Measuring Gender Equality in Education: Lessons from Trends in 43 Countries

Abstract

Gender inequalities in primary school enrollment have declined in recent decades in low-income countries, leading some observers to assume that disparities in educational attainment are now limited to secondary and higher education. Amid declarations of victory on school enrollment, policy and programmatic focus has also shifted from attainment to school quality and learning. Yet millions of young people are still not enrolling in primary school (UNESCO 2016a, 2016b).1 Sustainable Development Goals (SDGs) 4 and 5 focus on quality education and reducing gender inequality, respectively.2 Progress toward these goals requires a more complete understanding of current global patterns in school enrollment, grade attainment, and learning. Gender gaps in primary school enrollment narrowed substantially in many low-income countries during the post-colonial period, beginning in the 1960s (Lloyd, Kaufman, and Hewett 2000; Wils and Goujon 1998). In examining trends in enrollment and attainment in 24 sub-Saharan African countries from the late 1960s to the late 1990s, Hewett and Lloyd (2005) found that only 45 percent of girls and 66 percent of boys had ever attended school in the late 1960s. By the late 1990s these percentages had increased for both groups, but more so for girls: 73 percent of girls and 78 percent of boys had ever attended school. Over the same period, primary school completion levels increased from 46 to 58 percent for boys and from 26 to 53 percent for girls. By the 1980s and 1990s, however, the pace of progress slowed for both groups.3 These trends were consistent with national policies during that period, when many post-colonial governments prioritized expansion of educational infrastructures alongside strong economic growth. The subsequent combination of economic downturns, rapid population growth, and structural adjustment programs resulted in reduced investments in education (Lloyd, Kaufman, and Hewett 2000). Wils and Goujon (1998), who observed similar patterns of progress in expanding primary school enrollment between 1960 and 1990,4 argued that an overall “education transition” began to take shape during this period in some countries, in which primary enrollment increased for both males and females while gender gaps declined, followed by similar changes at the secondary school level. In 1990, in light of progress in increasing primary school enrollment and closing gender gaps, the Education for All (EFA) movement, committing world leaders to ensuring primary school enrollment for all children by 2000, was launched at the World Conference on Education for All, in Jomtien, Thailand. The need to focus on girls’ education was identified in the meeting's background report, which noted that gender gaps in both school enrollment and retention should be addressed, but that other aspects of gender norms were likely to affect students’ abilities to learn while in school (Inter-Agency Commission for the World Conference on Education for All 1990). The EFA movement was renewed in Dakar, Senegal in 2000, with a commitment by 164 countries to achieve six goals by 2015; Goal 5 focused on eliminating gender disparities in education (UNESCO 2015).5 Alongside strong policy commitments, additional progress was made between 1990 and 2015 in expanding primary school enrollment and closing gender gaps in low-income countries (Grant and Behrman 2010; UNESCO 2016a). The 2016 Gender Review accompanying UNESCO's annual Global Education Monitoring Report stated that, “in 2014, gender parity was achieved globally, on average, in primary, lower secondary and upper secondary education” (p. 14) (UNESCO 2016b). Importantly, these global estimates included high-income countries, where school enrollment is nearly universal but where boys are sometimes disadvantaged relative to girls. Based on analyses of Demographic and Health Survey (DHS) data from 38 low-income countries, Grant and Behrman (2010) found that, on a regional level, gender gaps in primary school completion were fully explained by gaps in enrollment rather than in grade progression. By 1990 to 2006, conditional on school enrollment, gender parity in primary school completion had been achieved in all regions studied, with an emerging female advantage (ibid.). In other words, at the regional level, once girls enrolled in school they completed as many grades as boys, if not more. The 2015 UNESCO Global Education Monitoring Report echoed this point (UNESCO 2016a).6 Rather than this advantage simply reflecting the selectivity of girls who had ever enrolled in school, as one might expect, the female advantage in attainment became stronger between 1990 and 2006, as a higher proportion of females enrolled in school. Grant and Behrman (2010) asserted that, if those trends were to continue, a female advantage in grade attainment would likely emerge in many regions and that programs aiming to close gender gaps in attainment should therefore focus on primary school entry. One might reasonably ask why continued investments in girls’ education are necessary, given that, on average regionally, girls progress as far as boys once they are in school (Grant and Behrman 2010), and that, on average globally, gender parity has been achieved in primary and secondary school enrollment (UNESCO 2016b). One reason is that these estimates mask substantial variations in school enrollment and grade attainment within countries and regions (Grant and Behrman 2010; UNESCO 2016b) and ignore other areas of potential gender inequality, such as learning outcomes and post-schooling opportunities (Grant and Behrman 2010; Lloyd 2009). Further, little is known about the changing patterns of gender gaps over time or whether the absence of a gender gap, particularly in countries with low grade attainment, means that gaps are unlikely to emerge (or re-emerge) as attainment increases. Also, as Grant and Behrman