Testing Rori at scale: a new RCT in Ghana

Rori is Rising Academies’ AI-powered maths tutor, designed to give students personalised practice and support through WhatsApp.

Previous research in Ghana, led by researchers from the University of Oxford and J-PAL, has given us promising evidence of Rori’s potential. In an eight-month randomised controlled trial involving around 1,000 students across 11 schools, students using Rori for two 30-minute sessions each week improved their maths performance by 0.36 standard deviations compared with the control group.

The next question is whether we can achieve similar results at substantially greater scale, and in government schools.

This academic year, we are launching a new randomised controlled trial across 100 public primary and junior high schools in Ghana, with support from the Learning Engineering Virtual Institute (LEVI), IPA and the University of Ghana.

The study will involve 50 treatment schools and 50 control schools across five districts. Students in Grades 6–8 in treatment schools will have access to Rori over the course of the academic year, while students in control schools will continue with business-as-usual teaching.

Baseline assessments will take place in September and October, followed by endline assessments in June and July. Student learning will be measured using the standardised Star Maths assessment.

The study is designed to answer a straightforward question:

Do students with access to Rori make greater progress in maths over the course of an academic year than students receiving business-as-usual instruction?

We will use these results to inform our next steps, understand what is required to deliver Rori effectively at scale, and share the findings broadly to help strengthen the global evidence base on cost-effective AI-powered tutoring.