Big data can transform learning – as long as lecturers take control …
Analysis of big data can provide valuable insights into how people learn. Photograph: Marcus Brandt
Big data has been eagerly embraced by the business world. Now it’s time to look at how it can be used in education.
The term refers to the trails we leave behind every time we use a website. We’re all familiar with sites such as Amazon suggesting that if we enjoyed one book, we might like another book on a similar topic. These recommendations are based on data collected from very large numbers of customers and as a rule work very well. It undoubtedly improves the shopping experience.
Until recently, such techniques could not have been used to improve the learning experience. Teaching in higher education has not typically handled thousands of students, but now we have massive open online courses (Moocs) – free courses that are open to all and aimed at an unlimited number of participants. Moocs often have six-figure cohorts, and students work through the same online activities.
Can we apply the principles we’ve learnt from business to online learning? By observing students’ patterns of behaviour online, could we help them learn more effectively?
Possibly, but it’s not that easy.
Unlike traditional research, the data collected from Moocs is not generated by a research question, but as a by-product of interactions with the system. So we have to consider what questions can be asked about the data.
This is not the usual order of things in science. The research question should determine the data gathered.
Asking the right questions
A methodologically respectable approach to big data is essential, as it can tell us many useful things.
On an ICT in primary education course I teach on Coursera, there are more than 2,000 participants from emerging economies. We know this group will need access to free ICT tools and resources in places where internet access is poor. That information affects the nature of the activities we design into the course. The platform will also tell us the rate of dropout each week, which activities participants spent most time on, which were ignored, and so on.
But that data can only tell us where to look to ask questions. It won’t tell us how to explain that behaviour or what to do about it.
Using data to inform the design-test-redesign cycle requires a design goal. The lecturer working with a class aims to achieve some learning outcome – and designs the activities, collects data on student performance, interprets it and so improves their design in order to better achieve the goal.
That’s all well and good, but this process does not generate big data. This is local data, tied to the specific learning design of the lecturer. It is not widely tested. There is no independent peer evaluation. It is not scientific.
Learning data does not have to be local
What is exciting now is that with large-scale learning platforms, such as the Open University’s FutureLearn, learning data does not have to remain merely local. If the learning design (the planning and management of learning activities) can be used by other lecturers all collecting the same data from their students, it can be widely tested, with many independent peers able to review, advise and redesign for a better outcome.
There are now learning design tools available on the web that enable lecturers to share their teaching ideas. If many lecturers run the same design through their local virtual learning environment (VLE), with students using the same digital tools to collect their performance data, we could crowdsource local data from designs shared on the large scale.
This kind of big data – crowdsourced local data – remains under the control of lecturers. It is lecturers who decide what kind of data they need to inform the future design of their courses.
Big data could improve teaching, but not without educators taking control of this extraordinary methodological gift. At present the field is being driven almost entirely by technology professionals who are not educators and have never taught online. Instead, we could be recruiting all lecturers everywhere to collaborate and generate their own large-scale data collection and analysis. Then big data could really make a difference.
Diana Laurillard is assistant director of Open Mode at the Institute of Education. Her most recent book is Teaching as a Design Science: Building Pedagogical Patterns for Learning and Technology. Follow her on Twitter @thinksitthrough.
Join the Higher Education Network for more comment, analysis and job opportunities, direct to your inbox. Follow us on Twitter @gdnhighered.
Visit source:
Big data can transform learning – as long as lecturers take control …
