Last updated: 2026-10-06

U
Undergraduate level

From Marks to Maturity: Should Degrees Assess How Learners Develop?

Why a degree measures what a student produced at particular moments, and what it would take to measure how the student changed

Most degree classifications are produced by aggregating marks from individual modules. A weighted average of coursework, examinations and projects becomes a first, a 2:1, or a lower class. That process tells us something real about what a student achieved at specific points in time. It tells us surprisingly little about one of the central purposes of higher education, which is how the student changed during the process.logs track the arc of change, not just points

This page starts from that practical observation. Assessment in most universities is episodic. It samples a learner's work at intervals and treats each sample as a separate piece of evidence. The question I want to explore is whether a degree could also be a judgement about development, and what such a judgement would need to look like. Near the end, the page moves to a more speculative question: what a qualification actually rests on. I treat that part as a thought experiment rather than a proposal.

We Assess Artefacts, Not Trajectories

The usual structure of assessment is familiar. Universities set examinations, assignments, projects and, at the end, a dissertation. Each produces a mark. The marks are combined into module results, the module results into a programme classification, and the classification is printed on the transcript.

The chain looks like this:

  • assignment results
  • module grades
  • programme classification

At every step, the evidence is an artefact: a piece of work produced at a point in time under particular conditions. The question that is largely missing from this chain is the one that matters most to the student and to any employer reading the transcript. What is the evidence that this learner has developed? The chain assumes that development has happened, because the later marks are often higher than the earlier ones, but the assumption is rarely examined directly. A rising profile of marks is a proxy for growth, and a rather indirect one.logs make the trajectory visible

Learning Is Inherently Longitudinal

It helps to ask what we would expect to see over three or four years of study. Some of the things a university hopes to develop unfold slowly and are hard to capture in any single piece of work.

Academic development. Synthesis becomes more fluent. Critique becomes more precise. Arguments acquire depth, and a student learns to recognise which of the available evidence actually bears on the question. Anyone who has marked a first-year essay and a final-year dissertation from the same student will have seen some of this.

Professional development. Judgement improves. Students take on more responsibility for decisions, begin to notice ethical considerations they would once have missed, and learn to explain why a choice was right rather than only that it was.

Learning development. Independence grows. Students become better at setting their own direction, monitoring their own progress and changing course when something is not working. This is the territory of self-regulation and metacognition, which the earlier pages on the self-model as a learning tool and on learning logs and narrative identity explore in more detail.see-also: self-model as a learning tool

Universities assume that such growth has happened. Graduate attribute statements say so, and many programmes are designed around it. Very few assess it directly. The developmental outcomes appear in the learning objectives and then disappear from the marking scheme, which records only the artefacts.

A Longitudinal Portfolio Model

One possible alternative is to treat the degree as a body of evidence accumulated over time, rather than as a set of separate samples. In this model, each student builds a portfolio across the programme. It would include coursework, projects, reflective writing, formative work and dissertation outputs, along with the feedback each one received and the student's responses to it.the portfolio is a narrative identity

The final assessment would then consider four things in addition to the marks:

  • Consistency. Does the body of work tell a coherent story about the student's learning, or do the pieces contradict one another?
  • Development. What has changed between the first piece and the last, and what evidence supports that change?
  • Maturity. How has the learner's judgement, self-management and handling of uncertainty evolved?
  • Independence. What evidence exists that the student has taken on more of the direction of their own work?

On this model, a degree would reflect both what was achieved and how the capability was developed. The marks would remain, because they carry information about disciplinary attainment that is worth keeping. They would be joined by a judgement about the trajectory.

This is a proposal for a way of thinking about the question, not a finished scheme. It raises practical questions about workload, about the reliability of judgements about development, and about how a longitudinal record can be assessed fairly across different students and different routes through a programme. Those questions are real, and I return to some of them below.

The Connection to Learner Maturity

A qualification currently measures disciplinary attainment. It does not measure, directly, self-management, adaptability, reflective capability or the capacity for independent learning. Those dimensions appear in graduate attribute statements, and they are often assumed to follow from a good degree. The assumption may be reasonable, but it has not been tested in any systematic way that I know of.

Maturity models try to describe these dimensions in terms of levels. The site's page on personal learning maturity applies the levels of the Capability Maturity Model Integration to a learner, and the page on bachelor's graduate attributes sets out what a degree is meant to develop. Putting the two together makes the gap visible. A learner can be at a high level of disciplinary attainment and a lower level of self-direction, or the reverse.CMMI: a maturity framework borrowed from software eng

This gives a plain observation: a first-class degree does not necessarily imply a highly mature learner, and a highly mature learner does not necessarily receive the highest marks. The relationship exists, but it is imperfect, and the imperfection is what a developmental assessment would try to see.

Why AI Makes This More Relevant

Historically, the central assessment question was whether the student could produce the artefact. That question is becoming less useful. A student can now produce a competent artefact with a great deal of help, and the artefact no longer tells us much about who produced it or how.

The more useful question is whether the student can direct, evaluate, understand and improve the process. That question is developmental by nature. It cannot be answered by a single submission, because a single submission cannot show how the student's judgement changed over time. A longitudinal record can, at least in principle, show evidence of genuine learning, of increasing sophistication and of developing judgement, more effectively than any individual piece of work.