acknowledge, gender differences in learning might arise due to experiences both in and out of school, especially in contexts where social promotion from one grade to the next is common. Beyond the magnitude of gender gaps in school enrollment and attainment, many argue that continued investments in girls’ education are warranted because girls face numerous additional challenges in receiving a quality education (Grant and Hallman 2008; Psaki 2016). Lloyd (2009) noted that, compared to boys, girls face more barriers to staying in school, including earlier marriage and pregnancy, and that girls face greater challenges in entering the labor force. Further, Patrinos (2008) argued that the social returns to girls’ schooling are higher, in part because of women's greater role in reproduction and childrearing and because the private returns to secondary schooling are higher for girls than boys. Following proclamations of global success in achieving gender parity in school enrollment (UNESCO 2016b), SDG 4 calls for universal completion of upper secondary education by 2030. The Incheon Declaration for Education 2030, in support of SDG 4, emphasizes quality education and lifelong learning for all (UNESCO 2016b). These are lofty goals, given the UN's estimates of current levels of attainment: by 2013, only 70 percent of children in low-income countries completed primary school, and only 14 percent completed secondary school (UNESCO 2016a). In fact, UNESCO (2016b) acknowledged that, “if past trends continue, not even the EFA goal of universal primary completion…is likely to be achieved by 2030. The target of universal secondary completion is clearly beyond reach” (p. 24). As the global community devotes resources to the post-2015 education agenda, a clearer picture of progress and challenges in achieving gender equality in education should inform discussions of which investments are needed. Using DHS data from 43 countries collected at two time points, we seek to address key outstanding questions, including: How do patterns of school enrollment and progression differ for girls and boys? What is the relationship between enrollment, attainment, and gender parity? Does gender parity in enrollment lead to parity in learning? Where should investments in girls’ education be directed to maximize impact? The DHS collects nationally representative data on health, population, education, nutrition, and livelihoods in low- and middle-income countries.7 In order to include recent estimates on gender gaps in school enrollment and completion, we limit our analyses to countries that conducted a DHS since 2007.8 We compare results from these most recent surveys with those from a survey conducted approximately ten years earlier in each country. Since the elapsed time between surveys varies between countries, we selected the previous survey in each country that would produce a gap as close to ten years as possible.9 Education data on children and young people aged 6–24 are drawn from DHS household surveys, during which an adult respondent provides information on all household members. Literacy assessments are administered to respondents aged 15–49 as part of individual surveys. DHS data are weighted to be representative at the national and urban/rural levels. Education levels were constructed using the highest level of schooling completed (attainment). Consistent with DHS practice and available data, we defined literacy as the ability to read a complete sentence (or sentences) out loud in a language of the respondent's choosing; those who could read none or some of the sentence were considered illiterate.10 Our analyses proceed in four steps. First, using data on primary school enrollment and attainment for 15–19-year-old females, we identify three stages of development in girls’ education, and group countries by those stages at time 1 and time 2.11 Second, using data from time 2, we describe patterns of gender gaps in school enrollment and attainment, based on each country's stage of girls’ educational development. We define a gender gap as a difference of more than 5 percentage points between the proportion of males and females entering or completing each level of schooling; this gap may reflect either a female disadvantage (positive difference) or a male disadvantage (negative difference). Third, we discuss several ways in which gender parity in educational attainment—a commonly used measure—may provide an incomplete and misleading picture of progress in achieving gender equality in education. We propose a broader approach to tracking progress that integrates information on the country's stage in girls’ educational development and when gender gaps emerge. As a final step, we assess the extent to which parity in educational attainment, when achieved, is likely to translate into parity in learning. We describe patterns of basic literacy among males and females with the same level of educational attainment (completion of primary school) at both time points. We use data from a subset of countries in which literacy was directly assessed in both DHS rounds to show gender disparities in the proportion of young people with basic literacy skills among those who completed primary school at time 1. We also assess changes between time 1 and time 2 in the proportion of males and females able to read a simple sentence, among those who completed primary school. Table 1 shows the countries and timing of survey rounds included in our analyses. Time 1 data were collected between 1997 and 2007; time 2 data were collected between 2008 and 2014.12 The mean gap between survey rounds is ten years, ranging from five years in Sierra Leone, Lesotho, Pakistan, and Honduras to 15 years in Togo and 16 in Comoros. The countries included represent multiple regions, but are heavily concentrated in sub-Saharan Africa given data availability. While our analyses are not globally or regionally representative, they include examples of low- and middle-income countries at different