London, United Kingdom
Many tools on the Internet are used without thinking about the social repercussions they might have. Lots of tools are also used without reading the terms and conditions and understanding what they are doing with our data and our use of the helps. In teaching and learning applications, this is a very important and sensitive point. One of the tools that has a lot of skepticism and ethical issues is learning analytics and predictive analytics.
Matt Lingard is the Head of Technology-Enhanced Learning at the University of West London, and as soon as the university decided to implement a learning analytics program, he thought it would be essential to also apply a policy of use to go with the analytics program. One of the reasons why analytics generates an ethical debate is because a negative prediction on a student might have repercussions on his or her feelings towards the course, influence the teacher’s grade, or make the student seem as less than an individual and more as a statistic. It’s important to understand how to use analytics to create a code of conduct that supports teachers to make appropriate use of the insights from learning analytics.
Learning analytics is a set of data, measurements, and analysis of different sources of information (demographics, use of Blackboard Learn as an LMS, and the student record system) that show the student’s progression through the course and different topics they might have trouble with. This allows teachers to reach out to the students to help them where they need it and figure out what is wrong, and, ideally, prevent dropouts.
Learning analytics works with something called Big Data, which is a set of algorithms and statistics that can mathematically add up components and make conclusions based on them. For example, the amount of times a student enters Blackboard Learn to see his/her classes, assignments, and feedback can show high engagement. That tied with high grades, but a low in person attendance rate might sh hat the student wishes to keep studying but might live far away and it’s being a problem. The predictive analytics tool used at the University of West London examines if that student is likely to continue studying next semester. It produces a ‘persistance prediction’ based on the analysis of the student’s scores, attendance, demographic data, and engagement with the LMS. Since all this data is very important, but also very sensitive the University of West London published a policy based on 10 principles to have in mind when using learning analytics.
Three of the 10 principles are particularly important and, as Matt explains, all intertwine with the each other. These 3 principles are:
1. Consent is one of the key areas of the policy, and its informed consent more than anything. When students enroll in any university, they are giving consent for the university to use their information. However since this is a sensitive topic, the objective of the learning analytics must be clear about how and why the data is used. The students must understand that analytics is used for the benefit of them in order to help them in certain areas and to understand how well the school is doing in teaching and learning. The implementation of learning analytics has been gradual and there are only certain areas of the university are using analytics as of now: the School of Computing and Engineering and the London Geller College of Hospitality and Tourism. These two schools are the first to use Analytics, since the university wants to make sure that all staff are properly trained before starting to use the system before implementing it in all the faculties of the university.
2. Transparency and openness are both pillars of the policy and are very tied to the consent. They want to be transparent in that they want to inform everybody about exactly how their information is going to be used. This is one area that Matt acknowledges that the university has some more work to do. Although students have been notified by email and a related webpage the University needs to ensure that students and staff are fully aware; a more proactive campaign is needed.
3. Access is a huge part of the reason why the policy is created. It is important that the data is seen only by the people who need to see it. In the UK, there is a personal tutor role. Each personal tutor is assigned twenty students, and supports them in academic matters. The personal tutor, has all the information and the predictive analytics, their attendance records, and their demographics in order to have and informed conversation with all their students . There is also a bigger set of analytics available to the student engagement team and the senior management, which shows analytics for the school’s population as a whole, not of any individual student. This will allow the university to make strategic decisions based on insights for particular groups of students, for examplefirst year students or mature, students.
It was always important to implement this policy so that Analytics doesn’t interfere or influence grading, and that the aspect of the individual remains intact. This means understanding that there will be individuals who do not follow the typical patterns, and will not turn out as the predictive analytics says.
In order for this policy to be written, Matt led a team of twelve people who wholly represented the university: teaching staff, student representation, senior management, the Pro-Vice Chancellor for Education, and the University Secretary (the person who is responsible for the legal matters of the university). After Matt did the investigation, he realized that it was very simple to form a new policy that was modeled on the existing policy of the Open University, which had had a learning analytics policy for three years, and the advice that the Jisc (Joint Information Systems Committee) put out on the issue. Matt explains that he did not see the point in reinventing the wheel when he had two superb fountains of information that additionally had a creative commons license.
Matt also explains that all this work to create the policy was completely worth it because of the benefits that a learning analytics program can do for the student’s lives. The whole purpose of analytics is to use data to help ensure student success and supporting students who might be at risk of not continuing. Dropout rates are a very important topic in schools and universities around the world. Retention has always been a priority, and this helps in achieving that because the personal tutors can see a red flag when it presents itself and helps students achieve a different outcome than those who have had similar results because tutors usually have a very close relationship with their assigned students. Personal tutors can also use the learning analytics tool to connect to their students through messaging so that if a student has not shown up to class in a week, for example, the advisor can reach out to them and asks them what is wrong and try to get them to come back.
Overall, learning analytics is a great tool with huge advantages, which, like all things, has to be used wisely, responsibly, ethically, and with the full consent of everyone involved. It has potential risks, which, if used well, should not exist. The University of West London is a very important example to look up to for any educational institution which is planning to have or already has an analytics program in place.
*Matt Lingard, Head of Technology and Enhanced Learning.
*Photos by: AFP Glyn Kirk