64499 July 2011 – Number 41 AN EXCEPTION TO THE GENDER GAP IN EDUCATION: THE MIDDLE EAST? Mourad Ezzine, Simon Thacker and Nadereh inequality, there is no gender gap in Chamlou 1 mathematics on average in these places,�3 the authors write. What we will see in fact is that Introduction: An interesting consequence of there is a reverse gender gap with girls the Arab Spring is that it is compelling the West outperforming boys in grade 4 results in the to re-evaluate its understanding of the Middle Middle East, a trend that continues into grade East. Stereotypes and misconceptions have 8, though there are some distinct exceptions abounded, but today the region can no longer beginning to be noticeable by that grade, too. be grossly dismissed as home only to extremists or the oil-rich. Instead, a more nuanced and accurate picture is emerging: one that is as full Grade 4 Math TIMSS by of contradictions as an image of any region would be. For a region not known for its Gender for 1995, 2003, equitable attitudes towards women, for 2007 instance, the Middle East offers up some surprising results for girls in school, results that 500 450 1995 ♂ are much better in some ways than the rest of 400 350 1995 ♀ the world. 300 250 200 2003 ♂ The Gender Gap in Mathematics: A recent Yemen Kuwait Qatar Dubai, UAE Algeria Tunisia Morocco 2003 F♀ Iran National Bureau of Economic Research (NBER) working paper by Roland Fryer and Steven 2007 ♂ Levitt, "An Empirical Analysis of the Gender 2007 ♀ Gap in Mathematics,"2 finds evidence for a gender gap in elementary school level Figure 1 mathematics in the United States, a gender gap that they find, extending their analysis to The study is based on data from the Early international results, in elementary- and Childhood Longitudinal Study Kindergarten Cohort secondary-level students around the world – (ECLS-K), a data set covering a sample of more except the Middle East. than 20,000 children from approximately 1,000 US schools entering kindergarten in the fall of “Surprisingly, although these Middle Eastern 1998, who were then subsequently re- countries have a high degree of gender interviewed in the spring of kindergarten, first grade, third grade, and fifth grade. 1 Mourad Ezzine, Education Sector Manager, Middle East and North Africa Human Development Department, The The study finds that on entering kindergarten, World Bank and Simon Thacker, Middle East and North Africa Human Development Department, The World Bank. girls and boys are observationally equivalent in Nadereh Chamlou, Senior Advisor, MENA Chief Economist’ Office, The World Bank. 3 Fryer and Levitt (2009), p. 7. 2 See www.NBER.org both math and reading. By the end of fifth performance by girls rather than an unusually grade, however, girls have fallen more than 0.2 weak showing by boys; nor does it appear to be standard deviations behind their male due to selection bias arising from a smaller, yet counterparts in math, a gap that is equivalent to academically stronger, share of girls in school roughly 2.5 months of schooling -- in either. comparison girls start .15 standard deviations ahead of boys in reading and retain that One intriguing hypothesis the authors put advantage. The study goes on to show that girls forward to explain this has to do with same-sex are losing ground in math in every region of the classrooms. In most of the world, most classes country, every racial group, all levels of the are mixed-gender. In a handful of the Middle socio-economic distribution, every family Eastern countries that participated in TIMSS structure, and in both public and private (Bahrain, Iran, Jordan, Palestine, and Saudi schools. Arabia), virtually all secondary schooling is gender-segregated, while in two other A range of explanations is put forward for the countries, Egypt and Syria, there is a mix of U.S. data, including less investment by girls in same and mixed-gender classrooms, but the math, low parental expectations, and biased majority are same-sex. As it is in these tests, but the authors find little support for any classrooms environments where girls of these theories. Paradoxically, the evidence outperform boys, this leads the authors to suggests that the gender math gap is notably contend that: “Cross-country data are large among children who attend private consistent with the hypothesis that mixed schools, have highly-educated mothers, and gender classrooms are a necessary component have mothers working in math-related for gender inequality to translate into poor occupations -- all factors that one would think female math performance� (Fryer and Levitt, would actually be in line with girls’ success in 2009, p.21). math. In this connection, it would have been When Fryer and Levitt turn their attention to interesting to know whether mathematics international results, probing data from PISA scores taken from the ECLS-K dataset for girls and TIMSS exams, what they find there and boys taught in same-sex classrooms in the confirms what they see in the US -- except in US would have corroborated this hypothesis or the Middle East. Referring to a paper by Guiso not, but this was not the authors’ primary et al. (2008)4 that uses the PISA exam for 15 year research question. Much evidence exists olds, the authors show that the results are demonstrating the improved performance of largely consistent with the ECLS-K data for the girls in mathematics in same-sex classrooms5, US with boys outperforming girls in math in though a clear counter-example is mentioned in the majority of countries, while girls the study: that of Scandinavian countries, outperform boys in reading in every country of considered to be gender-equal societies, where the sample. Guiso et al. shows however that the the gender gap effectively disappears. gender gap in math scores is strongly negatively correlated with various measures of However, there is much room for argument gender equality such as the World Economic here: not all primary schooling in the Arab forum’s Gender Gap Index. world is segregated. Yet it is in primary school For instance, TIMSS results from 2007 for grade that students are first tested with the grade 4 4 and grade 8 mathematics