Poster: "Representing and Predicting Student Navigational Pathways in Online College Courses"

Event: 2018 School of Education Research Poster Celebration
Date: Friday, September 28, 2018
Time: 3:30-5:00 pm
Location: School of Education Courtyard
 
Presenter: Renzhe Yu
Poster Title: "Representing and Predicting Student Navigational Pathways in Online College Courses"
Poster Advisor: Mark Warschauer

Abstract

Understanding and predicting how students navigate through course space is crucial to improving instruction yet challenging in educational research. Building on prior research on MOOCs, this study investigates students' navigational pathways by fitting neural network models on clickstream data of an online college course for residential students. We first learnt vector representations of resource pages from students' visiting sequences. Comparing their locations in the space to pre-designed course structure, we found that students who got different final grades exhibited different levels of adherence to the designed sequence. Next, we used a neural network architecture to predict the next page that a student visits given her prior sequence of visits. The highest accuracy reached 50.8% and largely outperformed the frequency-based baseline of 41.3%. These results show that neural network methods have the potential to help instructors understand students' learning behaviors and facilitate automated instructional support.