My favourite definition of learning is “a change in long term memory”, which means that something has been learned and then remembered permanently.
On the face of it this seems like common sense but the way in which school assessment systems are set up often works against this as an objective
The problem is that many schools confuse performance with learning. If learning can be defined as a permanent change, we can define performance as as a change that doesn’t last. The confusion is compounded by the fact that performance is an important part of a process that might, so long as it is accompanied by other strategies, eventually lead to durable learning.
In a school which confuses performance with learning, the progress of pupils is often measured by what they can do at best the end of the unit, and at worst at the end of a single lesson (or indeed after twenty minutes for those who remember the dreaded mini-plenary). This means that the data going into spreadsheets will only reflect what they have remembered over a short period of time. This can create a misleadingly optimistic picture which suggests that children are doing far better than they actually are. This is because, as we all do, pupils forget over time; just because someone can do something after a day doesn’t mean they will be able to do it a year later, especially if they have been given intensive support to get to that point in the first place.
Schools operating systems which are focused on performance incentivise the prioritisation of the short term over the long. This can mean too much emphasis on one off lesson observations and the progress pupils make in one hour. Teachers working in such systems might quite naturally concentrate on making sure pupils do well on tests focused on what has been covered most recently and may not spend time going back to make sure that topics covered in the past have been remembered.
This, in a context in which almost all Y11 outcomes are based on a set of terminal exams sat at the end of the course, can be a disaster. If KS4 lasts three years (9-11) as it does in many schools, pupils may not be supported to revise content they studied at the beginning of the course until right at the end. This means, of course, that many will have completely forgotten what they were taught three years before. It also means that pupils are not required or expected to revise large amounts of content until the stakes are terrifyingly high; far better would be to help pupils develop effective revision strategies earlier in their school lives so they are well embedded before Y11.
The most impactful way to do this is to make sure that all assessments are tests of everything that has been covered up to the point the pupils sit it. The table below, derived from an example shared by Michael Fordham a couple of years ago, shows how this might be done over one year in any subject.
This simple model means that teachers will be much more likely to focus on re-teaching and re-visiting the whole curriculum because if they do not, their pupils are likely to perform increasingly poorly on tests. It also makes it more likely that when pupils study at home, they will need to study the entire curriculum, which makes it more likely they will remember more of it.
In schools that follow this model teachers are much more likely to interleave and expose their pupils to more retrieval practice. This means that pupils are much more likely to remember more of what they are taught for longer.
While this model of assessment is a strong one it is not without its issues.
The first of these is that unless a school dramatically increases the amount of time it spends on assessment, each test will only be able to cover an increasingly small sample of what has been taught. Those feeling uneasy about this (as I was originally), may find it helpful to consider what exactly they are assessing. If we are testing true learning then we are actually assessing more, because we are drawing from everything that has ever been covered. This is also much fairer on pupils too, because a pupil can gain marks from a wider domain of knowledge and won’t be unduly punished for struggling on a single unit.
For once, exam accountability is helpful. Testing in the way I’ve outlined is authentic to GCSEs which sample from a wide domain in a deliberately unpredictable fashion; covering a topic does not guarantee that it will come up on the exam, which is why good pupils prepare for everything.
Assessments in history for pupils in my MAT, informed by the principles of learning I’ve outlined now look increasingly like this:
SECTION A: (50% of marks)
Short answer and multiple choice questions based on the entire domain of knowledge taught that year.
SECTION B: (15%) of marks)
One short extended writing question (a paragraph or two) based on the topic covered most recently.
SECTION C: (35% of marks)
One longer extended writing question from a choice of essay questions drawn from the entire domain of knowledge taught that year.
This, I feel, offers a reasonable balance between testing what has been covered most recently, and everything else that has been covered too.