Highlighted Session: Testing What Works: Refining Coaching and Teacher Support Models Through Implementation Research

Group Submission Type: Highlighted Paper Session

Proposal

Coaching and ongoing teacher support are consistently highlighted as critical for improving foundational literacy and numeracy (FLN) outcomes (Ganimian & Murnane, 2016; Kraft et al., 2018; Moore et al., 2017). Yet these structured pedagogy components remain among the hardest interventions to scale. Evidence shows that well-designed coaching can strengthen teacher practice and student learning, but the costs of sustained support, the intensity of training required, and the challenge of maintaining quality across large systems often limit effectiveness at scale (Kraft, Blazar & Hogan, 2018; Piper & Zuilkowski, 2015, Bagby et al, 2024). Many FLN programs struggle with high turnover among coaches, variability in delivery, and the difficulty of embedding quality, consistent support within government systems. This is where implementation research adds value: rather than asking only whether coaching and on-going teacher support works, it investigates how coaching and on-going support can be adapted, improved, and sustained in specific contexts. By testing and refining models in real-world conditions, implementation research helps implementers identify context-appropriate approaches to scale teacher support more effectively, while simultaneously contributing to broader system learning.

This panel will share findings and learning from four implementation research projects centred on coaching and on-going teacher support. First, we’ll hear from Meerkat Learning about how they used A/B testing to compare approaches for scalable teacher coaching in rural Namibian schools. Then, Impact Network will present their findings on how they improved support for teachers in government schools by testing and iterating the use of deliberate practice in coaching conversations and checklists in Zambia. Then VVOB will share results from an A/B test on peer-mentoring solutions to support TaRL implementation in Uganda. Finally, FHI360 will share findings on how it used evidence to improve their on-going teacher coaching model by testing and tweaking the role of a school-based mentor.

All the implementing partners on the panel were part of the first uBoraBora cohort, a fund and learning network that uses implementation research to fuel continuous improvement. The name uBoraBora, meaning “better, better” in Kiswahili, reflects the shared commitment to ongoing learning and the pursuit of stronger foundational literacy and numeracy outcomes through testing and learning. Finishing in December 2025, these projects will generate insights not only on teacher support models, but also on what it takes to “tweak” interventions in practice. Two organizations, FHI 360 and VVOB, emphasized evidence-informed action - meaning that they began with investigation to determine what actions to take - while Meerkat Learning and Impact Network pursued implementation-informed iteration - meaning that they began with already formed hypotheses about ways to improve implementation and went straight to testing them. Together, their work helps unpack the difference between these two approaches to learning in loops and shows how implementation research can begin in different places to drive continuous improvement to overcome common obstacles to delivering coaching at scale.

They will share the findings of their research, highlighting how evidence has shaped the design and delivery of coaching and peer-mentoring, and how these insights have guided their pathway to scale. The panel will also reflect on the systems needed to capture and use reliable coaching implementation data, ensuring ongoing teacher support is responsive, effective, and continuously improved, as well as the key lessons learned in the process. Finally, Clio Dintilhac from the Gates Foundation will open the discussion as the panel’s discussant, facilitating a dialogue on lessons learned in using implementation research to improve coaching and peer-mentoring alongside a discussion on wider implications for FLN programming at scale.

Sub Unit

• Global Literacy SIG

Chair

• Lea Simpson, Brink

Individual Presentations

• Optimizing Remote Coaching via government support to FLN programs in rural Namibia - Angelica Towne Amporo, Meerkat Learning

• Testing in-school coaching for sustainable teacher on-going support at scale in Nigeria - Emily J Koester, FHI 360; Zahra Abubakar; Rachel M Hatch, FHI360; Elizabeth Adelman, FHI 360