Marking system is failing

When I became a lecturer two decades ago, I had a problem. Faced with a big pile of essays, I didn't know how to mark them. I consulted an experienced colleague, soon to retire after a distinguished career. "Don't ask me," he said modestly, "I don't know either."

At that time, there were neither guidelines nor reference points. All I had was the university's marking scale (40 per cent = pass, 50 = 2:2, 60 = 2:1, 70 = first), my own experience as a student and a personal view of what degree grades meant. The challenge was to attach grades to what was before me.

I muddled through, losing a little sleep occasionally but trying to maintain integrity. The experience left me with a strong sense that marking is an imperfect art. I cannot accurately describe the difference between a 2:2 and a 2:1, but I do "know it when I see it". Over the years, review by external examiners suggests that my approach works, or at least does not differ markedly from their own. With the advent of teaching quality assessments and greater public scrutiny of degree courses, we started to be given charts detailing what marks meant.

These were intended to bring objectivity to assessment and agreement about grades. They are also given to students as part of the discourse about their progress, symbolising the openness that apparently characterises university education. But I have my doubts. Marking is scarcely more objective than it used to be. I have three difficulties. Firstly, marking is carried out back to front. Like my colleagues, I first assess a piece of work for the degree class it represents, then allocate a mark. I still use rule-of-thumb criteria I set years ago.

An exam answer, for example, passes the 2:1/first borderline only if the student gives me more than I gave them: however accurate, complete and informative the answer, they can reach the higher class only if they have read beyond the lecture material and show independence of thought. Having decided the class, I position the work within the mark range and award a value. Despite working backwards, the outcome invariably chimes with the marking scheme, but that just reflects the imprecision of mark descriptors.

My second problem is with the marking scale. External examiners constantly entreat us to use "a greater range of marks". They apparently forget that 30 per cent of the scale (70 per cent and above) designates first-class performance. Simply shifting more marks to that region is a sure-fire recipe for grade inflation. In fact, the hard work of discriminating different performances mostly takes place between 50 per cent and 70 per cent, a mere fifth of the range. The point is that the scale is not linear. Furthermore, setting the pass mark at 40 rather than 50 per cent means that a student can progress by understanding less than half the course material. Is that defensible?

My third issue is an arithmetical one. How can the twisted scale be used for linearly marked items? A typical science module includes multiple choice question papers (MCQs), numerical tests and digital assessments generating scores on a linear scale. To arrive at an overall mark for the module these somehow have to be merged with non-linear marks from coursework and exams.

No one seems to know how to do this. Cleverly set questions may distribute marks in the right place, but do not solve the problem. I don't know if students appreciate this irrationality, but the bright ones might enhance their degree prospects by choosing modules with a high proportion of linear assessment.

So my conclusion is that marking remains subjective. This brings little comfort to those defending the security of academic standards, but it's true. Like my long-retired colleague, I still don't know how to do it, although I feel less worried about it than I used to.

The marks I award depend on my judgment and I'm content if that reflects my professional skill and integrity. The problem of the bent marking scale is more tricky. It's been embedded for centuries, for reasons long forgotten. At some point it will need reforming. But I don't expect a higher education marking revolution any time soon.

Above all, we should acknowledge the subjectivity of assessment and stop pretending that marking is something it isn't.