Levels: where it all went wrong

Levels are dead! Long live… well, what? I am quite a fan of the new 2014 National Curriculum for history. I very much wish that Michael Gove had been brave enough to do what was originally intended in the National Curriculum and to have history compulsory to 16 (we can blame Ken Clarke for it being pegged back to 14) which would bring us in line with most of the rest of Europe. As a statement about what pupils should know about the past, however, I’m generally in agreement with its emphases. What I am most pleased about, however, is that the new National Curriculum ditched the level descriptors. I want to use this post to explain why.

To understand this, we need to take a trip back to the earliest National Curriculum. Let’s take a quick look at one of the Attainment Targets for the 1991 National Curriculum. It took a second-order concept (causation) and broke it down into ten ‘steps’, with each being supposedly harder than that which preceded it.

• Level 1 – Recognise everyday time conventions
• Level 2 – Demonstrate, by reference to stories of the past, an awareness that actions have consequences
• Level 3 – Demonstrate an awareness of human motivation illustrated by reference to events of the past
• Level 4 – Understand that historical events usually have more than one cause and consequence
• Level 5 – Understand that historical events have different types of causes and consequences
• Level 6 – When explaining historical issues, place some causes and consequences in a sensible order of importance
• Level 7 – When examining historical issues, can draw the distinction between causes, motives and reasons
• Level 8 – Produce a well-argued hierarchy of causes for complex historical issues
• Level 9 – Demonstrate an awareness of the problems inherent in the idea of causation
• Level 10 – Demonstrate a clear understanding of the complexities of the relationship between cause, consequence and change

These levels did not survive very long. Primarily this was because they were so unwieldy that they were almost impossible to use. Alongside the practicalities, however, was a deeper problem: the levels themselves don’t work historically. Try this as an exercise: cut out the levels without the numbers attached and then try and place them in order. It’s remarkably difficult to do. Why, for example, is ‘drawing a distinction between causes, motives and reasons’ necessarily harder than placing ‘some causes and consequences in a sensible order of importance’? Why does understanding that ‘events have more than one cause and consequence’ come higher than an ‘awareness of human motivation’? Why is the concept of ‘consequence’ easier than that of ‘cause’?[1]

Not only does the line of progression built into these levels make no sense but it also makes the mistake of turning a second-order concept into a skill. This is a distinction that is often lost. ‘Causation’ is not a skill; ‘writing a causal essay’ might be understood as a skill. The former is a concept; the latter is the application of the concept. The use of verbs such as ‘place’, ‘draw’ and ‘produce’ are indicative of this. And there’s another problem. The 1991 levels, in aiming to be generic, ripped the concepts away from the substantive periods that were being explained. The model assumes that an attempt to explain the causes of the First World War is essentially the same thing as an attempt to explain the causes of the Reformation, or the collapse of the Roman Empire. Obviously there are similarities that can be drawn, but adopting a common mark scheme for any question puts the cart before the horse: it sets out the hoops through which pupils need to jump, and then forces the substantive period into those hoops.

These problems were well understood by 1995. The 1991 levels were (a) too simplistic a progression model, (b) confused concept and skill and (c) separated substantive knowledge away from the progression model. The 1995 Level Descriptors were an attempt to resolve this. Below is the top level (Exceptional Performance) from the 2008 curriculum.

Pupils show a confident and extensive knowledge and understanding of local, national and international history. They use this to frame and pursue enquiries about historical change and continuity, diversity and causation, constructing well-substantiated, analytic arguments within a wide frame of historical reference. They analyse links between events and developments that took place in different countries and in different periods. When exploring historical interpretations and judgements about significance, pupils construct convincing and substantiated arguments and evaluations based on their understanding of the historical context. They evaluate critically a wide range of sources, reaching substantiated conclusions independently. They use historical terminology confidently, reflectively and critically. They consistently produce precise and coherent narratives, descriptions and explanations.

These level descriptors, first introduced in 1995, were designed to prevent the different parts of history being separated out. Substantive knowledge, second-order concepts and ‘skills’ were bound together into a holistic (horrible word – sorry – but it works here) statement. Importantly, these statements were designed to be used only at the end of a Key Stage. They were absolutely not designed to be used to mark individual pieces of work. The idea was that, at the end of a Key Stage, teachers would make a ‘best-fit’ judgement about what level to give to every pupil. They lack the specificity needed to mark a particular piece of work, but, as a broad statement, they work tolerably well.

Except this is not how they were used. Under demands to measure progress at smaller intervals, history departments began splitting the level descriptors up. They were, for example, split ‘horizontally’. Particular strands such as ‘causation’ and ‘evidential reasoning’ were once again separated, recreating the 1991 Attainment Targets, and thus recreating all of the problems this entailed. It did not take long before some schools started marking particular pieces of work with the levels, a task for which they are not at all suited due to a lack of specificity.

Then it got worse and the  separated strands were sub-divided vertically. Pupils were now not just ‘Level 5’ but ‘Level 5a’ or ‘Level 5b’. Take this nice example. The Level 5 descriptor included the  line ‘begin to make links’. The Level 6 descriptor included ‘make links’. So how might we sub-divide Level 5 further? Presumably pupils were ‘beginning to begin to make links’. It’s farcical and would be funny had not a generation of pupils had their work marked using this model. Under the well-meaning aura of assessment for learning this went even further, and, with the demand for ‘pupil-friendly’ assessment criteria, these bastardised level descriptors were turned into meaningless grids that were stuck in the pages of pupil books.

The Level Descriptors, from 1995 to 2014, were never going to be great. Had they been used for their original purpose – a ‘best-fit’ description of a pupil’s abilities at the end of a Key Stage – then they might have been fit for purpose. The way in which they were used, however, made a mockery of Key Stage 3 history assessment. In 2004 Sally Burnham and Geraint Brown wrote an article in Teaching History making just this point: it took a decade for the levels to be finally scrapped. Rumour has it that they are to publish an article soon entitled ‘We told you so…’.

So what now? The level descriptors have gone, and probably the worst thing we could do is to keep on using the old descriptors; I’d like to think that the profession of history teachers has desired their demise for years, and will now jump at the opportunity to ditch them. A void, however, is hard to fill, particular when there are stringent accountability demands that requires some system to be in place.

I have but one plea in this. Please, ***please***, do not let us recreate the 1991 Attainment Targets. We tried this in 1991 and it did not work. Ever since, schools have been trying to convert the Level Descriptors back into the 1991 Attainment Targets and, unsurprisingly, it once again has not worked. I shall stick my neck out here and say that any assessment model that takes second-order concepts (such as causation, change, consequence etc.) and tries to sub-divide them into discrete stages that can meaningfully be applied to pupil work is bound to fail.

What, then, should come in its place? I intend to write something in the next few days addressing this. In short, however, I think the answer lies in placing substantive knowledge at the heart of our assessment model.

[1] I’m grateful to Christine Counsell for this activity which is used at the beginning of the Cambridge History PGCE course.