Assessing Learning Behavior and Cognitive Bias from Web Logs
Bailey Braaten (The Ohio State University)
Co-authors: Rashmi Jayathirtha Rao (The Ohio State University), Christopher Stewart (The Ohio State University), Arnulfo Perez (The Ohio State University), Siva Meenakshi Renganathan (The Ohio State University)
Students who can link algebraic functions to their corresponding graphs perform well in STEM courses. Increasingly, early algebra curricula teaches these concepts in tandem. Tests, video analyses, interviews and other traditional methods that aim to quantify how students link concepts taught in school requires lot of time. In this paper, we use web logs collected by a smart classroom web server to infer learning as they are widely available, voluminous and amenable to data science. Our approach partitions the web interface into components related to data and graph concepts. We collect click and mouse movement data as users interact with these components. We used statistical and data mining techniques to model their learning behavior. We compared our models with traditional methods to assess learning behavior for a workshop presented in Summer 2016 in which participants were middle-school math teachers planning to use this curriculum created by The Department of Teaching and Learning at OSU in their own classrooms. As a part of this, they interpreted certain experimental observations using our web based e-learning portal. We used our models to assess participation levels, a prerequisite indicator for learning. Our models aligned with ground-truth traditional methods for 17 of the 18 participants. We also compared our complete models to simple models that eschew statistical machine learning. The number of interactions and the time spent with the two partitions of the web portal have been used to infer either data or graph oriented cognitive bias in the participants.