Patrick Lynch presented on the use of learning analytics at the University of Hull, which he defined as the process of developing actionable insights into learning and teaching which may be used to influence teaching delivery.
Hull’s approach is based on the creation of a statistical tool reporting on 47 different ‘events’ across their institutional eBridge virtual learning environment, which is run on Sakai software. By drawing out course reports from the institutional VLE (exported in Excel), the e-learning team can provide insights on student activity based on a sample of ten individuals from a large module. This data is delivered just-in-time, as the course is being delivered, enabling course leader to make changes and adjustments to the delivery of the programme if things aren’t going to plan. The statistics code student activity against 4 dimensions: simply logging in to the course site; accessing the reading pathway for the course – touching on the learning objectives for the week; access to course resources; activity on the discussion forum. Students who have displayed no activity during a specific week’s work are coded ‘red’. Whilst it is highly questionable whether this sort of data on its own can determine whether there are specific issues with the course design or issues with the commitment of individual students, the immediacy of the data enables instructors to reflect and take action – such as following up with students who are perceived to be struggling. It also offers an overview of what students are doing within their VLE module site -where they are going and the type of activities they are engaging in and the resources they are using. This may help instructors to judge whether online activity is matching their expectations of how students are meant to be working online. The weekly data can be useful in highlighting the most popular resources, and the reporting from the data can also present ‘temperature gauges’, comparing online and physical attendance for the module.
Building on the reporting formats, Hull are now looking at discourse analysis for discussion forums, and Patrick’s team have been developing Wordle representations to capture the behaviour and values which are being expressed by students in their contributions to online discussion forums. They are also looking at social network analysis to generate network diagrams for discussion forums, judging where the nodes of communication lie and the modes of interaction across a discussion activity. This is offering staff new insights into their teaching and the levels of student engagement within a task – for instance they will have a clearer picture of whether an activity is actually teacher-led or student-led, based on the nodes of communication and patterns of interaction – who is responding to whom and the types of responses that are being generated in the discussion.
Where next for learning analytics? Patrick noted that the commitment to learning analytics has been written in to the learning and teaching reform programme at Hull. It is seen as being closely associated with business intelligence work, which will be used to kick off a student retention project. Student experience officers will take the data and use it to intervene with students who appear to be at risk of withdrawal form their study programmes.