age groups than the statistical comparisons which are provided for all age groups. Because censuses tend to identify severe disabilities that are often observed at or soon after birth, the risk of bias may however not be too large. Despite these limitations, and given the lack of alternatives, census data remain a useful source of data for measuring for multiple countries and in a relatively comparable way the impact of exclusion related to disabilities on education outcomes, especially over time, and the impact of exclusion by type of disability. The analysis in this note is carried with census data from 19 countries: Bangladesh 2011, Burkina Faso 2006, Cambodia 2008, Costa Rica 2011, Dominican Republic 2010, Ethiopia 2007, Ghana 2010, Indonesia 2010, Kenya 2009, Liberia 2008, Mali 2009, Malawi 2008, Mexico 2010, Mozambique 2007, Peru 2007, South Africa 2011, South Sudan 2008, Vietnam 2009, and Zambia 2010. This note reports average trends in education outcomes over time across the 19 countries, as well as the marginal impacts of exclusion related to disabilities for these education outcomes. More detailed analysis at the country level is available in separate notes in this series. DECEMBER 2017 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | 4 FIGURE 1: PRIMARY COMPLETION RATES BY AGE AND GROUP (%) COMPLETION RATES BY GROUP (%) 70 65 60 55 50 45 40 35 30 25 20 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 AGE MEN, NO DISABILIT Y MEN, WITH DISABILIT Y WOMEN, NO DISABILIT Y WOMEN, WITH DISABILIT Y FIGURE 2: DISABILITY GAP FOR PRIMARY COMPLETION (%) 20 18 DISABILITY GAP BY SEX (%) 16 14 12 10 8 6 4 2 0 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 MEN WOMEN AGE ENROLLING Patterns observed in Figures 3 and 4 for ever enrolling in school are similar to those observed in Figures 1 and 2 for IN SCHOOL primary completion. For children without disabilities, the likelihood to ever enroll in school increased by 20.7 points for The gaps in primary school completion between children boys and 35.9 points for girls over the 47 years separating with and without disabilities may be due in part to the fact the youngest and oldest age groups. Girls caught up with that children with disabilities are at a higher risk of dropping boys, as is the case for primary completion. Smaller gains out before completing the cycle than children without are observed over the period for children with disabilities, at disabilities. However, many children with disabilities never 9.8 points for boys and 25.9 points for girls. Even if girls with even get to enroll in school. Indeed, the disadvantages disabilities have caught up with boys with disabilities, absolute faced by these children start in their early childhood years gaps in the likelihood of ever enrolling in school between including for enrollment at the primary level. Figures 3 and children with and without disabilities have grown over time. 4 provide data on the share of children who ever enrolled This is shown in Figure 4. At the time of the latest available in school. Statistics are computed for individuals aged 12 to census data, the absolute gaps associated with disabilities 58. Statistics for children aged 6-10 are not shown because were at 13.2 points for boys and 12.7 points for girls. some of these children could still enroll at a later age. This is less likely for children aged 11 and over if they have not yet enrolled by then. 5 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | DECEMBER 2017 Many children with disabilities are never enrolled in school. At the time of the latest census data, among children aged 11, the likelihood of having ever enrolled in school was 13 percentage points lower for children with disabilities versus children without disabilities. The disability gap has increased over time. FIGURE 3: SHARES OF CHILDREN EVER ENROLLED BY AGE AND GROUP (%) 90 85 80 SHARES BY GROUP (%) 75 70 65 60 55 50 45 40 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41 40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 AGE MEN, NO DISABILIT Y MEN, WITH DISABILIT Y WOMEN, NO DISABILIT Y WOMEN, WITH DISABILIT Y FIGURE 4: DISABILITY GAP FOR SHARE OF CHILDREN EVER ENROLLED (%) 16 14 DISABILITY GAP BY SEX 12 10 8 6 4 2 0 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 MEN WOMEN AGE DECEMBER 2017 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | 6 SECONDARY EDUCATION of progress but still lag slightly behind boys, probably in part due to the risks of child marriage and early childbearing in COMPLETION many countries, both of which lead girls to drop out of school at the secondary level. The gains for children with disabilities The analysis is repeated for secondary education completion were once again smaller, at 10.6 points for boys and 20.8 in Figures 5 and 6. Once again, similar patterns emerge, points for girls. While girls with