Using Hidden Markov Models to Characterize Student Behaviors in Learning-by-Teaching Environments

Using hidden Markov models (HMMs) and traditional behavior analysis, we have examined the effect of metacognitive prompting on students’ learning in the context of our computer-based learning-by-teaching environment. This paper discusses our analysis techniques, and presents evidence that HMMs can be used to effectively determine students’ pattern of activities. The results indicate clear differences between different interventions, and links between students learning performance and their interactions with the system.