Evidence-Based Tutoring in Karnataka

The initial evidence 

The COVID-19 pandemic disrupted learning for over a billion children worldwide. In its wake, the urgency of addressing foundational learning gaps became clearer than ever. In response, we tested ConnectEd, a targeted, accessible, and cost-effective remote learning program that delivers one-on-one math instruction to children via basic mobile phones. A five-country randomized controlled trial (India, Kenya, Nepal, the Philippines, and Uganda) found consistent learning gains across diverse contexts, and indicated that the program delivers one of the highest learning gains per dollar spent of any education program. These results highlighted the model’s adaptability, potential for impact and scalability. 

Recently, the Government of Karnataka introduced Ganitha Ganaka, an adaptation of ConnectEd, in the 2024–25 academic year. The program targets students in Grades 3 to 5 in government primary schools with the goal of improving foundational numeracy. The program is set to achieve a few goals: (a) enable students to receive targeted tutoring support in numeracy (b) support teachers to deliver hands-on practicals (c) engage parents in their child’s education and (d) deliver effective interventions during “normal times” and be ready to support education delivery during school disruption. 

Ganitha Ganaka is delivered by government teachers who select students needing additional support and tutor them via weekly phone calls in six-week cycles. Each student receives focused instruction in basic mathematical operations; addition, subtraction, multiplication, and division.

Government teachers training

Strong results delivered

The program was initially piloted in Soraba and Raichur blocks, involving 80 teachers and 240 students. Results showed that 4 out of 5 students learned at least one new math operation. Based on this success, the program scaled to 93 aspirational blocks across 17 districts, training 35,000 teachers and reaching over 150,000 students with individual tutoring in the last academic year. A live dashboard has been developed to monitor results in real time. The government has also budgeted for the program and the Chief Minister has announced a state-wide scale-up.

Even as we scale, learning outcomes remain stable and strong. It’s proof that the model is scalable without losing efficacy, and that rapid scale is achievable through close government partnership. 

Students’ levels before and after the program, showing huge improvements at scale. 

Scaling and integrating with the government 

We are proud to announce that on 29th May 2025, Youth Impact signed a formal partnership with the Government of Karnataka. This partnership includes a mandate to provide technical assistance in order to improve systems for learning outcomes and better foundational literacy and numeracy status in the State.

Formal partnership

The signing of this formal partnership marks a significant step in aligning public education policy with rigorous evidence on what works to improve foundational learning. This partnership signing marks the sixth government Youth Impact has formed a partnership with. For us, these partnerships are the holy grail of sustainable scaling, because they demonstrate government ownership of evidence-based programming.

Exploring AI for efficiency and scale

Further innovations are underway. We are now testing AI-driven tools to improve cost-effectiveness and reduce teacher workload. This includes automated assessments using voice-based IVR systems, real-time chatbot scheduling, and voice-to-voice AI tutoring—all aimed at improving data quality and delivery efficiency.

The way forward

Ganitha Ganaka offers a compelling example of how evidence can move from research to large-scale implementation, bridging the gap between innovation and policy in the effort to ensure all children access quality foundational learning.

The implementation of this program is a collaborative effort. We appreciate the leadership of Samagra Shikshana Karnataka (SSK) in executing the intervention at scale, funding support from donors including The Agency Fund, and the field-level support provided by Alokit. We also recognize the contributions of What Works Hub for Global Education and J-PAL, whose involvement ensured that the program design and scale-up were informed by rigorous evidence and ongoing research.