The studies I have discussed elsewhere on this site point in the same direction. Shen and Tamkin found that developers who asked for explanations and conceptual help kept their learning intact, while those who handed over the whole task did not[1]. The pattern of use matters as much as the output. A portfolio that records patterns over time is better placed to show that than a single submission. The argument is developed in How I Use AI in Teaching, and How Students Use It Responsibly.see the linked pages for how to track that pattern

An Unexpected Benefit

A longitudinal portfolio would naturally make discontinuities more visible. A dramatic change in capability between two pieces of work, a sudden change in voice, or reasoning that shifts style without explanation would all stand out against a body of work that is read as a whole.

I want to be careful about this point. It is not an argument for AI detection, and I would be uneasy about any approach that treated the portfolio as a surveillance device. Discontinuities are informative because they invite a conversation. A tutor who notices a sudden shift can ask about it, and the answer may be entirely innocent. Improved detection of misconduct would be a side effect of the approach, not its purpose. The purpose is understanding how the learner has developed. The earlier page on misconduct and the history of assessment makes the same point about proportionate assurance.the portfolio as a narrative

What Exactly Does a Degree Certify?

This is where the page turns to the more speculative material. If we are going to talk about assessing development, we need to be clear about what a degree is for in the first place. There are several plausible answers:

  • a certificate of knowledge in a discipline
  • evidence of particular skills
  • a judgement about capability to perform in a field
  • a record of development over a period of formation

Each of these answers suggests a different kind of assessment, and most real degrees probably combine several of them. The marks mostly speak to the first two. The third and fourth are harder to capture, and they are the ones a longitudinal approach would address.

There is a further possibility, which I find more interesting and less comfortable. A degree may represent trust. Employers trust universities. Universities trust their academics. Academics trust the evidence they have evaluated. The qualification is a point at which these trust relationships converge.a signal, not a seal

Accreditation as a Chain of Trust

This section is intentionally provocative, and it is a thought experiment rather than a policy proposal. Its claim is that qualifications derive their authority from human judgement, and that this is easy to forget.

A degree is not valuable because of the paper it is printed on, the certificate or the database entry. It is valuable because people trust the evaluation process behind it. The chain of trust can be drawn as a sequence:

  • student
  • assessed by academics
  • academic standards processes
  • university reputation
  • employer trust

The question I want to put is whether the academic component could be made more explicit. At present, the link between the individual academic who judged the work and the qualification that results is usually invisible. The judgement is aggregated, moderated and reported as a classification. The name of the person who made the judgement is not part of the record. Could it be?

Academic Lineages

There are precedents for thinking about expertise as a lineage. Doctoral supervision produces family trees of research students and supervisors, and academic disciplines often trace their ideas through such chains. Apprenticeship traditions and guild systems passed standards from one generation of practitioners to the next, and Lave and Wenger's account of learning in communities of practice describes how novices move towards full participation through their relationships with established members of a community[2].

Following that thought experiment, a qualification could in principle carry a record of the chain of academic judgement behind it:

  • assessed by Professor A
  • who was supervised by Professor B
  • who was supervised by Professor C

and so on, forming a chain through academic history. The chain would make the human basis of accreditation visible, and it would give a student a sense of belonging to a tradition of standards. I offer this as an idea to think with, not a recommendation. It would be expensive to maintain, it would be hard to make consistent across disciplines, and it would raise obvious questions about who gets into the chain at all.

What Institutions Actually Add

A pure lineage system would have serious weaknesses, and I do not want the argument to become a case against universities. Institutions provide things that individual judgements cannot. They provide quality assurance, moderation that reduces the effect of an individual marker's idiosyncrasies, shared standards across a discipline and across cohorts, resources for learning, governance that holds people to account, and continuity that outlasts any individual academic's career.

The purpose of the thought experiment is not to replace these things. It is to examine where the trust that the institution relies on originates. Institutions are the framework within which judgements are made, and the judgements themselves still have to be made by people.

A Hybrid View

A more practical synthesis would treat a qualification as the combination of three things, rather than as a sum of marks:

  • institutional accreditation, which provides the framework, standards and assurance
  • academic judgement, which provides the evaluation of the work and of the learner
  • evidence of learner development, which shows how the capability has changed over time

This view keeps the institution and the marks, and adds the human judgement and the developmental evidence that the current model leaves largely implicit. It is closer to what many academics already do when they write a reference or sit on a board, and it would make that work visible.

Conclusion

Perhaps the most important question a degree can answer is not "What marks did this student achieve?" but "What kind of learner has this person become?" A qualification based solely on module results offers only a partial answer. A longitudinal view of development may provide a richer picture of capability, maturity and professional readiness, and it would reward the kind of growth that universities claim to value most.

It also raises a deeper question about accreditation itself. Universities certify achievement, but that certification rests on chains of human judgement and trust. The institution provides the framework. The academics provide the evaluation. The learner provides the evidence. A degree may therefore be better understood not simply as a collection of marks, but as a statement that a community of scholars is willing to stand behind the development of a particular individual.

References

  1. Shen, J. H., & Tamkin, A. (2026). How AI impacts skill formation. arXiv:2601.20245. https://arxiv.org/abs/2601.20245
  2. Lave, J., & Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Cambridge University Press.