Categories
Other HEQCO Staff

On Our Radar – An algorithm for student success

On Our Radar features HEQCO staff and guest bloggers offering their unique perspectives on trends, new ideas and hot-button issues in higher education. The opinions are those of the authors.

 

For years, postsecondary institutions have collected massive amounts of data but have not used them to their full capacity. Postsecondary institutions have a vested interest in improving student experiences, completion times and retention rates, and if they don’t, they should. While there are no clear solutions to these problems, data mining and academic analytics, which tend to draw on administrative and other relevant student data, are being used by some postsecondary institutions and appear to be a step in the right direction.

Applications that rely on data mining and analytics can be used by administrators, instructors, advisors, students and others. How can students who are registered and about to start their postsecondary program benefit from data mining and academic analytics? In my opinion, one of the best applications is the virtual advisor. Although similar to traditional advisors that guide students through their postsecondary experience, it’s not meant to replace them but more to assist them and the students.

For example, Arizona State University provides a wide array of services and supports through its eAdvisor, which tracks students’ progress during their first four semesters, helps plan course selection and alerts students and advisors if a student is not on track. Students deemed off track must see an advisor prior to registering for any additional courses.  Should they remain off track for two consecutive semesters, they may be required to switch majors.

Purdue University’s Course Signals predicts student success in a course by combining information about their grades in the course, their time on task and past performance. Each participating course provides students with a colour based on these three elements. Green signals that students have a low risk for poor performance, yellow signals that students have room for improvement and red signals that students have a high risk for poor performance. Unlike eAdvisor, Course Signals is heavily dependent on the cooperation of course instructors as they decide how often to run signals in a course.

Although just scratching the surface of data mining and academic analytics for higher education, these are innovative approaches to the on-going institutional challenge of increasing student success. Some may be wary of the Big Brother aspects of applications such as eAdvisor, to which I would respond, show me something better that actually works.

While applications such as those at Arizona and Purdue are perhaps more invasive than a traditional advisor, they do create positive results. Increased student success including course grades and retention rates have been reported by multiple institutions. Innovation seems to be the buzzword of the day and while most of us just talk about it, the examples above show that some are actually doing it. We have far too many talkers in this sector and we should be embracing the doers.  If there are problems with the techniques, let’s offer some solutions rather than just simply criticize.  We need not look too far to see where that has gotten us.

– Lindsay DeClou, Research Analyst

One reply on “On Our Radar – An algorithm for student success”

In the post above, DeClou notes how virtual advising programs identify and support those students most at risk of poor academic performance, and recommends that other universities adopt such approaches.

In fact, the Faculty of Arts & Science at the University of Toronto is in its sixth year of the Early Alert Program, an advising approach similar to the Course Signals program highlighted in DeClou’s article. In late December, Arts & Science advising staff work with instructors and TAs to compile information on students’ performance in targeted first-year courses. This information is designed to indicate a pattern of challenges, rather than a single failed test or assignment, and the patterns are used to place students into appropriate ‘alert’ categories indicating their relative academic risk (from Extreme to Low). The program casts a wide net – beyond the students who might be seen as in serious trouble – to alert students to the services and advising offered, and to let them know advisers are there to help. In total, some 1000 students (approximately 20% of first-year students) are identified through the program each year to receive proactive personal contact from advising staff. The urgency of the message they receive is calibrated to the apparent magnitude of the student’s difficulty, varying from ‘Come in right away!’ to ‘We urge you to come and see us’ to ‘You might want to consider these available supports, and we are here if you need us.’

As in the programs DeClou profiled, the Early Alert Program at the UofT is considered a successful and important advising initiative. The Early Alert Program is designed to help students find their feet in university studies sooner than they might do if left on their own, and to communicate that the Faculty and its instructors and advisers are interested in their success and well-being. The students appear to appreciate this. The advising support provided by the Early Alert Program, combined with the academic support resources to which advisers direct students, help the Faculty of Arts & Science to maintain its very high retention and graduation rates: the year-to-year retention rate in the Faculty is usually between 92%-94% and the graduation rate over 6 years is roughly 85% – enviable by Canadian and North American standards.

The Early Alert Program is therefore an established, Ontario-based example of the innovative academic support initiatives DeClou identifies.

–Glenn Loney, Assistant Dean, Faculty Registrar & Faculty Secretary, Faculty of Arts & Science, University of Toronto

Leave a Reply

Your email address will not be published. Required fields are marked *

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.