Predictive Analytics for Online Education

Updated article originally published June 15, 2016.

A number of students balance their online education alongside work and family commitments. These factors, along with the lack of in-person interaction that defines online courses, often contribute to a loss of motivation in continuing with studies. Fortunately, an increasing number of institutions are using predictive analytics to change the culture and experience of online learning.

Predictive analytics alerts help educators to see the data that flags students as at risk. With this data in hand, instructors are better equipped to offer academic assistance and resources to students who need them.

Why Predictive Analytics?

Institutions that have shifted towards a culture of early warning and feedback have noted gains in retention and graduation.

In one experiment with predictive analytics Strayer University significantly increased online student retention. By identifying at-risk students and providing them with highly targeted and personal support, the overall number of students considered to be at risk within the experiment decreased by 17%.

Similarly, Georgia State University states that Black, Hispanic, first-generation, and low-income students graduated at rates at or above the rate of the student body overall since the widespread use of predictive analytics at the institution began.

These institutions are showing that with a focused effort, notable success can be achieved. Individual instructors can implement smaller scale analysis and begin the cultural shift needed for massive gains.

Classroom-Level Predictive Analytics: What Data is Important?

In a traditional class, an instructor is able to identify at-risk students through a number of factors, including in-class participation and assessments. The data provided by online student activity provides similar data, allowing instructors the ability to motivate and support students the moment they show signs of struggle.

Using information from an LMS or VLE, instructors can assess student progress. For instance, Strayer University is tracking how much time students spend on assignments, watching lectures, accessing optional resources, and posting on discussion boards. These insights provide a holistic view of entire classes and individual students, allowing instructors to provide personalized support.

When thinking about what data to watch, focus on engagement. How much time students spend with course materials is important. Alongside that information, traditional metrics such as attendance, class participation, and grades will help create a full picture of whether students are likely to be successful in a course.

Considerations for Intervention

Once you have established what data to base your interventions on, you will need to decide how best to improve your students’ chances of success. University of Central Florida’s Teaching Online Pedagogical Repository suggests a few steps for communicating with online students whose behaviour has been flagged.

First, think about when to message students. Instructors commonly connect with students during course milestones, including at the end of the first week, shortly after a major assessment, and during the final week to withdraw from classes. Picking the right times to discuss student progress ensures that course success is already top of mind for the student.

Next, decide whether your potential method of contact aligns with the student’s current behaviour. For example, if the student has not been logging into the LMS or VLE, sending a message via that medium is unlikely to have an impact. In this situation, sending an email may be more effective. In more severe cases, instructors may want to consider setting up one-on-one conversations with the student to review course expectations and concerns.

Randy Weinstein, Vice Provost for Teaching and Learning at Villanova University, suggests that the personal touch is key. While students may not respond to an automated message, having an instructor reach out personally is likely to garner a better response. Be thoughtful, and personalize your message where appropriate.

Finally, tailor your messaging to the behaviour of concern. For instance, if the student is receiving low marks, you may want to message them about tutoring services or remind them of office hours and resources that are available.

When thinking about reaching out, remember that positive reinforcement is as important as constructive feedback. Reward successful and improving students with encouragement.

Learn more about intervention and personalized instruction:

  • Buttons on an old machine with letters of the alphabet

    Focusing on Giving Online Students Feedback

  • Two students discussing work at a table with instsructor

    Assessment + Personalized Instruction = Learning

  • Student concentrating hard while studying

    How to Help Prepare Students for Exams