Research highlights
My research uses large-scale randomized evaluations to shape African development. My work transforms African schools into cost-effective hubs for agricultural technology diffusion, student-centered, experiential education, and routine experimental data collection—laying the foundation for equitable growth and structural transformation.
School-based agricultural extension (SBAE)
In collaboration with the NGOs AgriCorps and 4-H Liberia, I piloted and then led a large-scale randomized evaluation of school-based agricultural extension (SBAE) in Liberia (2021–23), securing $4,160,984 across 12 grants since 2019. SBAE tackles technology diffusion and school attendance together by integrating experiential science instruction, school farms as demonstration plots, and student home-farming projects. Given its success, it will scale to all rural schools by 2035.
Job market paper: SBAE's effects reverse with parental engagement
SBAE diffuses technology to rural households through two channels: school farms that demonstrate improved techniques to parents, and student home projects that test those techniques on the household's own soils. My job market paper shows that parental engagement—an informational component I designed, centered on a video clarifying how SBAE differs from typical "school gardens" (often associated with drudgery and student farm labor)—is critical to activating both. With it, SBAE is more cost-effective than most African extension programs at diffusing technology and twice as cost-effective as a widely cited, successful conditional cash transfers program at raising attendance. Without it, SBAE fails to improve technology adoption and sharply reduces students' study hours and attendance—the opposite of its goal. More puzzling, parents persistently avoid information that would resolve their uncertainty: over three years, despite 81% student participation, only 44% of parents visited school farms, and absent parental engagement, parents at SBAE schools reduced both school farm visits and parent-teacher meeting attendance (by 39%) relative to control. On their own, standard models of decision-making under uncertainty—such as Bayesian updating, selective attention, memory retrieval, prospect theory, or model misspecification—are difficult to reconcile with the joint patterns of information avoidance and strong responses to parental engagement. By integrating existing economic theories using insights from cognitive neuroscience, my ongoing work develops a behavioral framework that explains these results.
Scaling SBAE:
"Learning to learn" teacher training, routine data collection, and equitable structural transformation
One of SBAE's core goals is to offer a platform of experiential learning experiences in science and agriculture in environments where learning resources—such as textbooks and laboratories—are missing. Yet, most teachers in developing countries lack such experiences themselves, and typically respond to pedagogical interventions by memorizing fixed recipes of "experiential learning" instead of internalizing the spirit and adapting to the dynamic needs of their curriculum and students. My ongoing work with Vesall Nourani (Harvard GSE) integrates the highly effective, student-centered "Learning to Learn" teacher training into SBAE, unlocking SBAE's potential in experiential science learning to improve student learning outcomes while enabling rural communities to experiment with new farming practices.
Supported by a routine student data collection system I developed and lab-in-the-field experiments, SBAE further provides an efficient platform to test economic theories and generate experimental evidence that informs agricultural R&D design. By equipping rural populations with generalizable skills beyond agriculture and appropriate agricultural technologies, SBAE positions African countries for equitable structural transformation.
Looking ahead, I am designing three clusters of randomized trials for 2026–2030 to strengthen SBAE as hubs for rural development in Liberia, Côte d'Ivoire, and Kenya.