disabilities have caught up albeit with lower completion rates than for primary with boys with disabilities, gaps in secondary completion rates education. Estimates are provided for individuals aged 20 have widened between children with and without disabilities. and above to allow for a few additional years versus the They reached 14.5 points for boys and 10.4 points for girls typical age of on-time completion of 18 in most countries. aged 20 in the latest available census data. As for other For children without disabilities, completion rates at the indicators, children with disabilities did not benefit as much as secondary level increased by 20.6 percentage points for other children from the large gains in educational attainment boys and 27.3 percentage points for girls over the 38 years achieved by countries. separating the youngest and oldest groups. Girls made a lot Large gains for boys and girls without disabilities were also achieved in secondary completion rates, but gains were once again smaller for those with disabilities. This has led to disability gaps in completion rates of 14.5 points for boys and 10.4 points for girls. FIGURE 5: SECONDARY COMPLETION RATES BY AGE AND GROUP (%) 50 COMPLETION RATES BY GROUP (%) 45 40 35 30 25 20 15 10 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 AGE MEN, NO DISABILIT Y MEN, WITH DISABILIT Y WOMEN, NO DISABILIT Y WOMEN, WITH DISABILIT Y 7 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | DECEMBER 2017 FIGURE 6: DISABILITY GAP FOR SECONDARY COMPLETION (%) 16 14 DISABILITY GAP BY SEX (%) 12 10 8 6 4 2 0 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 MEN WOMEN AGE LITERACY The last indicator considered for the analysis of trends over time is literacy, as declared subjectively by census The last indicator considered respondents for their household members. Figures 7 and for the analysis of trends over 8 provide the trends in literacy over time by age groups, time is literacy. In part as a considering individuals from 12 to 58 years of age. For result of differentiated trends children without disabilities, the likelihood of literacy in educational attainment by increased by 14.9 points for boys and 31.2 points for girls over the 47 years separating the youngest and oldest age disability status, the disability groups. As expected, girls caught up with boys. The gains gaps for literacy also grew over for children with disabilities were small for boys at 1.3 points, time substantially, reaching 16.2 while they reached 19.4 points for girls. As shown in Figure points for boys and 15.5 points 8, the disability gaps grew over time substantially, reaching 16.2 points for boys and 15.5 points for girls. for girls. DECEMBER 2017 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | 8 FIGURE 7: LITERACY RATES BY AGE AND GROUP (%) 85 LITERACY RATE BY GROUP (%) 80 75 70 65 60 55 50 45 40 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 AGE MEN, NO DISABILIT Y MEN, WITH DISABILIT Y WOMEN, NO DISABILIT Y WOMEN, WITH DISABILIT Y FIGURE 8: DISABILITY GAP FOR LITERACY (%) 18 DISABILITY GAP IN LITERACY RATE 16 14 12 10 8 6 4 2 0 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 MEN WOMEN AGE SUMMARY AND Are the current gaps the result of exclusion associated with MARGINAL IMPACTS disabilities, or do they result from other characteristics of children that could be correlated with disabilities? To assess the The data presented in the previous sections suggest large likely impact of exclusion related to disability on educational and growing gaps in educational attainment and literacy attainment and literacy, regression analysis is needed to between children with and without disabilities over time. Table control for other factors that may affect education outcomes. 1 summarizes the key estimates for the youngest appropriate The term “impact” is used for simplicity (see Box 2 on what is age cohorts for each of the four indicators. The disability gaps meant by “impact”). are large, but this does not necessarily mean that they are due to disabilities as opposed to other characteristics that could be The analysis of marginal impacts considers the same four correlated with having one or more disabilities. education outcomes, but with slightly different age groups 9 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | DECEMBER 2017 Table 1: Disability Gaps for the Most Recent Appropriate Age Cohorts (%) Ever Enrolled Primary Completed Secondary Completed Literacy (11 years old) (16 years old) (20 years old) (12 years old) Men Women Men Women Men Women Men Women No disability 85.33 84.64 64.69 63.57 46.93 42.88 75.16 75.09 Disability 72.18 71.94 47.09 48.13 32.47 32.44 58.94 60.04 Disability gap 13.15 12.70 17.60 15.44 14.46 10.43 16.22 15.05 Source: Authors’ estimation using census data for 19 countries. mostly for sample size reasons: (1) whether a child ever in