Can test results lead to education reform?
@alenestour.bsky.social and @leecrawfurd.bsky.social show how PASEC results have improved learning in Francophone Africa.
At a few cents per child, the question is not cost but whether countries can afford to lose it.
cgdev.org/blog/pisa-shocks-africa
You might have heard of "PISA-shock".
You probably haven't heard of "PASEC-shock" in Niger
[new blog with @alenestour.bsky.social]
C’est bien qu’il y ait un chapitre sur le vote à 16 ans qui n’est pas assez discuté en France à mon avis. On devrait d’ailleurs aller encore plus bas si on veut que le corps électoral soit vraiment représentatif.
Au Royaume-Uni, le taux de pauvreté absolu est 60% du revenu médian de 2010/11, ajusté de l'inflation. Donc ça dépend bien du contexte. Le précédent seuil datait de 98/99. A mon avis, ça a du sens d'avoir un seuil fixe sur 10-20 ans.
Oui je suis bien conscient des limites des deux approches mais je trouve difficile d'apprécier les tendances avec un seuil qui change chaque année. Je préfèrerais que l'INSEE fasse comme le RU et reporte les taux de pauvreté absolue et relatif.
Comment le niveau de vie median a-t-il évolué sur la période ? Est-il possible de recalculer la série avec un seuil de pauvreté absolu ?
❓Why do the Nordics & Dutch speak English so much better than the Germans, Italians & French?
➡️ New Working Paper:
Out-of-School Learning: Subtitling vs. Dubbing and the Acquisition of Foreign-Language Skills
w/ F. Baumeister & E. Hanushek
www.nber.org/papers/w33984
A 🧵 1/12
New blog: @rglenner.bsky.social & I suggest 4 criteria for donors choosing between multilateral education funds (eg GPE/ECW/IDA/IFFED/UNICEF/UNESCO)
1. Cost-effective interventions
2. Foundational literacy & numeracy
3. Poor countries
4. Financial & policy leverage
www.cgdev.org/blog/four-cr...
I’m happy to share our latest Data Must Speak report for #BurkinaFaso.
We collaborated with the Ministry of #Education to analyze their admin education data and model the factors influencing school performance.
Main results in a 🧵
@unicefinnocenti.bsky.social
www.unicef.org/innocenti/re...
A 10 μg/m3 increase in daily PM2.5 air pollution causes a 5.7 percent increase in full-day student absences, a 13.1 percent increase in teacher absences, and a 28 percent increase in behavior referrals, from Sarah Chung, Claudia Persico, and Jing Liu https://www.nber.org/papers/w33549
"We might see a fall in ODA, by 2027, of at least 25% relative to 2023 levels. And that’s very much a best-case scenario." ☹️
devpolicy.org/burden-shedd...
There are many more findings in the report (textbooks, school distance, canteens, etc.). Our models use data from all schools over 5 years with school fixed effects. Education admin data (EMIS) offers huge potential to inform policy—let’s keep using it!
#DataMustSpeak #EduData #UNICEF
Regression table showing the differences in terms of promotion rate and success at the primary leaving exam between public schools and different types of private schools in Burkina Faso with no controls and after controlling for parental profession, location and violence in the area.
8) Private sector
Private schools outperform public ones on exam success, but the advantage largely fades when controlling for student background and location. Catholic schools still perform better after controls, while Muslim schools—often rural with poorer students—do worse than public schools.
Graph showing the association between active parental associations, mother associations and school committees in Burkina Faso and various school characteristics or actions. Community engagement is especially important for health actions (eg medical visits or deworming).
7) Community engagement.
Active parental associations (APE), mothers' associations (AME) or school management committees (COGES) play a positive role in students' schooling conditions, notably for the presence of health interventions : bednet use, medical check-ups or deworming.
