We will present first results from a project in which we test whether and under which conditions learning data from students' self-study with an adaptive digital retrieval practice programme (MemoryLab.nl) is a reliable predictor of subsequent test performance. As part of this project, we extended the existing computational model of fact learning (Memorylab, van Rijn et al., 2009, Sense et al., 2016) to make predictions about long-term retention of practiced information, based on the responses registered during adaptive retrieval practice. More specifically, we propose a mastery criterion that can be reached during practice, based on a cognitive model of the memory strength after a delay of 24 hours. We will use this to determine at what point learners have practiced an item sufficiently. The to-be-presented first experiment will test to what extent reaching mastery during self-study sessions indeed predicts subsequent recall, compared to no-mastery and compared to other descriptors of the practice phase (such as the number of (accurate) recalls, repetitions, or changes in response times over the course of practice).

van Rijn, H., van Maanen, L., & van Woudenberg, M. (2009). Passing the test: Improving learning gains by balancing spacing and testing effects. Proceedings of the 9th International Conference on Cognitive Modeling, 110–115.
Sense, F., Behrens, F., Meijer, R. R., & van Rijn, H. (2016). An Individual’s Rate of Forgetting Is Stable Over Time but Differs Across Materials. Topics in Cognitive Science, 8(1), 305–321. https://doi.org/10.1111/tops.12183