Improving Early Warning Systems with Categorized Course Resource Usage

Abstract

Early Warning Systems (EWSs) aggregate multiple sources of data to provide timely information to stakeholders about students in need of academic support. There is an increasing need to incorporate relevant data about student behaviours into the algorithms underlying EWSs to improve predictors of student success or failure. Many EWSs currently incorporate counts of course resource use, although these measures provide no information about which resources students are using. We use seven years of data from seven core STEM courses at a large university to investigate the associations between student use of categorized course resources (e.g., lecture or exam preparation resources) and their final course grade. Using logistic regression, we find that students who use exam preparation resources to a greater degree than their peers are more likely to receive a final grade of B or higher. In contrast, students who use more lecture-related resources than their peers are less likely to receive a final grade of B or higher. We discuss the implications of our results for developers deciding how to incorporate categories of course resource usage data into EWSs, for academic advisors using this information with students, and for instructors deciding which resources to include on their LMS site.

Open learner models and learning analytics dashboards

“…A key way of doing so is simply to display the student model in the software's student-facing interface. The earliest of these interfaces did not use the term "OLM" [8,13,16,30,52]; this terminology emerged in the late 1990s.…”

Contrasting prediction methods for early warning systems at undergraduate level

“…Gašević, Dawson, Rogers, and Gašević, Danijela (2016) note that researchers have produced prediction models by using classification algorithms such as EM, C4.5, Naive Bayes Classifier, and Support Vector Machines. Logistic regression and multiple regression modelling are often used as prediction models (Macfadyen and Dawson, 2010;Waddington et al, 2016), with logistic regression being considered the most popular prediction method for educational settings (Marbouti et al, 2016). Hierarchical mixed models (Joksimović, Gašević, Loughin, Kovanović, and Hatala, 2015;You, 2016), K-nearest neighbor (Marbouti et al, 2016), neural network models (Calvo-Flores, Galindo, Jiménez, and Pérez, 2006), and decision tree methods (Azcona and Casey, 2015) are also methods employed.…”

Modeling Consistency Using Engagement Patterns in Online Courses

“…As a solution, count-based methods using frequencies of different online learning activities for early warning prediction have been studied [4,14,17,53]. However, they suffer from the same shortcomings of other count-based methods mentioned before and using the sequential structure has the potential to provide a more holistic support.…”