Time to retire Big Brother from the learning analytics debate

The  question of ethics arises in most discussions about learning analytics. It came up in Eric Bohms’ blog where he called for the sector to take care in respecting and complying with the Government’s draft Investigatory Powers Bill, and it comes up continually in Jisc’s learning analytics work. The Higher Education Commission was also cautious about the opportunities data brings in its report ‘From Bricks to Clicks’.

But I believe that the discourse on learning analytics needs to move on from Big Brother. I have real concerns that those of a more conspiratorial outlook, with the best of intentions no doubt, may inadvertently be delaying progress rather than supporting the safe and considered development of what may prove a powerful approach to fostering student success.

I see three key arguments in making a robust ethical case for learning analytics.

First, there exists an undercurrent in which learning analytics and ethics are painted as mutually exclusive. They are not. It is possible to carry out learning analytics ethically and in compliance with the law.

Universities must abide with the Data Protection Act in all aspects of their business, including learning analytics. While pastoral care of students by mutual consent has been part of many universities’ offer for some time, learning analytics is not an excuse for snooping on students or collecting more data than is necessary.

Universities need to implement information management policies and procedures for learning analytics as they do in other areas involving handling sensitive and personal data. Whether that is by implementing an entirely new policy, or including learning analytics as part of an existing one, is up to the individual university.

Second, the collection, use and management of personal big data is not an issue confined to our sector. Higher education is arguably lagging behind other sectors in leveraging such data. In banking, publishing, retail, even healthcare, big data is used to understand customers and to innovate and improve services.

For example, in Singapore healthcare providers dig into data and analytics to better understand a patient’s condition, lifestyle choices, work and home environment. As a result, they can create personalised treatment plans tailored to individuals’ needs. Learning analytics offers these same opportunities.

Third, universities are already collecting substantial data on students, but often in a siloed manner, geared towards addressing internal administrative imperatives. Surely universities have a moral obligation to actively pool and analyse that data, for the benefit of students, as part of their duty of care?

Imagine a student has not accessed the VLE, been to the library or engaged in the university community for a number of weeks and has missed their last couple of deadlines. Data about that is already being collected, but without learning analytics, it is typically not collated or acted upon. Such behaviour may indicate a problem, one that a timely human intervention by a personal tutor or similar could mitigate. In some cases, this could prevent an individual student from dropping out of his or her course.

Using analytics techniques, this existing data (plus other new data if appropriate) can be rolled into a predictive model of student success or failure. For example, as outlined in our international review, a model at the New York Institute of Technology successfully identified 74 per cent of students who subsequently dropped out as having been ‘at risk’.

Could some of these drop-outs have been prevented if an appropriate intervention had been staged? My feeling is yes. I also believe any university which states that student success and wellbeing is core to its mission has a clear moral obligation here. As students increasingly assert their power as consumers in the marketplace, it will be interesting to see if any legal dimension emerges in the UK under consumer law, duty of care, or otherwise.

In the near future, I can see students becoming the driving force for learning analytics executed for their benefit, and under the guidance of NUS. If universities have a duty to learners, to offer an exceptional experience, support learning, and mitigate the risk of non-completion, then learning analytics can be central to delivering that. It’s only a matter of time before students, as fee-paying consumers, come to expect the sort of personalised, high-quality experiences that learning analytics can underpin as a matter of course.

Looking further ahead I would be surprised not to see similar demands for learning analytics by the Government. If the soon-to-be-released White Paper takes up the recommendations of the Green Paper, we should expect to see a big increase in the use of metrics, including the introduction of the teaching excellence framework (TEF). Although the details are still to be confirmed, such an initiative will depend on universities being able to demonstrate and measure exceptional teaching and engagement. A national learning analytics service could help them to do so, and form an important part of this ecosystem.

With learning analytics offering so much to so many, and there being a strong ethical case for its advancement, let’s hope we can finally put Big Brother to bed.