Thinking Critically and Realistically About the Platform University

Thinking Critically and Realistically About the Platform University

Peter Francis and Carly Foster

The concept of the ‘Platform University’ is gaining prominence as academics and social commentators writing from a variety of perspectives explore the nature, power and impact of soft technologies in Higher Education. Whilst competition and globalisation are undeniably reshaping the HE landscape, it is technology, and in particular digital platforms, which arguably are threatening to disrupt the very foundations of trust and public good on which the sector has built its reputation. But, why is this the case? Particularly when a platform in its simplest form is nothing more than a series of code enabling two or more groups to interact. Often described as agile, the platform enables coordination, collaboration and exchange. When successful it creates and proliferates demand for both itself and for the product or service which it provisions – in this case education, research and knowledge exchange.

The answer to the above question lies partly in the purported impact of digital platforms and partly in the manner in which they are used. At the heart of the debate is data. Platforms enable the extraction and control of individuals’ data providing proprietary access to traces of activity and to generating new forms of value from it. Yet, to what purpose, by whom and with what impact are digital platforms being used? These questions are being asked in the hallowed corridors, classrooms, laboratories and studios of Universities in the UK and overseas. Unsurprising perhaps, when phrases such as ‘platformisation’ and the ‘platform society’ are gaining credibility at a time when the sector is facing considerable political and economic challenge to deliver efficiencies, demonstrate relevance and add value.

Alongside the Virtual Learning Environment, Educational Analytics is one of the most visible digital developments in HE of the last decade. Educational Analytics utilises engagement, behavioural and, in some cases, demographic data to create student profiles to deliver timely communication (‘nudges’) and student support interventions. Three platforms exist within Educational Analytics. The data platform integrates various sources such as the student records and library loans systems and integrates with other platforms such as the virtual learning environment to host data in one centralized location. Although plentiful with raw data this platform usually requires a digital platform to visualise and operationalize insights in to student attributes and behaviour at scale. It is typically a dashboard used by academic and professional support staff to support activities. Between the two sits arguably the most controversial – the algorithmic platform. Algorithmic processing is a spectrum. At one end there is descriptive analytics – say the sum of a student’s library visits. Midway on the spectrum are derived insights on the frequency and timing of those visits. At the far end of the spectrum is predictive analytics. Usually based on historical patterns this approach weights a student’s library visiting behaviour in the context of a likely outcome such as whether he or she will progress or drop out.

Neoliberal Orthodoxies and Counter-narratives
When describing the ‘Platform University’ there are those who project a neoliberal orthodoxy on to platforms like Educational Analytics; presented as a stable object that marries the benefits of profiled, data-driven higher education with adaptive and scalable personalized learning environments. Educational Analytics is articulated as a means to standardise and deliver cost-effective solutions for improving student outcomes and support through engagement initiatives. Often referencing published research as ‘evidence’ of impact, these commentaries present the analysis of student data as objectively informing a fairer and more dynamic allocation of resources to those in need compared to traditional profiles such as gender or Widening Participation. Such thinking attests that profiling matches students to the most appropriate resources based on evidence and best fit. Embedding concepts such as self-regulated learning within student facing analytics and personalised learning models conveys the digital platform as the enabler of success and empowerment, placing student agency at the centre of educational pedagogies.

In contrast, counter-narratives bring to the fore the political economy of digital platforms relative to the data and algorithms underpinning them. Presenting the ‘black box’ of platforms such as educational analytics as an obfuscation of Higher Education’s wider social responsibility, these ‘alternative’ perspectives attempt to uncover the real impact of digital platforms on agency and identity for institutions, staff and students. With analysis located firmly in concepts such as ‘platform capitalism’, ‘digitalisation’, ‘surveillance’ and ‘control’, such perspectives view digital platforms as creating instability and inequity by exploiting students’ data to improve university performance and metrics. Rather than enabling effective risk-based interventions, data harvesting, algorithmic processing and prediction profiling are presented as exacerbating discrimination and bias by identifying, ranking and rating individual students relative to cohort norms. And the providers of such services – often private and operating at arm’s length to HEIs – are targeted as fuelling an environment where big data and machine learning methods provide a vehicle for massification. Platforms undermine students’ relationships with academic staff and university experiences by attaching a common value to their personal engagement style; they negate student agency by ranking and ‘pushing’ options and resources to them.

