Last updated: 2026-10-07

U
Undergraduate level

Beyond Marks: A Developmental Model of Higher Education

If development were the main thing a degree assessed, what would the degree look like, and would it be better?

The previous page, From Marks to Maturity, asked whether degrees should assess learner development as well as academic attainment. I suggested a longitudinal portfolio as a way of seeing that development, with the marks kept as they are. This page asks a more radical question. What would happen if the development became the primary object of assessment, and the marks became one source of evidence among several?logs as evidence of the journey

I should say at the outset that I do not think this is easy. But I think the idea is plausible, some current practice already looks like it, and many of the objections to it are weaker than they first appear. The reader will have to judge whether the current model is as obviously better as it seems.

A Contrast

A PhD is not usually awarded by averaging hundreds of marks. The candidate develops over several years, receives extensive formative feedback from supervisors and peers, produces a body of work, and then undergoes a final holistic assessment in which the examiners judge whether the person has become a researcher.the PhD model is already here

Undergraduate education takes almost the opposite approach. Most work is graded. Most grades contribute to the final classification. Development is assumed rather than assessed directly, and the feedback that would support development often arrives after the mark has already been recorded.

What would happen if we reversed that balance, and organised an undergraduate degree more like a doctorate? That is the question this page explores.

The Current Model

The prevailing system runs from assignment, to mark, to module grade, to degree classification. It has real advantages. It is administratively simple, it is familiar to students, employers and regulators, and it has the appearance of objectivity. A weighted mean produces a number, and numbers are easy to compare, defend and report.

The weakness is structural rather than personal. The system concentrates on performance at many isolated moments, and it pays relatively little explicit attention to the learner's overall trajectory. Classification draws only on the second and third years, and a student who struggled in the second year and found their footing in the third is classified on a mean that gives both years a similar say, so the shape of the change is lost. There is a boundary-case rule that can move a student up a class where their marks have improved, sometimes described as exit velocity, but it applies only within a tight margin of the grade boundary. It recognises a trajectory, but only for students who were already close to the line. The system can see that a student did well. It is much less able to see how they became someone who could do well.

A Developmental Model

The alternative is to make most of the activity formative. Students would still produce coursework, complete projects, undertake investigations and receive detailed feedback. Much of that work would be read and commented on, and some of it would be revised. But the primary purpose of each piece would shift from measuring performance to supporting development.

The final judgement would then draw on a portfolio of evidence accumulated across the years, a reflective synthesis written by the student, and a viva or professional discussion in which the student and the assessors examine the trajectory together. The result would still be a classification, or some other award, but it would rest on a different kind of evidence.the portfolio as a living map of growth

This is not a radical break with everything we already do. Many programmes already include a final-year project with supervision, a dissertation that asks for synthesis across the degree, and a reflective element. The proposal is to make the developmental part the core of the award rather than an extra component added to a mark-based structure.

A Test First

One element I would add is borrowed from software development. In test-driven development, the programmer writes the test that defines success before writing the code, watches it fail, and then builds until it passes. The failing test is useful because it fixes the starting point and says what the work has to achieve.

A developmental degree could use the same idea. An early summative test in the first weeks would establish each student's entry benchmark: what they can do, where their understanding is thin, and what they bring to the programme. That benchmark is not a mark that counts towards the award. It is the baseline against which development is measured, and it gives the later portfolio something to be compared with. Without it, the claim that a student has grown is hard to support, because we would have no record of where they started.

The analogy has limits, and two of them matter. First, a baseline test can easily become a source of anxiety, and a student who sees their first attempt as a verdict will behave accordingly. The test would need to be explicitly diagnostic, with the results used to plan support rather than to rank. Second, the baseline must not quietly become a target. If it is later treated as a yardstick for penalising the students who started weakest, the developmental model has reproduced the mark-based model with more paperwork. Used properly, the entry test is a measurement of change, and it should be judged by how much the student has changed, not by where they began.

Why AI Makes This Relevant

Historically, assessment asked whether the student could produce the artefact. Technology has changed what it takes to produce one. The more useful question may now be whether the student can understand, evaluate, improve, justify and integrate what has been produced, with or without help.

This matters for AI assistance, for contract cheating, and for plagiarism. All three become less useful as a basis for judgement when the assessment concerns isolated submissions. A single essay can be bought, generated or copied. Years of work, with the feedback, revision and conversations that go with it, are much harder to fabricate. Bretag and colleagues found that contract cheating was associated with the teaching and learning environment as well as with opportunity[1], which suggests that the structure of the degree is part of the problem and part of the answer. The earlier page on misconduct and the history of assessment makes the general case.the portfolio as proof of growth

Learning Versus Performance

This is the section I find most uncomfortable, because it describes what I suspect many students already know. Many students optimise for marks. The structure of the system encourages them to do so, and it would be odd to blame them for responding sensibly to incentives.

It helps to separate two orientations. A performance orientation asks how to maximise marks, how to avoid losing them, and what the examiner wants. A learning orientation asks what has been learned, what the student can now do, and how they have developed. Both orientations are reasonable responses to different systems. The first is what a mark-based structure rewards. The second is what a developmental structure would need.the 'performance' orientation

I do not think a developmental model would remove the performance orientation entirely, and I would be suspicious of any proposal that claimed it could. But I think formative-heavy systems align better with learning, because they remove the penalty for the mistakes that learning requires. A student who can try something, get it wrong, receive feedback and try again is in a better position to learn than one who must get it right the first time.

The Portfolio

The final portfolio would contain a selection of the student's work across the degree. It might include selected coursework, project outputs, reflective pieces, peer feedback, supervisor feedback, evidence of revision, and examples of failure and recovery.