which the child resides, the mother tongue of the child, the enrolled in school (the sample for the regression analysis quintile of wealth of the households to which the child belongs, consists of children ages 6 to 11 to account for conditions at the the religion of the child, the size of the household in which time of the implementation of the censuses); (2) whether a the child resides, the sex of the household head, a number of child completes primary education (sample of children 15 to 18 characteristics for the household heal (age according to 10 years old); (3) whether a child completes secondary education years intervals, education level, and type of work), the same (sample of youth ages 19-22); and (4) whether a child is characteristics for the spouse of the household head, and the considered by the parents as literate (sample of children ages leave-out-mean of the dependent variable. This leave-out- 15-18). For each regression, the sample of children included mean variable is computed among all other children in the area is a bit older than the normal threshold to complete a level of where a child lives and is meant to capture local conditions schooling or be literate to account for children who may start that affect education outcomes for children in the more school late or repeat grades. disaggregated area where a child lives. While census data have limits in terms of the variables that can Table 2 are almost always statistically significant as well as be used as controls, a number of important variables known large. Across the 19 countries, the average reductions at to affect educational outcomes are available in the data. The the margin for children with disabilities in the probabilities regression analysis includes the following variables (with minor of ever enrolling in school, completing primary schooling, differences between countries) as controls: the sex of the child, completing secondary schooling, and being literate are whether the child has a birth certificate, whether the child estimated at 11.9 points, 16.8 points, 13.9 points, and 16.4 has a disability and the type of disability observed (in addition, points, respectively, in comparison to children with no an interaction effect is included to assess if a disability has a disability. These estimates are of a similar order of magnitude differential impact for boys or girls), whether the child is an to the statistical measures mentioned in the previous section, orphan on the side of the mother, the father, or both; whether suggesting that most of the differences between children the child resides in an urban or rural area, the geographic area with and without disabilities are indeed associated with the BOX 2: WHAT DO WE MEAN BY “IMPACTS” OF EXCLUSION RELATED TO DISABILITIES? The aim of this section is to estimate the impact of exclusion associated with a disability on education outcomes for children with disabilities in comparison to children without a disability but with otherwise similar characteristics. The term “impact” is used for simplicity, but one must be careful about not necessarily inferring causality. Estimates of impacts are obtained through regression analysis to control for other variables that may affect education outcomes. What is measured are statistical associations, and not necessarily impacts as could be observed for example with randomized control trials. Since a disability cannot be randomized, we must rely on regression analysis to estimate likely impacts, but there is always a risk of bias in the measures of the impacts on outcomes of exclusion associated with a disability. DECEMBER 2017 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | 10 disability status of the children as opposed to other variables that could be correlated with that status. Although this is Across the 19 countries, the not discussed here, as noted in other briefs in this series, average reductions at the margin there are differences in the impacts associated with different for children with disabilities types of disabilities. For example, children with cognitive, in the probabilities of ever psychosocial or multiple disabilities often fare worse than enrolling in school, completing children with physical disabilities. primary schooling, completing To provide perspective, it is useful to compare the marginal secondary schooling, and being impacts on education outcomes of exclusion related to literate are estimated at 11.9 disabilities with the impact of other factors affecting those points, 16.8 points, 13.9 points, outcomes. In many countries, girls continue to have lower educational outcomes than boys, but the magnitude of and 16.4 points respectively. the gender gaps is much smaller than for disability gaps, as already suggested by the analysis in previous sections. Being an orphan is also associated with a lower likelihood of completing