Thinking Seriously About the Platform University
Whilst headline grabbing, such polarised narratives do little justice to the complexity of platforms in HE. Theoretically limited and empirically light, they lack precision, offering partial operational understanding, and are reductionist in tone. Searching for an equipoise, we favour an approach able to comprehend the platform through human experience, opportunity and fallibility. Thinking seriously about the platform university involves allowing specificity, context and theory prominence in any assessment of impact and purpose. This is essential in order to assess the opportunities and challenges of digital platforms in HE and for preserving student and staff agency at macro, meso and micro levels of university structure.

Reflecting on questions of impact and purpose, critical realism is an approach that supports a constructive discussion on digital platforms which goes beyond the naive neoliberal assertion of ‘its harmless’ and ‘it benefits everyone’. It also equips us to reject the crude narrative that the algorithmic processing of students’ data leads to malignantly capitalist mentalities enacted in and through marketized university spaces. Critical realism offers the opportunity to reassess orthodox and counter-narratives, opening up alternative viewpoints which are empirically and theoretically grounded. This has been our approach to implementing educational analytics in one large UK HEI. In our view, thinking critically and realistically about the Educational Analytics platform means:

  • understanding EA as part of a whole university approach where the algorithm is transparent and proportionate to a human-led intervention strategy to improve outcomes for all Simply mining educational data can be used to predict student outcomes but no intervention is assumed until staff engage with the outputs. Actuarial based approaches can promote positive differentiation where risk can be an opportunity factor. Insights can be coupled with effective interventions not only for individual students but also programmes and subjects within an inclusive framework to impact positively on those outcomes for all students.
  • accepting that data itself is benign; but its manipulation and analysis has the potential for bias and discrimination but inclusivity and regulation can be outcomes of any defined implementation plan. GDPR asks for informed consent but it creates challenges to inclusivity and equality of opportunity and engagement and should be managed with students as stakeholders.
  • acknowledging that agency can be improved within the right context of student community, values and culture, by moving from a position whereby the relationship between agency and structure cannot be presumed. Human relationships are a necessary input and outcome of Educational Analytics.
  • recognising that the current research ‘evidence base’ for improving student outcomes through Educational Analytics is weak and the need for more empirical evidence stands counter to both orthodox and critical narratives. Its use and promotion must be carefully managed through further peer reviewed research involving qualitative and quantitative methods, and collaboratively with students.
  • Challenging the theories of change for educational analytics which are currently unable to explain the impact on student outcomes relative to a platformised university. This is particularly the case in relation to positive and thriving student communities.

So What?
Whilst exposing platformisation to fair critical scrutiny in the context of university organization, ‘the platform university’ undermines an emerging understanding of digital platforms in HE. Overlapping social contexts at the macro, meso and micro levels are essential to understanding changes brought about by big data and technology in HE. What is required is a nuanced approach, informed by an agenda able to provide a richer understanding of the complex and dynamic nature of data analytics for learning, teaching and the student experience. Without this the binary narratives, with their predilection for reductionism and essentialism, negate much needed specificity and agency in practice and hamper the adoption and development of Educational Analytics. No institution wants to miss out on the opportunity to help students in need; nor do university leaders want to create barriers to success. We will continue to develop and share our systematic and empirical critical realist research on educational analytics with the sector as we aim to offer insight and evidence based on experience rather than conjecture.


Peter Francis is Deputy Vice-Chancellor and Professor of Criminology at Northumbria University.

Carly Foster is Insight and Performance Manager at Northumbria University and PhD student in Educational Research at Lancaster University

IMAGE CREDIT: Photo by Markus Spiske on Unsplash