The portfolio is not a showcase. A showcase presents the best work, and a developmental portfolio should also show the weaker work and the route from one to the other. The point is evidence of growth, and growth is visible in the difference between the first attempt and the last, and in the student's account of what changed between them. This is close to the learning-log approach described in Learning Logs and Narrative Identity, scaled up to a degree.the 'route' is key: growth needs the stumble

The Viva

The viva in this model is developmental rather than defensive. It is not an exercise in catching a student out. The question is not "Prove you wrote this." It is "Explain how you became capable of this."

A discussion of that kind might cover the key learning experiences in the degree, the changes in the student's understanding, the mistakes they overcame and how, the capabilities they have developed, and the learning they intend to do next. Each of those topics is harder to bluff than a single question about a single essay, and each gives the student a chance to show what they have actually become.logs make this visible

Graduate Attributes as the Main Target

Most universities claim that their graduates should be critical thinkers, communicators, ethical professionals, independent learners and effective problem-solvers. Those claims appear in programme specifications and in the marketing. Yet these attributes are mostly assessed indirectly, through the marks for content, which were never designed to measure them.

A developmental model would assess them directly. The central question would be whether this graduate can operate effectively as a learner, a professional and a thinker. That is a different question from the one a transcript answers, which is what average mark the graduate achieved. The earlier page on bachelor's graduate attributes sets out what the degree is meant to develop, and this model would test whether it does.

Relationship to Learner Maturity

The developmental model connects directly to the maturity levels discussed in Personal Learning Maturity, which applies the logic of the Capability Maturity Model Integration to a learner. A learner might move from being dependent, to being managed, to being self-directed, to being self-improving. Those levels describe the same kind of growth that a portfolio would be designed to show.

I would put the central observation like this. The most important outcome of higher education may not be the knowledge acquired during the degree. It may be the ability to continue acquiring knowledge after it has ended. A developmental assessment is the only kind that could even attempt to measure that.

Expected Objections

There are several objections to this model, and I want to take them seriously rather than wave them away.

Objection 1: It would be too subjective

Portfolio and viva assessment involve judgement, and judgement is said to be unreliable. The response is that current assessment already contains a great deal of human judgement. Marking schemes are interpreted by people, and second marking, moderation and external examining exist because the judgement is known to vary. Those safeguards would remain available in a developmental model. The real question is not whether judgement exists, but whether it is directed at the right thing.

Objection 2: Students would stop working

If individual assignments carry no marks, the argument goes, students will not engage. The response is that doctoral students are largely assessed in this way, and they work very hard. Many professional apprenticeships also rely heavily on developmental feedback. This objection may reveal more about our assumptions about motivation than about the limits of the model. Motivation that depends entirely on a mark is not the same as motivation to learn.

Objection 3: Employers want simple grades

Employers do use classifications as a quick signal, and a shift away from them would have costs. But employers also say they value adaptability, independent learning, communication, judgement and resilience, and a classification tells them very little about any of these. A developmental assessment could provide richer evidence of exactly those characteristics, although it would take more effort to read.

Objection 4: Professional requirements demand competency checking

Disciplines such as medicine, pharmacy and nursing require explicit verification of specific skills, and rightly so. Nothing in this model prevents those competencies remaining mandatory. The model changes the overall structure of assessment. It does not remove discipline-specific requirements, and it would have to coexist with them.

Objection 5: Students could fake development

Students might construct artificial portfolios. That is possible, and I do not want to dismiss it. But maintaining a coherent developmental trajectory across several years is likely to be much harder than producing a single misleading assignment. The portfolio, the viva and the historical evidence reinforce one another, and an inconsistency in one of them is the kind of thing a conversation can bring to the surface. As the previous page argued, the aim is not detection but understanding.

Why Institutions Might Resist

Institutions have good reasons to be cautious. Regulation, comparability between institutions and between cohorts, scalability, existing classification systems, league tables and administrative complexity are all real concerns. A developmental model would need answers to each of them, and I do not have complete answers.

It is worth noticing, though, that universities evolved around industrial-era approaches to assessment partly because those approaches scale. A mark can be produced for ten thousand students by a standard procedure. A judgement about development is harder to produce at that scale. That observation explains why the current model persists. It does not show that the current model is the best possible educational design.

A Different View of Accreditation

The previous page suggested that accreditation rests on chains of academic judgement, and that this is easy to forget. This model makes that point more concrete. The implicit model at present is that a university certifies achievement. A developmental alternative would be closer to a community of scholars evaluating growth and certifying development.

The qualification would then say something closer to this: the person has undergone a transformation which we are prepared to endorse. At present it says something closer to this: the person accumulated enough marks. The first is a more demanding claim, and it is also a more honest one about what higher education is trying to do.

Conclusion

Universities often claim that their purpose is not merely to transfer knowledge but to develop people. Yet most assessment systems remain focused on measuring performance in individual tasks. A developmental model would take the transformative claim more seriously.

Students would still work. They would still be challenged, and they would still be assessed. But the primary question would change. Instead of asking "What marks did this learner achieve?", we would increasingly ask "What evidence is there that this learner has become more capable, independent, reflective and professional through the process of higher education?"

I am not sure the answer to that question would be easy to produce. I am fairly sure it would tell us more about the value of a degree than any weighted average of module results. That is probably enough reason to take the idea seriously, even if it is difficult to implement.

References

  1. Bretag, T., Harper, R., Burton, M., Ellis, C., Newton, P., Rozenberg, P., Saddiqui, S., & van Haeringen, K. (2019). Contract cheating: a survey of Australian university students. Studies in Higher Education, 44(11), 1837–1856. https://doi.org/10.1080/03075079.2018.1462788