corroborate these TIMSS math exam and the gender gap is findings. In Bahrain and Iran, which are among already apparent. In Tunisia where classes are the worst in terms of gender equality, girls are not segregated, TIMSS results clearly easily outperforming boys on math, a demonstrate statistically significant differences phenomenon due to relatively strong favoring girls who score 337, on average 18 4 5 “Culture, Gender, and Math,� Luigi Guiso, Ferdinando see http://www.singlesexschools.org/research- Monte, Paola Sapienza, and Luigi Zingales, Science, 320, No. singlesexvscoed.htm 5880 (2008), 1164-1165. July 2011 · Number 41 · 2 points better than boys at 319 (TIMSS 2007 in Figure 2 below. Of the many conservative mathematics for Gr. 4). In Figure 1, we see the societies with same-sex classrooms, all have results by gender of those Middle Eastern girls outperforming boys, but in more secular Countries participating in the TIMSS Grade 4 cultures, many more exceptions show up: mathematics. Girls outperform boys in all cases, Morocco’s results continue to show boys but two: in Morocco where boys did better than outperforming girls, as they did in grade 4 girls in 2003 and 2007, and in Iran in 1995, results, while Tunisia, whose results in grade 4 where boys did better than girls (but where favored girls, has now done an about-face with subsequently in the next two testing cycles the boys outscoring girls. That said, other less trend was reversed). conservative countries, like West Bank and Gaza (WBG), show girls dominating the scores. However, the trend is less apparent when we consider the grade 8 TIMSS mathematics results Figure 2 Grade 8 Math TIMSS by Gender for 1999, 2003, 2007 475 455 435 415 1995 ♂ 395 375 1995 ♀ 355 2003 ♂ 335 315 2003 F♀ 295 275 2007 ♂ 2007 ♀ A perfectly satisfying explanation for all these particularly true of the Middle East where we differing results remains elusive -- perhaps observe what seems to be a reversal of the results from TIMSS 2011 will help. It is gender gap sometimes. Clearly much more fascinating how social context and environment research is needed. can frame academic performance. Unfortunately, it is still not clear what the root In the meantime, one inescapable observation is causes are or how to effectively structure that girls are now doing better now than boys incentives to systematically attenuate the in most subjects in the region, with the result gender gap. This is true in the US and also that many more of them are going on to tertiary July 2011 · Number 41 · 3 education. According to UNESCO figures, in 19 women was 52.1% and 52.7% for these same countries of the region that were surveyed, the years).8 As women's educational attainment in gross enrolment rate for women in tertiary MENA countries has increased, more women education is increasing annually, and in 15 of have tried to move towards the job market. But those countries, the women’s rate is now equal there, not only are employment prospects dim, to or exceeds the men’s rate.6 advancement prospects, when they are finally employed, are poor too. World Bank figures for This is leading to its own set of complications. unemployment suggest that educated women In the 2010 edition of Women's Rights in the have the highest rates of unemployment in the Middle East and North Africa 7 one of the many region. As table 1 demonstrates, in three secular findings deals with women’s advances in countries (hence those with lesser constraints education. Over the past 10 years, women in all placed on women), the unemployment rates of MENA countries except Yemen have made women with secondary and tertiary education gains in access to education, literacy, university is anywhere between one and five times as high enrollment, and the variety of subjects open to as their male counterparts with comparable study. There has, as well, been an increase in education. For instance, 29.1% of female the availability of vocational training schools university graduates in Egypt are unemployed, and business colleges for female students. almost three times higher than their male counterparts. Yet once employed, women of the But these advances are coming at a price: in the region also face significant and often fields of mathematics and science, the report insurmountable obstacles to their advancement. points out some glaring imbalances: Some have suggested that one cause of the Although women are generally still Arab Spring uprisings was that the educated encouraged to study in such traditionally populace in Tunisia and Egypt were held down female disciplines as education and too long. As women in the region progress medicine, in many countries women’s educationally, one wonders how long they numbers have increased in the fields of might be held back too. science and engineering. Particular progress has been visible in the Gulf States, where Contact MNA K&L: women are now joining new professions in Emmanuel Mbi, Director, Strategy and Operations. MNA Region, The World Bank substantial numbers and are increasingly going abroad on government scholarships. Regional Quick Notes Team: There has, however, been something of a Omer Karasapan, , Roby Fields, and Hafed Al- backlash in a few MENA countries against Ghwell women’s involvement in non-traditional study areas. In Kuwait, women who want to Tel #: (202) 473 8177 study in certain traditionally male fields, The MNA Quick Notes are intended to summarize such as engineering, must achieve a higher lessons learned from MNA and other Bank Knowledge and Learning activities. The Notes do not grade point average for admission than men. necessarily reflect the views of the World Bank, its In Oman, women students often must board or its member countries. postpone university study for one year, a limitation not applied to men. Women’s participation in the labor market is low by world standards but is slowly increasing. It was 22.6 % in 2000 rising to 24.8% in 2009 (in comparison the world average for 6 http://stats.uis.unesco.org/unesco/TableViewer/tableVie w.aspx 7 Found at 8 http://www.freedomhouse.org/template.cfm?page=163 ILO, Trends econometric models, October 2010 July 2011 · Number 41 · 2