The pandemic has cast many assumptions about how education could or should unfold into sharp relief. Like many others, I’ve been wondering about the positives we might find in our current situation and how – or whether – we can salvage anything when schools eventually return to normal. One area that seems to beg for reform is the way the exam season currently plays out.

Here are some of the factors to consider:

  1. Accountability creates huge pressures on teachers which are, inevitably, passed on to students. Is there a way to break this chain?
  2. Along with these pressures, the quantity of examinations students have to take make Year 11 an uncomfortable experience for most students with too many reporting crippling levels of stress.
  3. Accountability pressure also encourages gaming. Schools are encouraged to focus resources on Key Stage 4 and Year 11 in often unhelpful ways.
  4. Teacher assessment is vulnerable to unconscious biases that disproportionately disadvantage children from lower SES profiles and ethnic minorities. Because of this, exams are the ‘best worst option’.
  5. This year, due to the impossibility of students sitting exams, GCSE grades are being awarded via teacher assessment mediated by the examination boards and Ofqual.

Last year, Becky Allen outlined a system which could end the warping pressure of high-stakes accountability. In her post The Ungameable Game, she proposed the idea that each year a lottery would be held in which the measures used to hold schools to account would be randomly selected. Each year, schools would be assessed against different metrics, resulting in an ‘ungameable game’. This was a fascinating idea but as Becky herself acknowledged, “It is unlikely that any regulator or politician would choose to adopt such a system of messy regulation.”

But, could we perhaps opt for a less messy, more modest version of her proposal?

Here’s the idea my wife, Rosie Sisson has come up with. What if schools were asked to repeat the internal ranking of students necessitated this year by the pandemic every year? Let’s say that by the end of Easter, schools submit their rank orders for exam boards to mediate and then exam boards randomly allocate which students have to sit exams in each subject. If schools were given two weeks’ notice on which students would be sitting which exams there would be less scope for gaming and intervention. As long as the results of these exams are broadly in line with teacher assessments, the rank order produced by the school is used to determine students’ grades, but if the sample sitting the exam produce wildly divergent results, the rank orders would be adjusted to better align with those produced by other schools. For whole cohort subjects like English, maths and science this would be more straightforward, but for subjects with only a small number of entries it might be that each student entered has to sit the exams (This reflects what used to happen when coursework was requested for moderation: samples were randomly selected for subjects with large numbers of candidates, whereas as smaller subject would send off the entire sample.)

The potential benefits are:

  1. As schools would know in advance that they would have to compile a rank order for all subjects students might be more inclined to take interim assessment seriously.
  2. As students would only take a sample of their exam papers, the stress of sitting 20+ different exams would be reduced.
  3. As schools would not know which students would be sitting which exams it would harder to justify using curriculum time on ‘teaching to the test’.
  4. The weaknesses of teacher assessment would be balanced against the greater reliability and validity of sampled exam results.

Potential problems:

  1. What if an individual student does much better (or much worse) in their sampled exams than in their teacher assessment? Would their individual rank in the order be changed? Is this fair on those students not sampled? Should students be given the option to sit any exams where they believe they can outperform their teachers’ assessments?
  2. Inevitably, when the samples are announced, schools would be likely to spend the next two weeks in furious preparation and intervention. But, this is probably preferable to the current situation and, as all schools would be in the same position, shouldn’t be any more or less unfair.
  3. What if the students due to sampled fall ill, fail to show up, or suffer some other unanticipated mishap? For instance, what if some students had no exams and others had a full compliment? There would have to be some room for negotiation, some leeway offered.
  4. This year’s test case will indicate how difficult it is for schools to produce accurate rank orders. Presumably, with preparation and trining this would get easier over time.

There are no doubt myriad other problems I’ve failed to anticipate, so I’d be grateful if you could add your thoughts or concerns in the comments below. But what do you think? Would this be workable? Might it be an improvement on what we currently do? It also occurs to me that this would be equally possible – in fact, much easier – to implement for Ks2 SATS.