Quasi-experimental study results on Kenya primary student learning outcomes

• In Event: Blended Learning for Professional Development of School Leaders and Teachers: a cost-effective approach to improving learning outcomes

Proposal

A three-year quasi-experimental study in Kenya found children may gain as much as half a year equivalent of additional schooling linked to leader and teacher professional development using a blended learning approach, and school improvement loans at their schools. Key findings included:

• In English Literacy, the study found that students coming from lower socioeconomic backgrounds stand to gain the most from the blended learning school intervention – gaining the equivalent of 0.53 years of school more than their peers throughout the study period.
• Similar outcomes were observed for girls, who gain 0.47 years of English literacy schooling and all students gain 0.26 years from the intervention.

To assess learning improvements throughout the three-year program with affordable non-state schools, a baseline was collected using the Uwezo learning assessment in early 2021 from a sample of 131 schools that had recently joined the program. The exercise included students from a further 70 control schools. This 2021 baseline found that just 50% of Grade 2 students could read a story in English at Grade 2 level, 49.5% could read a story in Kiswahili, and 41.9% could do basic division.

The schools continued to participate in the program through 2023 and as schools were completing the program, an external data collector conducted the endline to conclude the longitudinal quasi-experimental study. The endline included 4,173 students of similar grade levels.

Research questions included:
• Is there a difference in learning outcomes for students attending schools participating in the blended learning program?
• Is there a difference in learning outcomes for students attending schools that make infrastructure investments?
• Is there a difference for either that extends to girls and/ or students coming from lower socioeconomic backgrounds?

To define the socio-economic status of learners without directly interviewing each parent, the research team created an index for socio-economic groups during the study design process. Questions had to be asked of 8-year-old students that would enable the study to distinguish and estimate economic status. These questions included: How many people live in your house? How many rooms are in your house? What are the walls made of? Which household items (of a list) are in your family’s possession? The results were then stratified to enable analysis of each group.

The study’s positive findings may indicate opportunities for ancillary services providers in the education sector to further explore integrating blended learning into teacher professional development models aimed at improved learning outcomes and addressing equity, even in low-connectivity environments.

Author

• Jack Tait, Center for RISC @ University of Chicago