various levels of schooling as well as being The marginal effects on literate, but with smaller effects than exclusion related to education outcomes of disability. There are also wealth effects at work, with children exclusion related to disability from better off households more likely to do well. When are often larger than the effect comparing children from the lowest quintile of wealth to those from the top quintile of wealth, marginal effects can of other child or household be large, but again, socio-economic differences tend to characteristics. For example, the have smaller effects at the margin that exclusion related marginal effect of a disability to disability. Finally, there are also marginal effects on is often larger than that of education outcomes associated with the education level of the quintile of wealth of the the household head and the location of the household, but these impacts are smaller again than those associated with households in which a child disability even if differences between urban and rural areas resides. can be substantial. 11 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | DECEMBER 2017 Table 2: Impact of Exclusion Related to Disability on Educational Attainment and Literacy Ever Enrolled Primary Completed Secondary Completed Literacy (6-11 years) (15-18 years) (19-22 years) (15-18 years) Bangladesh -34.3 -39.8 -22.9 -41.8 Burkina Faso -13.9 -18.7 -3.8 -7.8 Cambodia -18.3 -28.8 -11.3 -18.9 Costa Rica NS -7.5 -8.4 -5.1 Dominican Republic -8.5 -20.4 -20.4 -10.1 Ethiopia -10.1 -6.7 NS -21.8 Ghana -7.7 -13.3 -18.2 -6.2 Indonesia -27.5 -42.6 -46.6 -32.3 Kenya -9.9 -17.3 -15.9 -17.9 Liberia -7.8 -11.1 -8.2 -9.3 Malawi -5.8 -6.4 -4.3 -8.1 Mali -12.9 -7.8 -1.9 -10.0 Mexico -14.9 -27 -35.7 -22.2 Mozambique -13.8 -7 -1.2 -25.0 Peru -2.3 -3.6 -5.6 -2.2 South Africa -1.7 -1.6 -3.7 -23.5 South Sudan NS NS NS -5.2 Vietnam -27.8 -37.8 -38.4 -31.3 Zambia -6.9 -21.4 -16 -13.7 Average Impacts -11.9 -16.8 -13.9 -16.4 Source: Authors’ estimation using census data. Note: NS = not statistically significant. The average impact includes coefficients that are not statistically significant. CONCLUSION This note has provided a simple analysis of education While progress continues to be achieved to reduce outcomes for children with and without disabilities gender and wealth gaps in education, progress is much using census data for 19 countries. Children with too slow for children with disabilities. Ensuring that disabilities have been shown to be at a substantial children with disabilities have opportunities to go to disadvantage for ever enrolling in school, completing school and learn should be a top priority to end the primary or secondary education, or being literate. The persistent crisis of so many of these children remaining gaps between children with and without disabilities out-of-school or dropping out prematurely, with often have been increasing over time. In essence, children dramatic consequences for their prospects later in life. with disabilities are being left behind by global efforts to improve education opportunities for all children. The Sustainable Development Goals call for building and upgrading education facilities that are child, disability and gender sensitive and also provide safe, inclusive and effective learning environments. DECEMBER 2017 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | 12 Recommended citation for this note: Male, C., and Q. Wodon (2017). Disability Gaps in Educational Attainment and Literacy, The Price of Exclusion: Disability and Education Notes Series, Washington, DC: The World Bank. The authors are with the World Bank Group. This series of notes was launched by the Education Global Practices at the World Bank in collaboration with the Global Practice Social Urban, Rural, and Resilience which houses the World Bank’s disability team. Comments on this note from Hanna Katriina Alasuutari and Charlotte Vuyiswa McClain- Nhlapo are gratefully acknowledged. The series is part of a broader work program that benefitted from the support of the Global Partnership for Education. The findings, interpretations, and conclusions expressed in this note are entirely those of the authors and should not be attributed in any manner to the World Bank, its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this work. Information and illustrations contained in this report may be freely reproduced, published or otherwise used for noncommercial purposes without permission from the World Bank. However, the World Bank requests that the original study be cited as the source. © 2017 The World Bank, Washington, DC 20433. 13 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | DECEMBER 2017 DECEMBER 2017 | THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY | 14 THE PRICE OF EXCLUSION: DISABILITY AND EDUCATION DISABILITY GAPS IN EDUCATIONAL ATTAINMENT AND LITERACY