Proceedings of the Annual Meeting of the Cognitive Science Society

250 million children worldwide lack basic literacy and numeracy skills, many of whom have no access to regular schooling.Inexpensive tablet computers have the potential to scale up the distribution of intelligent tutoring systems to children inneed. We introduce a collection of tablet games presenting core literacy and numeracy concepts in a way that enablesself-directed learning, reinforced by a shared content engine with an adaptive algorithm that re-prioritizes content basedon the accuracy and timing of the learner’s responses to effectively space and distribute practice. The difficulty of eachgame’s dynamics adjust to the learner over time. We analyze response data from school children in Tanzania, examininghow they distribute their attention across the games and as a function of performance within each game. We also evaluatedifferent methods for determining their knowledge state and learning progress based on their responses, and examine howself-direction influences stimulus spacing.