stages of girls’ educational development. At time 1 (1997–2007), on average, 77 percent of females aged 15–19 had ever enrolled in primary school, 49 percent had completed primary school, and 38 percent had enrolled in secondary school. Despite high levels of female primary school enrollment in some settings, none of our 43 countries had achieved universal primary school completion for females at time 1 (not shown).13 By time 2 (2008–2014), levels of enrollment and attainment had improved overall. On average, 87 percent of females aged 15–19 had ever enrolled in primary school, 65 percent had completed primary school, and 56 percent had enrolled in secondary school. Universal primary school completion had been achieved in three countries (Colombia, Indonesia, Jordan). Although improvements have likely continued since time 2, the SDG target of universal secondary school completion by 2030 appears unattainable in most countries without a dramatic acceleration of progress. As is the case with UNESCO's global estimates, mean values based on these 43 countries mask important variations. Even within this sample of low- and middle-income countries, schooling trajectories for females aged 15–19 followed different patterns. Based on the patterns in our data, we grouped countries by three stages of educational development based on enrollment / completion levels observed at time 1: 1) low enrollment in primary school (less than 75 percent of young people); 2) high enrollment in primary school (75 percent or more of young people) followed by high dropout (fewer than 75 percent of young people complete primary); and 3) high enrollment and high completion in primary school (75 percent or more of young people completed primary).14 Figure 1 shows the mean levels of primary school enrollment and completion, and secondary school enrollment at both time points, by countries’ stages of girls’ educational development at time 1.15 Thus, for example, among countries classified as low enrollment at time 1, mean primary enrollment was about 50 percent at time 1 and 70 percent at time 2. Categorizing countries this way shows that, perhaps not surprising, the most dramatic improvements between time 1 and time 2 occurred in low enrollment countries. Countries in this group at time 1 experienced an approximately 20-percentage-point gain in mean primary school enrollment and completion, as well as mean secondary school enrollment. The high dropout countries also made notable progress in primary school completion (16 points) and secondary enrollment (18 points), and the high completion countries made substantial progress in secondary school enrollment (13 points). That is, on average, countries made the most progress at the schooling levels where they had the most room for improvement. At time 1, on average across all 43 countries, an 8-percentage-point female disadvantage existed in primary school enrollment, as well as a 3-percentage-point female disadvantage in primary completion and secondary enrollment. By time 2, the average female disadvantage in primary enrollment had decreased to 4 percentage points, and gender parity had emerged in primary completion and secondary enrollment. Although our sample includes only low- and middle-income countries, this pattern of diminishing female disadvantage is consistent with UNESCO's findings that global gender parity had been achieved in primary and secondary schooling by 2014 (UNESCO 2015). Given different meanings and implications of gender gaps, global summaries mask regional and country-level variations. Gender gaps should be assessed within the context of each country's stage in educational development, as well as the level of schooling at which gaps emerge and grow. Figure 2 shows mean gender gaps at each level of schooling at time 2, with countries now grouped by stage of female educational development at time 2.16 Panel A shows that, among all 15–19-year-olds, countries in the low enrollment group had a mean female disadvantage in primary enrollment (12 percentage points), primary completion (10 percentage points), and secondary enrollment (10 percentage points). There were no mean gender gaps in the high dropout group. In the high completion group, despite mean gender parity in primary enrollment, there were male disadvantages in primary completion (4 percentage points) and secondary enrollment (5 percentage points). However, conditional on enrollment or completion of the previous level, the picture of mean gender gaps changes (Panel B). In low enrollment countries, although more males than females enrolled in school on average, among those who ever enrolled there was no notable gender gap in primary completion or secondary enrollment at time 2. That is, on average the gender gaps in primary completion or secondary enrollment in low enrollment countries were completely attributable to gaps in primary enrollment, not dropout. In the high dropout and high completion countries, the patterns, while less pronounced, indicated that male disadvantages in secondary enrollment were attributable, in part, to gaps in primary completion. The preceding sections showed that a country's stage of educational development provides critical context for understanding progress in increasing female attainment and achieving gender parity, and understanding the level of schooling at which gender gaps emerge is important context for interpreting gaps in attainment. Yet, in practice, progress in achieving gender equality in education is often measured much more narrowly, without attention to these two contextual factors. Gender parity in educational attainment is often used as the main measure of progress in achieving gender equality in education, including progress toward Sustainable Development Goal 4.17 This choice of indicator is intuitively appealing, as it gives an immediate sense of relative levels of school participation, and is also simple to track using available data. In theory, interpretation