Optimizing Government Support to FLN Programs in Rural Areas

• In Event: Learning in groups, learning in loops using data for development: Evidence-based action and implementation-informed iteration

Proposal

Meerkat Learning supports the Namibian government to scale-up a foundational literacy and numeracy remediation program nationally. In partnership with uBoraBora, we launched a rapid impact assessment (aka A/B test) on different implementation strategies for teacher coaching, specifically targeting the most rural and marginalized communities.

Meerkat Learning specializes in implementation research that uses a quasi-experimental A/B testing methodology. A/B testing is an experimental methodology traditionally used by the technology sector (Kohavi et al., 2020; King et al., 2017; Siroker and Koomen 2015). We have adopted the Youth Impact approach to A/B testing since they have used it successfully for the past six years on Teaching at the Right Level (TaRL).

A/B testing is procedurally similar to a randomized controlled trial (RCT) in that we randomly allocate participants to different groups; because allocation to the intervention type is random, we can interpret differences in outcomes between the groups as caused by the intervention. While many RCTs compare a “no program” pure control group to an intervention group, A/B tests usually allocate two treatment groups – A and B – to the same main base program and introduce a small program variation in one group. Additional distinctions between RCTs and A/B tests include running multiple tests in rapid succession rather than one high stakes test every several years. These tests help answer the question “what works most optimally” – often the most relevant question to scale – not just “does the program work.”

Our A/B testing question is: Can government education officers provide remote teacher coaching that is equally effective to in-person school visits? The challenges teachers face while adopting new pedagogies are predictable–there is an initial increase of workload, decrease in teacher confidence, and demand for instructional coaching (Fullan, 1982). However, Namibia, similar to many countries, does not have the structures for education officials to provide FLN instructional coaching to teachers; namely, a 1:15 ratio of education officers and transport for 3 classroom visits (Piper et al, 2018). Therefore, with uBoraBora support, we are studying cost-effective alternative instructional coaching strategies.

Our 2023 A/B pilot test revealed that education officers can deliver 1-on-1 coaching phone calls with teachers and these 15 minute coaching calls reduced student attrition by 23% and increased learning gains by 25%. The current trials compare the use of low-tech conference calls, formal progress report letters, whatsapp groups, peer coaching, and the use of automated nudges. We will present the results of these trials including comparative cost, effect on learning outcomes, and feasibility for delivery at scale. The results will inform practitioners and contribute to the field of implementation research.

Author

• Angelica Towne Amporo, Meerkat Learning