Wednesday, 4 July 2012

Education Emptiness

I have read too much Roger Schank. And I've been thinking too much. This is something best avoided, especially when marking student essays. Why? Because it makes you question what it is we're doing in Higher Education.

Assessing Essays

Marking essays is a soul-destroying task, as you have too little time to spend on each essay, and you have a large pile of essays to process.  Most students spend days and weeks on preparing their essays, so it always feels wrong to read and assess them in about half an hour, including writing up your feedback. This is very unsatisfactory, but otherwise one simply cannot turn around the marking in the allocated time.

But the worst thing about marking is its reductionist nature. An essay is a complex piece of writing, comprised of style, argument, expression of knowledge, understanding, interpretation, analysis, discussion etc etc. And all these different dimensions get conflated into a single point on a one-dimensional scale: a grade between about 40 and 70. This is just not right.

Many essays end up having the same numerical grade assigned to it, but they are not really comparable. One student might write eloquently but superficially, another provides deep insights with terrible grammar. One student might have a great idea, but not much understanding of the underlying concepts. Another one has solidly learned all what was required but lacks the creativity to apply the principles to a given problem. Yet, they all get the same numerical value. Different feedback, sure, but that does not really count for anything.

School children get detailed reports (besides a few simple letter grades), but in HE there are simply not the resources to do this, as there are too few staff and too many students, unless you are in Oxbridge. In principle that should not be an insurmountable problem.

You're doing it wrong!

After grading a mathematical crime is committed: the numerical grades are added up and averaged. This is simply not possible. The numbers are not numbers, they are labels that look like numbers. If we assigned the essays letters, then it would be more obvious: what is the average of A and B? But that is a completely different issue to be discussed on another day...

Essentially, then, after a lot of adding and averaging, the whole three years a student spends at university is reduced to a single label again, the degree classification. This is again an enormous reduction of a multitude of information into a single point out of four. And this point decides what possible career a student can then pursue...

As the degree class is so important (and expensive, especially for the incoming cohorts of students), this tends to be at the forefront of students' minds. This is of course a wild generalisation, and there are many exceptions, but from my experience a lot of students are primarily interested in getting good grades. Learning becomes secondary, and only the means to the end of achieving good grades. That means, curiosity, a central ingredient for successful learning, suffers, or rather, is redirected into finding out how to get grades. One cannot really blame students for trying to game the system, which is essentially what they learn to do in the end.

It does work elsewhere...

Postgraduate work is different, though, as it is less regimented. And, more importantly I think, PhD students do not get a grade. It's pass or fail. You either get a PhD, or you don't. There are of course, differences: you could get through with major corrections, minor correction, or no corrections. But nobody will know whether you scraped through with a 'revise and re-submit' or sailed through without any required corrections. If it works for PhDs, why not for UGs as well?

The problem is that we've got too many undergraduates, so there needs to be some differentiation. But why? Who wants it? Presumably those who employ graduates, so that they can see who is better or worse. But does the degree class really reflect vocational ability? I would doubt that, but in the end it is just another filter to reduce the 52 applications for each graduate job to a manageable number. With PhDs this is not so much of an issue, as you can get a more rounded picture by looking at their previous grades or even publications.


So what is the solution to this dilemma? It basically requires a system change, which is probably not feasible. Employers want differentiation, universities want to climb up the league tables (which nowadays tend to include employability metrics), and students want to have something that distinguishes them from the crowd. But in the process, education suffers. Learning is not really the focus of HE, and we're just churning out graduates who are good at spotting what is needed to get a good grade and doing just that.

We're assessing far too much, and it destroys what I think universities are all about: expanding your horizons, applying your knowledge and curiosity to interesting problems, be able to fail tasks without jeopardising your future career, and generally maturing and learning stuff.


  1. The 'essay' is the problem. It's like the Olympics only having the 1500 metres event and scrapping everything else. HE assessment is one dimensional, lazy and wrong-headed. The same applies to lectures, overlong and poor pedagogy. Good post.

  2. I think you have some very good points here, Oliver. From an employer viewpoint, I can say that I rarely notice the grades of applicants. We look mainly at the written content of their application, their experiences (if any) and then how they respond to the questions we ask them. Some of those questions relate directly to the work we do.
    So why not pass/fail indeed? Interesting thought.

  3. This comment has been removed by a blog administrator.