of gaps in primary school enrollment is straightforward because both boys and girls are starting from a pool of 100 percent of young people eligible to enroll.18 However, seemingly comparable gender gaps in grade attainment may, in fact, reflect different patterns of progress and present different challenges for policy or other interventions. By highlighting selected examples from our data, all from time 2, Figure 3 shows four ways in which gender gaps may provide a misleading picture of progress in achieving gender equality in education: A) gaps in attainment may reflect disparities in enrollment and/or progression; B) gaps in dropout may be an artifact of enrollment gaps; C) parity in low attainment countries may mask inequalities; and D) comparable gaps may occur in very different education settings. As was shown in Figure 2, and demonstrated at a regional level by Grant and Behrman (2010), in many countries gender gaps in primary school completion or secondary enrollment largely reflect disparities in primary school enrollment. While this may be true on average, there are exceptions—countries where gender gaps reflect both enrollment and progression through school. The policies and interventions needed differ depending on the level(s) at which gaps emerge, yet data on gender gaps in attainment may mask these differences. For example, Panel A shows two countries, Benin and Nepal, with similar gaps in primary school completion levels. In Nepal, the gender gap in primary completion is almost completely attributable to a gap in primary enrollment (i.e., conditional on enrollment, the gap almost disappears), whereas in Benin the gender gap in attainment reflects female disadvantages in both primary enrollment and dropout. At time 2 in our full sample, 14 countries had female disadvantages in primary school completion among all 15–19-year-olds. While on average, as noted above, the gender gaps in low enrollment countries were attributable to gaps in primary enrollment, three countries (Benin, Côte d'Ivoire, Togo) also had female disadvantages in dropout during primary school. Data on gender gaps in education must include information on the level of schooling at which gaps emerge and then grow or narrow. At time 2 in Niger, 24 percent of males compared to 14 percent of females dropped out during primary school. This can be seen in Panel B, where the large female disadvantage in primary enrollment narrows by primary completion. This comparison alone, without information on levels of enrollment, may lead some to believe that males were at a disadvantage in schooling. However, a higher proportion of males than females completed every level of schooling. This pattern of higher male dropout was common in our data and contrary to common assumptions about school dropout during adolescence. However, in several countries with higher male dropout, a female disadvantage in primary school enrollment was also observed. Thus, in those settings, males simply caught up with—or dropped down to—the low levels of female enrollment by the end of primary school. In most countries with higher male dropout, nearly all males and females enrolled in school, and a high proportion of both groups—but relatively more males—dropped out by the transition to secondary school. In such cases, the most important policy challenge is likely the overall high dropout levels. Data on gender gaps in dropout, without information on enrollment levels, may create a false impression of which group, if either, is at a disadvantage. Absence of gender gaps is not always an indication of gender equality. Panel C shows two countries, Malawi and Senegal, with no gender gaps but high levels of primary school dropout at time 2. In settings like this, there may be different gender-related barriers facing males and females, such as pressure on males to earn an income and pressure on females to marry. Alternatively, some barriers, such as high student/teacher ratios, may affect males and females differently (Chin 2005). Finally, some sex- and gender-related experiences, such as menstruation, may present important challenges, even if they do not cause school dropout (Sumpter and Torondel 2013). In countries where attainment is low overall, the lack of a gender gap may not be very informative with regard to gender inequalities in schooling. Further, gender gaps in enrollment or attainment may emerge as enrollment increases, as occurred in two countries in our sample (Benin, Côte d'Ivoire). Gender gaps alone may be misleading indicators of where education investments are most urgently needed. Panel D shows comparably sized gender gaps in Liberia and Colombia. In Liberia, attainment levels were low for both males and females. Efforts in these settings should address overall low attainment as well as gender-related barriers to schooling. Gender gaps in high enrollment settings, such as the male disadvantage shown in Colombia, likely have different origins and implications than comparable gaps in countries with much lower levels of attainment. For example, male disadvantages in schooling in high enrollment countries may reflect male advantages in access to labor markets. To address the drawbacks of using gender gaps alone as indicators of progress in achieving gender equality in education, we assess progress between time 1 and time 2 on two dimensions: gender gaps and levels of female enrollment and attainment. We divide countries into the following mutually exclusive groups, based on whether gender gaps exist at two key schooling transitions: primary school enrollment and/or progression through primary school. The resulting groups are: 1) lower female enrollment and higher male dropout; 2) lower female enrollment and parity in dropout; 3) lower female enrollment and higher female dropout; 4) parity in enrollment and higher male dropout; 5) parity in enrollment and higher female dropout; and 6) parity in both enrollment and dropout. For simplicity, we do not include