This blog post is based on an article originally published on LinkedIn on 16 August 2020 – click here to read in full.

The fallout from A-level and GCSE results will be uncomfortable for government and upsetting and challenging for teachers and students alike. Arguments over whether this year’s results are robust and fair miss one key issue.

Put simply: "Has the exam system in England ever been robust and fair for individual pupils?"

For those of us who did well in exams and whose children also did well, it is too easy to be confident. Accepting that our success and others’ failure is a systemic problem, not a result of competence and capability, is not easy.

Let me be clear: I do not have confidence in the exam system in England as a measure either of success or capability. 

[…] Try this as a thought experiment. Imagine that I gave an exam paper submission to 100 examiners. Let me assume that it "objectively" is a C grade.

Would all 100 examiners give it a C? If not, what is the spread? Is the spread the same for English literature, physics and geography, as just three examples? If you cannot provide clear evidenced answers to these questions, how can you be confident that the system is objective?

If we look at the examiners, the same challenge appears. Are all examiners equally consistent in their marking, or do some tend to mark up or down? Where is the evidence, reviewed and published to demonstrate robustness?

We also know that the month you are born still has an effect on GCSE grades. What is robust about that?

[…] I have known children who have missed out on grades after divorce, separation and death of parents, siblings and pets. I cannot objectively give a measure of the impact, but then neither can the exam system. I would add that I suspect a classmate of mine missed out because of hayfever. Children with health issues such as leukaemia and asthma whose schooling is disrupted have had their grades affected every year, not just this one.

So, the high stakes exam system is, for me, a winner-takes-all loaded gun embedding inequality and privilege in the outcomes.

Can we do better? Well, if we want to use exams, then each paper needs to be marked by say five independent assessors. If they all agree on a "B" then that is a measure of confidence. This is often a model used for assessing loans, grants and investments in businesses. It does not guarantee success of course, but what it does is reduce reliance on potentially biased individuals. If I was an examiner and woke up today in a foul mood, would I mark a paper the same today as yesterday? I would not bet on it.

The really interesting cases in my experience are where you get 2As, a C and 2Ds, for instance. In my experience, I've seen it more often in "creative subjects", but some non-traditional thinkers in subjects like mathematics (a highly creative discipline, by the way) often don't fit the narrow models of assessment of our exam system. The problem with this example of bringing people together to try get a consensus on a "B" is that it eliminates the value that comes from the diverse views and the richness of the different perceptions.

So, for me, for a system to be robust it has to have more than one measure to allow the individual, parents, universities, FE and employers access to a richer view of an individual. If someone got an ABBCD in English that is as interesting as someone who got straight Bs. 

[…] There are already models that command respect in grading skill levels. Parents are quite happy if a child is doing grade 6 piano and grade 2 flute at the same time. They are quite happy for a child to sit when ready and have the chance to resit. Yet in the school setting the pressure is there for a child to be at level 8 say for all subjects. That puts unnecessary pressure on pupils, teachers and schools.

Imagine how society would react if you could only take the driving test once at 17 and barriers were raised to stop you retaking it.

[…] This year’s bizarre algorithmic system is not robust, but then we have never had a robust system as far as I am concerned. Let's open our eyes and build something that we should have more confidence in. Carpe diem.

Join the discussion: share your views in the comments below (member login required).