The rise of online tools for delivering courses enables educators new opportunities to harvest and use data. Buckingham Shum and colleagues at the OU presented the ways that data can be used in education as three overlapping spheres;
- Customer relations management
- Business intelligence and
- Learning analytics
This post examines just one of these, the possibilities of data to support the learning process.
Examples of ways that data might be used;
- To diagnose those not engaging in their studies
- To diagnose those who have not done a critical task eg downloaded the assignment brief by a certain point
- To display to a student how they are doing in relation to the rest of the cohort
- To signpost a learner to resources to their particular needs.
Daphne Koller’s presentation describes the features of the Coursera platform which she has been involved in developing. The platform has many features which make use of large scale data to help to support the learning process.
- The platform automatically takes the top graded submissions and presents them as exemplars in a gallery.
- Seeing frequently occurring wrong answers and using them to provide targeted support for students (Koller, 2012, 28-31 mins)
- Feeding back to lecturers on common mistakes
- Understanding what the best students are doing and advising others appropriately.
- improving the usability of forums through enabling students to vote on the most useful questions, so that valuable questions ‘bubble up to the top’. In addition if a student is about to post duplicate question to one that has already been formed the students is pointed to this before they post. (28min)
She argues that using online learning in this way means that the face to face time in the classroom can be used for more engaging learning activities. The notion of the flipped classroom, where listening to videos occurs at home, and face to face time is focussed on interactivity is relevant. She provides examples from her practice eg using classroom time to focus on the questions that students got wrong – facilitated by the immediate feedback of MCQs, higher level discussion, real life examples from experts, problem solving in small teams, immediate feedback. (37 min).
Of course such advances raise more questions than they answer:
- there are ethical debates about what we capture about our students and how we make this available to them to their funders etc:
- there are questions about whether we use what is easy to measure rather than what is important or valuable
- once data is harvested and repurposed to make it meaningful the challenge is how do make sense of it. (Staff and students will need a new literacy in this)
- What roles do institutions need to enable them to generate learning analytics (the data wrangler)
- Ensuring that there is time and space in the curriculum to make sense and use LA
- Dangers of Senior Managers with a love of stats!
- Predictor systems closing down choice rather than opening up choice to students
- What pedagogic models will underpin developments in LA?
This is an issue of growing importance. Some blogs to watch
Sheila MacNeil CETIS