Professor Dylan Wiliam introduces his forthcoming pamphlet, to help schools design their own assessment systems…
The premise of my forthcoming publication – Redesigning Schooling – 8: Principled assessment design – is that any assessment system should be designed to support the curriculum in place in a school, rather than the curriculum being designed to fit the assessment system.
Or to put it another way, assessment should be the servant, not the master, of the learning.
For two decades, assessment in schools has been driven by the eight- (originally ten-) level structure that was proposed by the Task Group on Assessment and Testing (1987). The model proposed by the Task Group represented a radical departure from existing assessment and reporting practice, and may well have been different from what the government had expected. It seems that what Kenneth Baker, the Secretary of State for Education and Science at the time, had envisaged was a series of ‘benchmarks’ at the ages of 7, 11, 14 and 16. What the Task Group proposed was that all students should be on the same ladder, with a common achievement scale for all students of compulsory school age.
While the model had many strengths, the increasing pressure on schools to improve student achievement on national tests and examinations, combined with the behavior of school inspectors, resulted in a situation in which pursuit of levels (or sub-levels!) of achievement displaced the learning that the levels were meant to represent. It is therefore not surprising that the National Curriculum Expert Panel (of which I was a member) proposed that the use of the TGAT assessment model should be discontinued (James, Oates, Pollard, & Wiliam, 2011). This recommendation was accepted by the Secretary of State for Education, and in March 2014, the Department for Education announced that while statutory assessment would continue at the end of each key stage, levels of achievement for assessing achievement during key stages would not be replaced and it would be up to each school to decide how to monitor students’ progress towards the expectations for the end of each key stage (Department for Education, 2014).
Understandably, this decision was far from universally popular. Many have argued that the parents were, at last, beginning to understand the system—indeed, many parents had themselves experienced the system at school when they were children. Others have argued that for all its faults, the system allowed achievement in all school subjects to be reported on a common scale.
However, against this, it does seem that the idea of age-independent levels of achievement was better suited to some subjects than others, and schools are now free to design assessment systems that meet their needs and the needs of their students. Formal assessments will continue at the end of primary and secondary schooling, in the form of key stage 2 tests and GCSE examinations, but it will be up to the schools to decide how to get their students to the highest achievement possible in these formal assessments. Schools will therefore be able to plan for the long term. Primary schools will receive children at the age of four or five, and it will be for them to decide what kind of curriculum will suit their local needs best. Secondary schools will receive students at the age of 11, and be able to plan five-year programmes of study that will allow their students to get the best possible grades at GCSE, and to prepare them for the complex ever-changing world of the 21st century.
Many schools are currently choosing, for the time being at least, to continue with the current levels of achievement, and this may be entirely appropriate in terms of managing workload. But it is important to note that there will be no straightforward way to carry the existing levels of achievement forward into the new national curriculum, since the new national curriculum will not provide descriptions of the levels. Even when schools choose to use something similar to the current levels of achievement, it may be appropriate to consider whether such levels of achievement are equally appropriate in all subjects. It may also be useful to consider whether dividing levels up into sub-levels makes sense for all subjects, and even where subdivisions do make sense, it may be that two, or four subdivisions may be more appropriate than three, as is currently the case.
The freedom presented by the recent changes is obviously a little daunting, since schools are losing supports on which they have relied for many years – indeed the majority of our teachers have known nothing else – but the opportunity to design an assessment system that works for the school, rather than the other way round, represents an extraordinary opportunity for schools.
The aim of my pamphlet, then, is to provide schools with some ‘tools for thinking’ as they begin the task of designing their own assessment system.
A school’s assessment system has many jobs to do. It must provide useful information for students, parents, teachers and leaders about the progress being made by students. At all times, the system should provide information about important checkpoints or milestones in learning but, as students approach high-stakes assessments such as key stage 2 tests and GCSE examinations, it should also provide information about the likely outcomes of those assessments. Some of the information generated by the system will be carefully recorded while other assessment information will be of a more ephemeral nature, informing decisions about next steps in teaching, for example, where formal recording could be unduly burdensome, or even counterproductive. The information also needs to be generated in a timely manner. For some kinds of assessment information, the ‘sell-by’ date is very close to the collection of the evidence, while other kinds of assessment information has a longer ‘shelf-life’. Inevitably, these different demands on the assessment will create conflicts and trade-offs, many of which cannot be resolved easily, if at all.
The aim of my new pamphlet is not, therefore, to tell schools how to design their assessment system, but rather to help them understand better what trade-offs they are making in designing their assessment system, and, in particular, what the costs and benefits of various decisions will be.
It is also important to realise that there is a great deal of science here. By that I mean that there are some issues in which personal opinions play very little part, and this is rather rare in education. As many people have pointed out (see, for example, Lagemann, 2000) education lacks strong disciplinary foundations, and there is relatively little that could be regarded as ‘reliable knowledge’ (Ziman, 1978). In the field of assessment, however, there are some very well-understood technical aspects where failure to appreciate the technical details mean that one can do things that are just wrong. This is particularly true for the issue of reliability, where schools routinely say things to students or parents that are not just unprofessional (in that they contradict professional standards or codes), but actually incorrect.
In making my selections on which aspects of assessment to consider in the pamphlet, I have steered clear of trying to propose a model for a school assessment system—such a model might be suitable for some schools but not for others. Instead, I have tried to focus on some key principles for assessment that will help school leaders and others to think about assessment in a more disciplined way.
These principles will often, if not always be in tension, so there is unlikely to be any assessment system that satisfies all the principles. But by making these principles explicit, my hope is that the compromises that are being made are clear and explicit, and above all, planned, rather than just being the unintended consequences of decisions taken by different people, in different parts of the school, at different times.
Dylan’s pamphlet will be sent to all SSAT member schools in the coming month. Copies or Redesigning Schooling are available to purchase from our website.