Graph showing the number of students affected by school closure and students categorized as internally displaced by year (2019-2023) in Burkina Faso. In 2019, there were about 100k students affected and there were more than 800k in 2023. The low number of internally displaced students (<100k) show that most students affected by school closure can't reenrol.
6) Security crisis
1 in 3 schools closed in 2022/23 due to insecurity, affecting 800k+ students; few reenrol elsewhere.
Linking with ACLED data shows violence also raises dropouts in schools that remain open, hitting girls 50% harder.
5) Multigrade
Students with multigrade teachers perform better (+3.2 pts promotion, +3.5 pts exam pass). Yet only 1% in public schools vs. 20% in private benefit. Multigrade teaching may hold insights for improving education quality.
Graph bar showing large disparities in average classe size in public primary schools between regions in Burkina Faso. From 100 in the Sahel region to about 60 in the Centre, Sud-Ouest and Plateau. Dots show that class sizes are always larger in rural areas (up to 150 in the Nord region).
4) Class size
Public classrooms in Burkina Faso have 68 students on average. Reducing class size by 10 could improves promotion (+2 pts) and exam pass rates (+1.5 pts). The issue isn’t teacher shortage but lack of classrooms—especially in cities and conflict-affected regions.
3) Teacher Training
1 in 3 private school teachers—and 3 in 4 in private Muslim schools—have no formal training.
Untrained teachers are associated with lower student promotion (-0.9 points) and exam pass rates (-1.6 points).
Bar graph showing the non-adjusted and adjusted difference of the success at he primary leaving exam between children of civil servants, formal workers, other jobs and merchants and children of farmers. The difference is as large as 38 points when not adjusted for children of civil servant but decrease to 9 points when adjusted for school characteristics and location. For the other categories, the difference is close to 0 when adjusted.
2) Inequalities
Children from more advantaged backgrounds score higher on the primary exam. But most of this gap fades when accounting for school type & location.
🚨 Exception: Civil servants’ kids still hold a strong advantage (+9 pts), even after controlling for school factors.
Bar graph showing total enrolment for girls and boys in primary school in Burkina Faso in 2022-23 by grade and the number admitted to end of primary leaving exam (CPE). Numbers drop from 300k in grade 1 to less than 200k in grade 6 and there are more girls than boys enrolled by grade 4.
1) Gender
➡️ More boys enter primary school, but their enrollment drops rapidly. By Grade 4, there are more girls, and more girls pass the primary exam (CPE).
➡️ Girls benefit more from female teachers, proximity to school, and access to canteens/latrines—factors that matter less for boys.
I’m happy to share our latest Data Must Speak report for #BurkinaFaso.
We collaborated with the Ministry of #Education to analyze their admin education data and model the factors influencing school performance.
Main results in a 🧵
@unicefinnocenti.bsky.social
www.unicef.org/innocenti/re...
À quoi ça sert de mettre le genre sur le passeport maintenant qu’on a des passeports biométriques ?
If VAT is a tariff, si is sales tax....
Part 1: Bitcoin is the most secure way to hold assets!
Part 2: buying a landfill and going through 100,000 to get your money back.
Oui je veux bien croire que c’est galère mais de là à financer 120% du salaire pendant 2 ans, ça me paraît beaucoup.
Honnêtement j’ai du mal à comprendre l’intérêt social d’une telle mesure. Pourquoi subventionner le salaire des plus diplômés ?
Interesting new paper on unemployment of college graduates in Burundi. Using a field experiment, Ndayikeza shows that college graduates are more likely to find a high-skill job if they have low-skill job experience. www.sciencedirect.com/science/arti...
Bonne initiative.
PEPFAR is *currently providing* lifesaving HIV treatment for 20.6 million people, including 566,000 children, as of Sept. 30, 2024.
That treatment is now in serious jeopardy.
I'm sure people in the administration don't want half a million kids on their conscience.
www.nytimes.com/2025/01/24/u...
AI has amazing prospects for education but, as for other edtech interventions, costs and logistics can be the real bottlenecks.