Learning analytics is increasingly in the news. Recently the Higher Education Commission launched their Bricks to Clicks report, which demonstrates the enormous potential for the use of learning analytics in UK universities. It also highlights the dangers around the potential misuse of students’ personal data.
Learning analytics involves deciphering trends and patterns from data about students and their learning activities. It’s being used to provide more personalised learning pathways, to enhance the curriculum and to identify students at risk of attrition.
Last year I put together a list of 86 ethical, legal and logistical issues about learning analytics that were concerning institutions. Many of these were holding up developments at universities and colleges, and denying students the potential benefits of better and more personalised advice and support.
So we worked with the sector and with the NUS to develop a code of practice for learning analytics. The document aims to ensure that the many ethical and legal issues are addressed by universities and colleges. Its underlying philosophy is that learning analytics should be carried out for the benefit of students. The code is now being used as a checklist for the issues they need to think about when rolling out a project.
Let me share with you a series of podcasts I’ve recorded to provide more in-depth guidance on some of the difficult, but often fascinating, legalities and ethical dilemmas to do with learner analytics. I’ve interviewed staff at a number of universities which are fairly advanced in their use of student data for analytics, so they’ve already had to consider and address most of the issues.
Student consent for learning analytics
In the first podcast I quiz legal experts, researchers and practitioners at universities on one of the most controversial areas in learning analytics: student consent. Should you always ask students for consent to use their data? Some universities and colleges assume that their students have already granted their consent by signing computing regulations when they begin their studies. But is this enough?
Bricks to Clicks suggests that obtaining the informed consent of students is vital. But how many people actually read the terms and conditions that they sign? Meanwhile, what happens if students don’t give their consent to the use of their data? Can they still use the virtual learning environments? Surely this wouldn’t make sense without the constant collection of data on student activity?
Giving students access to their data
Are there ever circumstances where universities or colleges can or should withhold learning analytics from a student? If a student is deemed to be at risk of failure or attrition, do we have a moral obligation to tell them, or do we risk further demotivating them by giving them an ‘at-risk’ label? These are merely predictions of potential risk or failure, and we need to avoid creating self-fulfilling prophecies.
Meanwhile, if students have the legal right to see all the data held about them, should we simply provide them with a dump of everything, or does it have to be presented in meaningful formats?
Intervening effectively with students
Arguably, learning analytics isn’t constructive unless accompanied by interventions aimed at changing student behaviour or improving the course somehow. But how do you decide when to make an intervention with students based on the analytics?
There’s also a risk that we ignore personal circumstances when intervening with students. I’ve asked how we can ensure that students are treated as individuals and not as numbers. I’ve also discussed with the experts how likely it is that learning analytics will favour one group of students over another.
Stewardship of student data
How should universities and colleges look after student data to comply with ethical and legal requirements? The Data Protection Act requires us to look after learning analytics data in the same way as other data, but there are particular issues for learning analytics data. I’m grilling the experts again on whether we can use anonymisation to overcome some of the privacy issues, and I look at whether we should be outsourcing the stewardship of our students’ data to third parties.
We’ll discuss how universities and colleges are ensuring that they comply with key aspects of the Act such as minimising the amount of data they hold and not keeping it for longer than required - and whether they need to destroy all the data held about a student if they request that.
Institutional readiness for learning analytics
What factors need to be in place for an effective learning analytics project? Staff at the University of Edinburgh and the Open University discuss the objections that crop up amongst their colleagues about learning analytics, and how to overcome them. I’ve looked at which stakeholders need to be consulted, how policy is being developed, and what kind of expertise is needed to kick off an initiative in learning analytics.
We’re now rolling out our architecture for learning analytics, and working with a large number of institutions in the UK to develop their capabilities in the area. The same fears and objections crop up time and again in universities and colleges, but these podcasts show that they needn’t impede progress indefinitely.
There are concrete steps institutions can take to address the valid concerns staff and students have to ensure that learning analytics is carried out ethically and legally – and ultimately for the benefit of students.