Education
Schooling, learning and school programs. Real interventions, each with its sources: what worked, what didn't, and why.
6 worked3 mixed2 didn't work
What worked
- WorkedEducation
Teaching at the Right Level
ChallengeMany Indian children couldn't read or do basic math even after years in school.
Grouping children by what they can actually do, not by their grade.
0.70 SDlanguage gain from learning camps, Uttar Pradesh, all enrolled studentsRCTRead case - WorkedEducation
School-based deworming in Kenya
ChallengeIntestinal worms made schoolchildren sick and kept them out of class.
Treating schoolchildren for intestinal worms to improve health and attendance.
+9.3 ptsschool participation in year one, treated pupils vs comparison (J-PAL)RCTRead case - WorkedEducation
Project STAR: smaller classes in Tennessee
ChallengeNo one knew if smaller classes really improved learning for young children.
A statewide experiment that randomly assigned children to small or regular classes.
11,600students randomly assigned to class sizesRCTRead case - WorkedEducation
The Perry Preschool Project
ChallengeDisadvantaged children started school far behind and fell further behind over time.
High-quality preschool for disadvantaged children, followed into middle age.
7–10%estimated annual return to societyRCTRead case - WorkedEducation
Bogotá's Ciempiés walking school routes
ChallengeUnsafe streets and long commutes were pushing children to change or skip school.
Guided walking caravans that take children safely to school.
~$285per child per year, cheaper than school busesQuasi-experimentalRead case - WorkedEducation
National School Feeding Programme (PNAE): Brazil
ChallengeChild hunger and poor diets hurt learning, while small farmers lacked stable markets.
Free meals for all public school pupils, with 30% of federal food funds spent on local family farms.
~40mpublic school pupils served each yearObservationalRead case
Mixed results
- MixedEducation
Chile's nationwide school vouchers
ChallengeChile wanted competition between schools to raise educational quality.
Public funding that follows each student to any public or private school.
+20 pprise in private enrollment rate; no gain in average test scoresQuasi-experimentalRead case - MixedEducation
No Child Left Behind
ChallengeAchievement gaps between groups of students were large and often hidden in averages.
US federal law requiring annual testing and penalties for schools that missed targets.
0.23effect size on fourth-grade math scores (NAEP) by 2007Quasi-experimentalRead case - MixedEducation
Providence Talks: closing the early word gap
ChallengeChildren from low-income homes heard far fewer words before kindergarten.
Coaching parents to talk more with babies, using small recorders that count words.
11th → 42ndpercentile of adult words heard, home-visit families below median (control: 11th → 35th)Quasi-experimentalRead case
What didn't work
- Didn't workEducation
One Laptop per Child in Peru
ChallengeRural students lacked access to technology and learning materials, and test scores lagged.
Low-cost laptops for students in rural primary schools.
0.003 SDestimated effect on average math and language score after 15 monthsRCTRead case - Didn't workEducation
D.A.R.E. school drug education (original curriculum)
ChallengeYouth drug use was rising and schools wanted a way to prevent it.
Police officers teaching kids to resist drugs in the classroom.
≈ 0effect on later drug useMeta-analysisRead case