Toward Learning at Scale in Developing Countries
summary
A 15-month controlled study of 2,000+ children in 168 Tanzanian villages tests tablet tutoring under six broken assumptions (no reliable power, internet, or one-per-child devices) using Solar battery stations, Pixel C tablets, local Wi‑Fi/FTP, shared‑tablet login software and RoboTutor with speech fallbacks, spoken prompts and automated placement. All five finalist apps outgain control by ~3×; treatment raises scores by 13.4 percentage points (effect size 0.82); children with literacy-oriented aspirations gain ~4.3 points; correct attempts predict gains while incorrect attempts predict losses.
AI-written summary · about
chapters (4)
- 0:00Start
- 0:04Lessons from the Global Learning Xprize Field Study
- 9:39Effect Size
- 13:05Remote User Testing
related talks
13mMobile Education App during COVID
21mSlow.tech: it's urgent to hack the system!
17mPredict, Prune, Play: Efficient Video Playback Optimization Under Device...
19mInvestigating Effects of Post-Selection Feedback for Acquiring Ultra-Small Targets on Touchscreen
44mTransforming Farmer's Lives Using Android in Kenya
36mLucky
34mReal world connectivity: Beyond the bubble
30mLet’s Be Actors of Sustainable (Mobile) Development