Luminos Program Impact Evaluation

The Luminos program is an accelerated learning program that teaches children basic reading and numeric skills and supports socioemotional development. Luminos focuses on children who have never been to school or have been out of school for several years. The program uses a learning approach that is adaptive to individual learning levels, offers diverse interaction and play-based learning, and is centered around intensive structured instruction. Luminos does not charge any fees and provides free daily lunch.

Our impact evaluation consists of a randomized controlled trial (RCT) that compared out-of- school children (OOSC) in 50 treatment communities where the program was offered in the 2022-23 school year to 50 control communities where the program was not offered. We assessed 1502 OOSC at baseline and endline on literacy (using the Early Grade Reading Assessment, or EGRA) and numeracy (using the Early Grade Math Assessment, or EGMA). We administered an extended survey that included a module on socioemotional learning (SEL) to 324 OOSC per community. We also assessed 348 government schoolchildren (GSC) in grades 1, 2, and 3 from the nearby primary school (and administered the SEL module to 3 GSC) in every community to provide a benchmark for learning gains in the OOSC sample.

The results show large, significant learning gains on all tasks in literacy and numeracy in the treatment communities compared to the control communities.

On average, treatment OOSC were able to read 29 words per minute compared to 7 words per minute for control OOSC at endline. Treatment OOSC correctly answered twice as many addition questions and twice as many subtraction questions than control OOSC. 93% of children in our treatment sample who were offered the program attended at least some classes, and so results are similar for children offered the program (intent-to-treat, or ITT estimates) as for children who attended the program (treatment-on-the-treated, or TOT estimates). Effects were similar in size for girls vs boys, younger vs older children, children who were previously enrolled in school vs dropouts, and children who started with lower baseline learning levels vs higher baseline learning levels. We also observe positive learning gains in all treatment communities that exceed learning gains in almost all control communities, indicating that the program is having positive impact in all communities where it is implemented.

Compared with government school children in the same communities, children in the Luminos program started the school year with much lower literacy and numeracy scores, but ended the school year with similar numeracy scores and substantially higher literacy scores than their peers in school.

We find few significant treatment effects or changes over time in the SEL results. However, we expect that these null results are due to the validity of the research instrument in this specific context rather than reflecting a true null effect of the program.

Our results show that the Luminos program has large impacts on foundational literacy and numeracy skills for out-of-school children after ten months. Effect sizes are on the upper end of effects measured in RCTs of other remedial initiatives and structured pedagogical programs. We plan to conduct a follow-up round of data collection in one year to assess the persistence of these large learning effects and to assess whether students in treatment communities have successfully enrolled in government schools.