Building Teacher Capacity or Outsourcing Pedagogy? Experimental Evidence from Science Education Reform

We study how the mode of delivery shapes the effectiveness of curiosity-based pedagogy in low-capacity public education systems. Using a randomized controlled trial in 150 government middle schools serving 7,185 students in rural Uttar Pradesh, India, we compare two implementation models delivering identical, curriculum-aligned, inquiry- driven science lessons. In the teacher-led model, the incumbent science teachers receive intensive training and ongoing pedagogical support; in the facilitator-led model, trained external instructors deliver the same curriculum directly to students. By holding instructional content constant and varying only the delivery mechanism, the design isolates trade-offs between implementation fidelity, learning outcomes, and institutional integration. We measure impacts on students’ curiosity, critical thinking, growth mindset, and academic achievement, as well as changes in classroom practices. The study provides policy-relevant evidence on whether pedagogical reforms are more effective when they build internal teacher capacity or rely on externally delivered instruction in resource- constrained education systems in low- and middle-income countries.