Universities have been significantly disrupted by the ongoing COVID-19 pandemic, and are likely to remain in a fragile, uncertain condition for months if not years to come. The very rapid shift to online teaching seems increasingly likely to be just a first step on a long path to the expansion of digital or hybrid technologies in higher education. For many education technology (edtech) vendors, the pandemic is not just a health crisis and an educational emergency, but a market opportunity fuelled both by private capital calculations and by desperate university customers. With the very continuation of higher education teaching at stake as universities recover from the coronavirus crisis, companies providing vital digital infrastructure for distance education are attractive prospects for educational and market institutions alike.
Our special issue on ‘The datafication of teaching in higher education’ was already in production as coronavirus spread around the planet. The issues confronted by many of the authors, however, anticipate discussions now occupying universities as they work out how far to increase their digital delivery, and what to do about the huge quantities of data these technologies collect about their students, staff and institutional performances. Although the use of statistics, metrics and data to measure student achievement, staff outputs and university performance is not new, as we show in the editorial introduction, digital forms of data are becoming increasingly prevalent with the widespread introduction of digital technologies for teaching and learning. Predictive learning analytics, learning management systems, online learning platforms, performance dashboards, plagiarism detection, library resource management, student experience apps, attendance monitoring, and even artificial intelligence assistants and tutors all depend on the persistent collection and analysis of data, and are part of a rapidly growing edtech sector and a multibillion dollar education market.
The datafication of teaching in higher education is transformative in three key ways. First, it is expanding processes of measurement, comparison and evaluation to as many functions and activities of higher educationas possible, through increasingly automated systems that run on highly opaque and proprietary code and algorithms that are based on specific technical understandings of education. Second, datafication privileges performance comparison more than ever, and thereby reinforces longstanding political preoccupations with marketization and competition, as the comparative performances of students, staff, courses, programs and whole institutions are made visible for evaluation and assessment. And third, datafication fuses higher educationto the business models of a global education industry, which then reshapes higher educationto fit its preferred ideas about what constitutes and measurably beneficial university experience. In other words, technologies of datafication are the material embodiment of particular measurement practices, political priorities and business plans, and reshape institutions of education to fit those forms.
The collected papers in the special issue tease out a number of key concerns. Paul Prinsloo foregrounds issues of ‘data power’, arguing that data systems define what ‘counts’ as a good student, an effective educator, or a quality education. He raises significant questions regarding the ‘data colonialism’ of edtech companies from the Global North pushing into Global South contexts to reveal ‘truths’ about education, students and teachers. Data analytics and the dashboards that present information about students are the focus of Michael Brown, whose article identifies the role of dashboards in ‘seeing students’ and shaping educators’ pedagogical strategies. Educators, he reports, may find their normal pedagogical routines stymied by the demands of datafication, and struggle to make sense of the data presented to them by their dashboards. This, for Michaela Harrison and coauthors, raises the issue of how ‘student data subjects’ are created from data in ways that make them visible to the educator’s eye as digital traces, which they argue may result in a ‘process of (un)teaching’ rather than meaningful teacher-student interaction.
Learning management systems have acquired some of the most extensive databases of student information on the planet. Roxana Marachi and Lawrence Quill draw specific attention to the learning management system Canvas, arguing it enables ‘frictionless’ data transitions across K12, higher education, and workforce data through the integration of third party applications and interoperability or data-sharing across platforms. They make the important call for greater public awareness concerning the use of predictive analytics, impacts of algorithmic bias, and enactment of ethical and legal protections for users who are required to use such software platforms. Juliana Raffaghelli and Bonnie Stewart suggest that building educators’ ‘data literacy’, with an emphasis on critical, ethical and personal approaches to datafication, is an important response to the increase of algorithmic decision-making and data collection in higher education, enabling educators make sense of the systems that shape life and learning in the twenty-first century. Extending a critical data literacies approach a computer science classroom, Mary Loftus and Michael Madden report on an experimental teaching module where students both explore the construction of machine learning models and learn to reflect on their social consequences as ‘students who will be building the autonomous, connected systems of the future’.
A number of the papers examine how datafication reinforces logics of marketization and performativity. Annette Bamberger, Yifat Bronshtein and Miri Yemini, for example, argue that as social media has become central to university marketing and reputation management, techniques of datafication help produce persuasive information that can be circulated as social media marketing material in the context of competitive struggles for the international student market. Aneta Hayes and Jie Cheng then examine the shortcomings of international teaching excellence and higher education outcomes frameworks, arguing that ‘epistemic equality’ and non-discrimination should be officially considered as indicators of teaching excellence, and show how evaluating universities on epistemic equality could work in practice. Such an approach stands in contrast to the surveillance techniques of the ‘smart campus’ analysed by Michael Kwet and Paul Prinsloo, who foreground the risks of normalizing surveillance architectures on-campus, call for a ban on various forms of dataveillance, and argue for decentralized services, ‘public interest technology’ and more democratic pedagogic models.
Rounding out the special issue, Neil Selwyn and Dragan Gasevic stage a dialogue between critical social science and data science. They add a computational dimension to familiar social criticisms of data representativeness, reductionism and injustice, as well as exploring social tensions inherent in technical claims to data-based precision, clarity and predictability, and finally highlight opportunities for productive interdisciplinary exchange and collaboration. Their paper offers a productive way forward for research on datafication in higher education. But significant challenges remain to reimagine and reshape the role of HE in the 2020s, both during the coronavirus recovery and in the longer term. We hope the special issue helps to catalyse debate about the limits, potential and challenges of the datafied university, and about the role of datafication in higher educationfor the future.
Ben Williamson (University of Edinburgh), Sian Bayne (University of Edinburgh) and Suellen Shay (University of Cape Town)