all possible combinations of gaps in enrollment and progression because some patterns are rare or do not occur in our data.19 The excluded patterns may, however, exist in other settings, or emerge in the future in these settings. In Figure 4, we combine the above gender gap groups (columns) with the stages of educational development (rows) presented in the previous section to create a two-dimensional profile of progress in absolute and relative educational attainment between time 1 and time 2.20 The expected progression in this framework is roughly from the bottom left boxes (low enrollment overall, lower female than male enrollment) to the top right box (high completion, parity at all levels). Although most countries made progress in either gender parity or attainment, only Comoros, Ghana, and Sierra Leone made progress in both areas.21 Many more countries made progress between time 1 and time 2 in closing gender gaps than in substantially increasing female attainment levels. The number of countries with a female disadvantage in primary school enrollment was nearly halved (from 21 to 12) during this interval, and the number with gender parity in primary school enrollment, completion, and secondary school enrollment increased from 12 to 20. Less progress was made in increasing attainment: only six countries shifted to a higher stage of girls’ educational development (female attainment) between time 1 and time 2. Patterns of progress in this two-dimensional framework further highlight the limitations of using only gender gaps in grade attainment to measure progress in achieving gender equality in education. At time 2, 12 countries had lower levels of female than male enrollment in primary school;22 three of these countries (Togo, Benin, Côte d'Ivoire) also had higher levels of female than male dropout during primary school. While patterns of gender gaps in enrollment and dropout may change as enrollment increases, in most countries at time 2 female disadvantages in attainment were due to gaps in primary school enrollment rather than to gaps in grade progression. Out of the 12 countries at time 2 with female disadvantages in primary enrollment, eight also had low levels of female enrollment, and three had high levels of female dropout. That is, in nearly all countries with female disadvantages, efforts are needed both to substantially increase attainment and to close gender gaps. Figure 4 also underlines the fact that comparable gender gaps may occur in different education settings. At both time 1 and time 2, female disadvantages were concentrated in countries with low attainment levels, and male disadvantages were concentrated in countries with high attainment. The causes and implications of these gender gaps are likely different, and they should not be equated. Both males and females were worst off in countries with female disadvantages. That is, female disadvantages are a sign of a weak education system overall, whereas male disadvantages are more common in relatively stronger education systems. Many other researchers have similarly noted this relationship between overall enrollment and the direction of gender gaps (Eloundou-Enyegue, Makki, and Giroux 2009; UNESCO 2016b; Wils and Goujon 1998). Therefore, targeting policies and programs to countries with female disadvantages, and ensuring they address broad challenges in the education system as well as gender-specific challenges, is likely to benefit both girls and boys. Gender parity in grade attainment is the measure of progress in girls’ schooling that has received the most attention from policymakers and practitioners, but little is known about the extent to which parity in attainment is likely to translate into parity in learning or post-schooling outcomes. Figure 5 shows the proportion of 15–19-year-olds who were able to read a simple sentence, among those at time 1 who had completed primary school but did not continue to secondary.23 Assuming that students received a quality education regardless of sex, and that skills were retained or strengthened after leaving school, we would expect all 15–19-year-olds who completed primary school to have basic literacy skills; this is clearly not the case. In nine of the 24 countries shown at time 1, less than 50 percent of females who completed primary education had basic literacy skills. Also, among young people with primary school attainment, gender gaps in basic literacy existed in about half of countries at time 1. Female disadvantages were more common than male disadvantages. A similar pattern is evident in the apparent stagnation—or deterioration in some countries—between time 1 and time 2 in the proportion of young people with a primary school education who had basic literacy skills (see Figure 6). At time 1, on average 70 percent of males and 65 percent of females who had completed primary education could read a simple sentence. These levels appear to have stagnated by time 2, when 69 percent of males and 62 percent of females with primary attainment had basic literacy skills. In ten countries both males and females with primary school attainment were less likely to be literate at time 2 than at time 1; this pattern was observed for females only in an additional two countries, and for males only in an additional two countries.24 These results for literacy underline the importance of a three-part strategy of promoting educational attainment, closing gender gaps, and improving the quality of education. Using data collected at two time points in 43 countries, we explored progress in increasing female educational attainment levels and closing gender gaps. We found that, on average, female disadvantages in primary school enrollment have decreased over time, and gender parity has been achieved in primary school completion and secondary school enrollment. However, our findings raise two caveats about those achievements: first, global averages mask important variability between countries; second, despite prog