ABSTRACT
This chapter describes an Early Warning System that accurately and consistently identifies at-risk students before or during a semester by assessing their progression through foundation-level courses at a large, metropolitan university with a significant online course presence. The study predicted student risk using variables from both the student information (SIS) and learning management systems (LMS). Hierarchical logistic regression solutions were developed as the baseline using SIS data. Then, weekly progression models were constructed using combined SIS and LMS data to validate the consistency of the initial findings. Training was based on student data captured during Fall 2017, followed by test data from the Fall 2018 semester. The results at each week were compared, demonstrating that baseline models can predict student risk with a high level of precision at or before the semester begins and consistently throughout the semester. Both the accuracy and consistency of the baseline model were confirmed using a weekly iteration procedure about factors leading to the reduction of students’ academic risk. Although the proposed initial model can be developed based on SIS data alone, the consistency of this system can be improved with weekly progression analyses. The findings of this study can be beneficial for faculty and others who wish to develop intervention support mechanisms and resources that help at-risk students.
