Almost twenty years ago, the Burgess Group concluded that the UK honours degree classification system was no longer fit for purpose. The group, convened by Universities UK, recommended the introduction of a Higher Education Achievement Report, a document designed to capture the richness of student learning rather than reducing years of study to a single classification. The HEAR would sit alongside the degree certificate, offering employers and postgraduate tutors a fuller picture of what graduates had actually achieved.
That was 2007. The recommendation was endorsed. Trials were conducted. Implementation guidance was published. And yet here we are, still debating fundamentally the same question: what are degree classifications actually trying to communicate?
This week, Wonkhe featured Mark Peace calling for a more robust interrogation of the classification system’s purpose, and what a better signal might look like. Alongside this, Abby Osborne and Robert Eaton observed that critical thinking, arguably among the qualities that matter most in a graduate, remains “signalled but invisible” in degree outcomes. A student might demonstrate exceptional analytical capability throughout their studies, yet leave with the same credential as someone whose First derived from rather different strengths.
The intervening years have not been idle. The sector has invested considerably in defending and calibrating the existing system, particularly as concerns about grade inflation have intensified. Office for Students data shows that the proportion of graduates awarded first-class degrees nearly doubled between 2010-11 and the pandemic years, prompting regulatory scrutiny and institutional soul-searching. Recent figures suggest some stabilisation, with 29.6 per cent achieving Firsts in 2022-23, down from the pandemic peak but still roughly double the proportion from a decade earlier.
Much energy has been devoted to ensuring that a First in 2025 carries comparable weight to a First awarded previously. The OfS has examined classification algorithms, identified providers at increased risk of inflationary practices and published guidance on ensuring awards credibly reflect achievement. This work matters. Employers and the wider public need confidence that qualifications mean something consistent over time.
But there is a different question that has received less sustained attention: whether the classification itself, even when perfectly calibrated, tells us very much at all.
Consider what a First actually communicates. It indicates that a student performed at the highest level according to their institution’s assessment regime. It does not reveal whether that performance derived from technical mastery, creative insight, collaborative skill or sustained intellectual resilience. It does not distinguish between someone who excelled from day one and someone who struggled initially before achieving remarkable growth. It cannot capture the student who took academic risks that did not always pay off, nor the one who contributed meaningfully to the learning of their peers.
I graduated with a First in 2001, when the proportion achieving that classification was roughly a third of what it is today. I am not sure it told prospective employers much about the particular qualities I might bring to their organisation. It signalled something, certainly. But what, exactly?
The Burgess Group understood this limitation. Their proposed HEAR was not merely an enhanced transcript but an attempt to render visible qualities that single classifications obscure. Bologna-aligned countries with comparable diploma supplements have grappled with similar questions. The challenge has always been implementation: building institutional systems capable of capturing richer information, and persuading employers that such information warrants their attention.
The arrival of generative AI in higher education may be forcing the conversation back. As Mark Peace noted in his Wonkhe piece, in a future where AI tools become routine in knowledge work, outputs alone will be a much less effective signal for the competences we associate with graduateness. The essay, the report, the analysis: all can be produced with substantial AI assistance. What distinguishes the graduate who used such tools wisely from one who used them as a substitute for thought?
This is not principally an academic integrity question, though it has integrity dimensions. It is a question about what higher education claims to develop in students, and how that development might be evidenced. Critical thinking, the capacity for reasoned judgement under uncertainty and the ability to evaluate sources and synthesise perspectives: these qualities matter more, not less, when AI can generate fluent text on demand. Yet they remain, as Osborne and Eaton observed, largely invisible in our credentialing systems.
There are no easy solutions here. Richer achievement records create administrative burden. Employers may lack the time or inclination to engage with nuanced documentation when a simple classification suffices for initial screening. The sector has limited collective capacity for coordinated change, and previous attempts to move beyond classification have not achieved critical mass.
Yet the question of what degrees communicate feels more pressing than it did when Burgess first raised it. The combination of grade inflation concerns, AI integration and an increasingly sceptical public discourse about graduate value creates a context in which the limitations of single-number credentialing become harder to ignore.
Perhaps the conversation needs reframing. Rather than asking how to supplement classifications with additional documentation, we might ask what employers and society actually need to know about graduates, and whether our current systems are capable of telling them. The answer may involve not just richer records but different approaches to assessment throughout a student’s journey, approaches designed from the outset to surface the qualities that classifications cannot capture.
I do not know what that might look like in practice. What seems clearer is that defending the classification system’s calibration, whilst important, addresses only part of the challenge. The deeper question, what a degree actually tells us about the person who holds it, remains as relevant as it was two decades ago. Possibly more so.
If higher education is to articulate persuasively what makes human learning distinctive in an age of artificial intelligence, it may need to find better ways of showing what it knows about its students. A single number, however carefully calibrated, seems an increasingly inadequate answer to an increasingly important question.