Analytics can help universities better support students’ learning

Advancing technology and digitalisation have resulted in universities experiencing a “data explosion”, much like what is happening in the business world.

Just consider the amount of data that universities collect when applicants fill in their application forms, when students interact with the learning management system and participate in academic and professional activities, and when universities measure the students’ performance.

What can universities do with these data about student demographics, prior educational background, study behaviour and academic performance?

First, they can “mine” educational data to uncover unknown patterns and determinants of academic performance in them to facilitate teaching and learning.

Let me illustrate how this would work.

For a start, universities can use the data to generate predictions of students’ performance, which are then fed into early alert systems with other information such as the students’ current semester study load and their previous semesters’ cumulative grade point average.

These can, in turn, provide information for faculty members to formulate intervention strategies to support individual students in their learning.

For example, students who are expected to perform well in the semester can be further encouraged to not only achieve their potential but to surpass it.

On the other hand, students who are expected not to perform as well can be advised to adopt good study habits and strategies and seek help early when they experience difficulties in their learning.

These can contribute towards a positive learning experience and supportive learning environment.

Incidentally, students’ perception of a caring learning environment is also a strong motivator for them to perform better.

As a case in point, the Purdue University in the United States has implemented a Signals system where its lecturers can provide learning interventions by posting a green, orange or red traffic signal to individual students accompanied by an email or text providing academic advice.

This has resulted in positive outcomes such as better student performance and retention, as well as better communication between students and faculty.

In the United Kingdom, the Open University has experimented with weekly predictive analytics reports on selected students.

These reports have helped the lecturers implement a more systematic monitoring of students’ learning and in determining the appropriate timing for interventions.

At home,  the Singapore University of Social Sciences (SUSS) has identified determinants of academic performance, such as pacing of study progression, having a study plan, being disciplined and focused, and engaging with online materials throughout the semester.

Several determinants of academic success stood out from the analysis. Learners who are able to achieve a cumulative grade point average of more than four out of five include those who take an average of four to five modules a semester, clear their core modules during the early part of their studies, access their online learning materials earlier and regularly.

In addition, learners with more working experience have benefitted from the applied learning approach of the university.

These insights are then translated into student advisories that are exposed to students at opportune times, such as at student orientation events and during course registration.

The advisories are also disseminated among the study materials and at lift lobbies on campus.

The advisories include tips such as not overloading on credit units, spreading out core university courses over more semesters and accessing online course materials earlier.

These prompts may sound “common sense” to some but for the majority of SUSS students — many are working adults who have left the mainstream education system for a few years — these are valuable reminders. 

The schools, with the early support system, also identify at-risk students to advise and support them with peer mentoring and customised coaching.

The use of learning analytics in universities is, of course, not without limitations. Providing student learning support is not just about deploying models and findings. There is a human element to the interactions between lecturers and their students.

Fundamentally, lecturers have to be able and willing to provide a caring and nurturing learning environment. Given this, they can then use analytics to help them do a better job.

In applying educational data mining and learning analytics to provide learning support to students, there are two risks that high education institutions have to bear in mind.

First, an overly aggressive use of such techniques can be overbearing for students. Second, there is a danger of adverse predictions/expectations leading to self-fulfilling prophecies.

To avoid these, it is important to be mindful that analytics is a tool that bears productive results only when appropriately used.

Hence, care should be exercised in ensuring that student advisories are handed out in a gentle and persuasive manner while early alert systems are used in a non-threatening but supportive and nurturing way.

In these ways, students would not feel that they are placed under a microscope or be stressed. Instead, they are likely to respond positively when they know that their university and faculty care about their learning and progress.

Finally, students too need to do their part in a world where self-directed and lifelong learning is a necessity and not a luxury.

Hopefully, lecturers and students — together with technology and tools (including analytics) — can work towards better teaching and learning. However, as the saying goes — give a man a fish and he eats for a day; teach the man how to fish and he eats for a lifetime.

Universities can facilitate learning but students have to do the learning.

ABOUT THE AUTHOR:

Professor Koh Hian Chye is the Director of Business Intelligence and Analytics at the Singapore University of Social Sciences. He has wide interest and experience in the areas of institutional analytics, learning analytics, data mining and business analytics applications.