Guided authentic scientific inquiry activities give students a clearer picture of the nature of science and how fields operate in practice, but they are difficult to do well. Virtual environments can facilitate such instruction by having built-in scaffolds such as feedback that reacts dynamically to learners’ actions. These methods of feedback are based on real-time automated assessments of student learning while using technology-based curricula. There is sparse experimental research on how formative feedback can best be deployed to students in open-ended immersive environments compared to more constrained tutoring systems or games. Additionally, teachers must be included in the design and implementation if the systems are used in complex classroom settings. This talk describes novel analytical techniques used to model and predict student learning based on logged actions in the virtual world of EcoXPT and how these models might inform scaffolds designed to give groups just-in-time support.
Joseph Reilly is a doctoral student at the Harvard Graduate School of Education. Prior to that, Joe taught middle school science and high school chemistry. His research interests include technology-based learning environments, stealth assessment, science education, and learning analytics.