'Re-imagining ‘Learning Analytics’ … a case for starting again?' by Neil Selwyn in (2020) 46
The Internet and Higher Education comments
This brief paper develops a series of provocations against the current forms of Learning Analytics that are beginning to be implemented in higher education contexts. The paper highlights a number of ways in which Learning Analytics can be experienced as discriminatory, oppressive and ultimately disadvantaging across whole student populations, and considers the limitations of current efforts within educational data science to increase awareness of ‘ethics’ and ‘social good’. This culminates in a stark choice: is it possible to substantially improve the field of Learning Analytics as it currently stands, or should we abandon it in favour of new forms of applying data science that are aligned with the experiences of non-conforming ‘learners’ and un-categorizable forms of ‘learning’?
Selwyn argues
Much of this special issue is understandably concerned with thinking the best of Learning Analytics. After all, the idea of Learning Analytics raises a number of seductive promises for all stakeholder groups currently involved in the development and implementation of these technologies in higher education settings. What educational data scientist would not want their work to foster potentially powerful forms of active learning across large student populations? What higher education leader would not want rich, detailed insights into key institutional ‘performance points’ such as student performance, retention and engagement?
In contrast, then, this brief article deliberately considers the premise of this special issue in a contrary manner. Instead of ‘reading with’ the promises of Learning Analytics, what happens if we choose to ‘read against’ them? What are the fundamental social problems inherent in the ways that Learning Analytics products and practices are being realized in higher education contexts? Who is most likely to be experiencing these issues? How might we ‘think otherwise’ about the application of analytics in higher education – particularly if ‘we’ is taken to include the many interest groups not currently included in the notion of Learning Analytics ‘stakeholders’? As I hope this paper will show, ‘thinking the worst’ can be a useful means of ‘stress testing’ the core premises, principles and politics that the current implementation of learning analytics into education is currently built around.
Of course, what readers choose ultimately to do with these insights will depend on their own underpinning agendas, values and ideologies. Yet regardless of one’s background, this is a highly appropriate moment to be introducing an element of pessimism into proceedings. People working in the area of learning analytics, education data-mining and other forms of ‘educational data science’ find themselves at a crossroads. On one hand, the vast majority of people working along these lines are clearly very thoughtful and well-intentioned - developing products, protocols and practices that they genuinely hope (if not believe) will considerably improve learning and learners’ experiences of engaging in education. On the other hand, going by the forms of ‘analytics’ that we see being implemented in higher education contexts, there might well be a strong case for radically rethinking how ‘Learning Analytics’ is playing out beyond the confines of LAK, SOLAR and the other academic manifestations of Learning Analytics.
To be blunt, then, this paper starts from the contention that something is surely amiss if the main aim of academics working in this area is to be “caring and supportive” (Prinsloo, 2019), but significant numbers of people continue to experience Learning Analytics tools and techniques as “data being used against me to screw me” (Essa, 2019). This tension has been highlighted in the recent Twitter-controversies over the propensity of learning analytics tools and systems to be used for purposes of institutional surveillance rather than individual support (e.g. Kovanovic, 2019). As I have argued elsewhere:
“All told, there is an emerging suspicion (warranted or not) that students are not the primary beneficiaries of the Learning Analytics technologies they are subjected to during their school or university education” (Selwyn, 2019).
So, in this brief paper I want to reflect a little further on these tensions – especially the question of what academics working in the area of Learning Analytics consider the political intent of their work to be. If we take the politics of Learning Analytics seriously, then perhaps we need to begin thinking along more radical lines than simply embracing ‘ethics’ and trying to foreground possible ‘social goods’ that educational data science might support. Instead, it might be a useful thought experiment to pursue a more radical logic – what Paul Prinsloo (2019) identifies as “question[ning] the very existence of Learning Analytics” (or, at least, questioning the very existence of the forms of Learning Analytics that are currently being implemented in educational settings around the world).
In working through this prospect, I want to draw on various recent provocations from within the broader data science community – all data science ‘insiders’ who are voicing informed frustrations over the politically uninterested malaise that they see pervading their field of work. These writers are beginning to argue that it is not good enough for data scientists to presume that technology is essentially neutral, that data is objective, and resort to all-absolving claims of ‘I am just an engineer’. In my view, these insider critiques offer some interesting future directions for educational data science to pursue.
'Is Data Dark? Lessons from Borges’s “Funes the Memorius”' by Alfred Essain in (2019) 6(3)
Journal of Learning Analytics 35–42 comments
In 'Funes the Memorius' Jorge Luis Borges tells the tale of an Argentinian man who falls off a horse, becomes paralyzed, but with his misfortune acquires the strange gift of infinite memory. Funes remembers everything, which is to say he forgets nothing. l will use Borges' story as the backdrop for my response to Professor Selwyn.
My commentary is in three parts. First, I begin by stating some core areas of agreement, of which there are many. Second, I examine Selwyn’s use of the word “data”. I argue that it perpetuates a number of common misconceptions about statistics and the scientific method. We cannot understand the importance of learning analytics without first clarifying these misconceptions and moving beyond them. In the course of my argument I challenge Selwyn’s central thesis that “Education is inherently social, inherently contextual, inherently subjective; it means you can’t objectively rate it, measure it, indicate it.” Third, I turn the table on Selwyn. As a critic of learning analytics Selwyn suggests that data “disadvantages large numbers of people”. I argue that the root problem in education is the status quo, which Selwyn unwittingly represents, and not learning analytics. If we care about equity in education, as part of a broader interest in social justice, then learning analytics and the use of educational data can be a powerful instrument for empowering the disadvantaged.
'Is Data Dark? Lessons from Borges’s “Funes the Memorius”' by Alfred Essain in (2019) 6(3) Journal of Learning Analytics 35–42 comments
