Working Paper: NBER ID: w27298
Authors: Annie Duflo; Jessica Kiessel; Adrienne Lucas
Abstract: We partnered with the Ghanaian government to test simultaneously four methods of increasing achievement in schools with low and heterogeneous student achievement| assistant led remedial pull-out lessons, assistant led remedial after school lessons, assistant led smaller class sizes, or teacher implemented partial day tracking. Despite implementation issues, the interventions increased student learning by about 0.1SD, about 0.4SD when adjusting for the imperfect implementation, with no effects on attendance, grade repetition, or drop-out. Test score increases were larger for girls and gains persisted after the program ended. Fidelity of implementation decreased over time for the assistants but increased for the teachers.
Keywords: education; learning interventions; randomized controlled trial; Ghana
JEL Codes: I21; I25; I28; J24; O15
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| fidelity of implementation (L15) | educational outcomes (I26) |
| four interventions (C90) | attendance (I29) |
| four interventions (C90) | grade repetition (A23) |
| four interventions (C90) | dropout rates (I21) |
| treatment received by students (I24) | student learning (A21) |
| treatment received by students (I24) | student test scores (I21) |
| four interventions (C90) | student learning (A21) |
| four interventions (C90) | student test scores (I21) |