Ben Williamson Seminar: Doing Educational Research in Web Science
The moment the notification email went around advising that Dr Ben Williamson was going to give a talk, I booked my place. As a ‘resting’ secondary school teacher taking a sabbatical to do a PhD, anything edu-related is always of interest. When it’s also embedded in Web Science, that’s a double bonus.
Dr Williamson has written extensively about ‘the digital age’ and education. Many publications are listed here (he’s currently a lecturer at the University of Stirling), and he has a web site where he writes extensively about ‘code acts in education’. The first blog post (from September 2013) explains what it’s all about
Ben’s talk was focused on a very new project that’s happening, which will eventually affect higher education (HE) in terms of data gathering and analysis. So far so boring, you might think, but it raises some big questions while addressing a genuine issue. Simply put, the focus on data and its importance has increased dramatically over the last few years. This has coincided with the arrival of ‘big data’ (data sets “…that are so large or complex that traditional data processing application software is inadequate to deal with them.” Wikipedia), research focusing on aspects of AI (Artificial Intelligence) and the introduction of Computer Science in primary and secondary schools. In short, it’s a hot topic.
HE institutions gather data about many things, for a variety of different purposes. An audit revealed that the whole system was, to put in bluntly, a mess and needed standardising and streamlining, with the student at the heart. This raises some immediate questions, which I’ll address in a minute. Incidentally, a report just released by the IHEP in the US (funded by the Bill and Melinda Gates Foundation) says “Accurate and complete data can empower college choices, promote student success, and inform federal, state and institutional policies. Yet existing postsecondary student data systems are disconnected, duplicative and incomplete.”
In 2014 HEDIIP carried out an audit and confirmed that things were far too complicated. Not all institutions collected the same data, in the same format, or organised and stored it in the same way. Inevitably, much of it is duplicated. Given that the student is now regarded, rightly or wrongly, as a ‘customer’, HE institutions need to attract and retain as many of these customers as they can. Access to the data will be provided via a ‘dashboard’ that will display the data in a way that’s informative and easy to read via interactive displays. So far, so good and laudable.
The whole project – worthy as it is – raises some big and important questions. There are very real concerns about data privacy and security. Students already have a unique student ID, but it’s probable that this system will eventually be able to track students from primary school onwards, as they are issued with a pupil identifier number at entry to full-time education. How will their data be used? How we can ensure that it will be used ethically? Who actually owns the data in the first place? The university? HESA? The government? The value of the data is not necessarily that which belongs to an individual student – it won’t make any difference if a few students opt to refuse to let themselves be tracked – but in the collective body of students, and especially when several cohorts have been through the system.
There’s also the question of how the data will be presented and interpreted. The need for training is acknowledged. Even though the data will be presented visually (and data visualisation is a big thing in data science), that doesn’t mean that it’s easy to interpret the information that’s being presented. Some of the finer points, or more subtle but nonetheless important nuances of the data may be missed. Jamie Bartlett (who presented on documentary on BBC2 recently about ‘Silicon Valley’) and Nathaniel Tkacz wrote a great article on this very subject, ‘Governance by Dashboard’. We all know that the current government has been warned more than once about cherry-picking data to support a position, when in fact the overall conclusions were very different.
Technical experts and/or data scientists are privileged in this new data-driven age. If that isn’t you, you may find yourself side-lined.
Some of the questions asked of Ben were naturally focused on the more challenging aspects of the data collection. A student missing a lecture may be emailed to ask why they didn’t attend, as a pattern of missing lectures may be a signal that a student is ‘at risk’. However, what if they were there? Or not there but had arranged a way to catch up?
I should explain that, as an ex-secondary school teacher, I am deeply cynical about anything that appears to hover big-brother like above teachers and, in the case of HE, lecturers and anyone responsible for delivering education. There are lots of notes in my notebook along the lines of data becoming a stick to beat people with. One of the constant battles in HE, it seems, is access to research grants. Will the data be used to rank faculties or staff with regard to the amount of funding they’ve been able to attract?
Ben himself acknowledged that this is a highly contentious space, and as such we should wait until the results begin to emerge. Certainly he presented the talk as a neutral observer, which is commendable. I know from conversations between edu-professionals on Twitter that any system that impacts education can provoke some pretty lively discussions, and there are plenty of people out there who are a lot more cynical than I am. Nevertheless, if the data is collected ethically, privacy and security are respected, it presented well and everyone is provided with training, it may well prove to be a very useful thing.
Overall, then, this was a very interesting talk covering some current work that will have an effect on us all before long. What matters now is that we must be prepared to make sure that the edu-data tail doesn’t wag edu-data dog.