NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Examining the impact and cost-effectiveness of supplementary math courses with a focus on girls on Benin
Initial registration date
March 28, 2017
March 28, 2017 4:39 PM EDT
African School of Economics
Other Primary Investigator(s)
Additional Trial Information
In Benin, as elsewhere in the world, education of girls lags behind that of boys in a number of dimensions, e.g. enrollment, promotion and graduation rates. While these measures have seen some improvement in the recent years, they do not guarantee that neither girls nor boys obtain quality education. One area where quality of education is especially critical, particularly for girls, is numeracy skills, as a solid quantitative foundation could highly increase their post-secondary education paths and employment prospects. Notwithstanding the importance of these skills, Beninese teachers are insufficiently trained in science subjects and are overburdened with work. Temporary supplementary teachers who have been trained to provide math classes and who can spend three months at a school represent a straightforward solution to this problem. This evaluation will examine the impact of temporary supplementary teachers in improving the learning outcomes in junior and secondary high schools located in rural Benin. In a second treatment arm, with only female supplementary teachers give supplemental sessions that are only available to girls and designed to cater to the specific needs and learning levels of female students. The female supplementary teachers are also trained to give life skills training to the girls.
Olapade, Markus and Leonard Wantchekon. 2017. "Examining the impact and cost-effectiveness of supplementary math courses with a focus on girls on Benin." AEA RCT Registry. March 28.
Girls' education lags behinds boys' cross Sub-Saharan Africa, with only 21 percent of girls completing secondary school (compared to 28 percent of boys) and 66 percent able to read and write (compared to 76 percent of boys). Better numeracy skills could give girls the opportunity to pursue higher education in traditionally male-dominated fields such as science and engineering, and thereby improve their employment prospects. However, teachers are often insufficiently trained and overburdened with work. In such contexts, adding temporary supplemental teachers trained to teach math could improve learning outcomes and employment prospects. Can supplementary math teachers improve numeracy skills and professional aspirations among girls?
In Benin, 42 percent of secondary-school aged youth are enrolled in school and girls are one-third less likely to be enrolled than boys. For those who do attend school, quality of instruction is often low. Despite the importance of quantitative skills in the 21st century, Beninese students' enrollment in math and science courses has declined since the 1980s and, in some schools, the math and science track has been cancelled completely.
In partnership with the Institute for Empirical Research in Political Economy (IERPE), the National Institute for Math and Physics (IMSP), and the Ministry of Secondary Education, researchers are evaluating the impact and cost-effectiveness of temporary supplemental math teachers on girls' numeracy skills, professional aspirations, and early career labor market outcomes. From a subset of schools that the Ministry of Secondary Education classified as having inadequate math instruction, researchers randomly selected 90 schools to compare how different variations of this intervention affected girls' math skills. Researchers assigned schools to one of two supplementary teacher program variations or to a comparison group.
• Supplemental math teachers: IERPE and IMSP trained older male students to serve as supplementary teachers for three-month periods. Supplemental sessions were available to all students (both boys and girls) in an eligible class.
• Girls-only math and life skills supplement: IERPE and IMSP trained female supplemental teachers, but supplemental sessions were only available to girls and designed to cater to the specific needs and learning levels of female students. The female teachers in this group were also trained to give life skills training to the girls.
• Comparison group: No supplemental sessions were made available to students.
In 2018, researchers plan to conduct an endline survey and gather school administrative data to measure each intervention's effects on student performance and aspirations. Researchers also plan to compare the cost-effectiveness of the two approaches.
Intervention Start Date
Intervention End Date
Primary Outcomes (end points)
Math test score, Knowledge of HIV/AIDS, Professional Aspirations
Primary Outcomes (explanation)
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
In a sample of 90 randomly selected schools stratified by geographical location within the country (North, Center, South), we randomly select the level at which the teachers intervene. Further, we alternate the treatment levels across the two intervention years.
We have three groups of schools, A, B, and C in which we intervene in years one and two. In year 1, we allocate the 90 teachers to three different groups of schools, i.e. A1, B1 and C1:
Group A1. 30 male teachers intervene at 6ieme (first year of junior high school) in 30 schools.
Group B1. 30 male teachers intervene at 3ieme (last year of junior high school) in 30 schools.
Group C1. 30 female teachers intervene at both 6ieme and 3ieme but only for girls.
In year 2 we keep the same groups of schools, A2, B2, and C2, but change the level at which the teachers intervene. A2. 30 male teachers teach 3ieme
B2. 30 male teachers teach 6ieme
C2. 30 female teachers teach 6ieme and 3ieme but only girls
Experimental Design Details
in office by a computer
We randomize schools into the three different groups and within schools randomly select classes that receive the complementary teacher.
Was the treatment clustered?
Sample size: planned number of clusters
Sample size: planned number of observations
Regarding individual sample size, we plan to interview 40 students (20 boys and 20 girls) per school/per level, which will allow us to attain a sample of 7200 pupils in year one and 7200 in year two. This sums up to a total of 14400 students, half of which are girls and half of which are boys, over two years.
Sample size (or number of clusters) by treatment arms
In 60 schools male teachers will intervene at two different levels, the first year of junior high school (30 schools) and the last year of junior high school ( 30 schools). In the second intervention year the same schools will be maintained but the level of intervention will be switched. Since we alternate the level at which the teachers intervene from one year to the other we consider that, in 2017, there are 30 schools that receive the teacher in the first year of junior high school and 30 schools that receive the treatment in the last year of junior high school. In 2018, the 30 schools that received treatment in the first year of junior high school, now receive it in the last year of junior high school and vice versa. Accordingly, while we are intervening in 60 schools only, we argue that we are treating a total of 120 schools (60 schools in the first year of junior high school and 60 schools in the last year of junior high school) over the two years. The number of students in this arm is 9600 over the two years.
For the second treatment, the female teacher intervenes in both levels, first and last year of junior highschool, in both years and thus we are intervening in 30 schools and interview a total of 4800 students over the two intervention years.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
For the main outcome, i.e. math test scores, we are expecting to be able to detect MDEs of 0.16 standard deviations in test scores at 80 percent power.
For the female teachers we are expecting to detect an MDE of .20 standard deviations at 80 percent power.
INSTITUTIONAL REVIEW BOARDS (IRBs)
Post Trial Information
Is the intervention completed?
Is data collection complete?