Computerized adaptive literacy learning

In this chapter, we introduce computerized adaptive practice (CAP) and its implementation in adaptive learning environments for practicing learners’ native language, Dutch and English as a Foreign Language. In CAP, the difficulty of items or exercises is adjusted to the ability of the student. This allows for deliberate practice, and individualized intensive training with direct feedback, which is known to enhance expertise development. CAP can be implemented using techniques developed in computerized adaptive testing (CAT), which is based on item response theory. However, several limitations of CAT must be solved. CAT assumes known item difficulties, thereby requiring pretesting item banks. Since the learning environments Language Sea and Words & Birds contain tens of thousands of items, pretesting is impossible. Furthermore, practicing low levels of success would lead to high drop-out rates making item selection techniques used in adaptive assessment not suitable. We solved the pretesting problem by using an algorithm based on the Elo rating system and allow for high success rates by incorporating response times in the measurement of ability. After describing the algorithm, we sketch the setup and working of some learning environments in which the algorithm is implemented. The games target different language domains such as spelling, vocabulary, reading fluency, and comprehension, but also working memory. We discuss some scientific results obtained from analyses of the data from the system, and end the paper with reflections on potential future directions for research.