Thinking about assessment and accountability again. I adapted this from a letter I wrote to the then Ed Select Committee.
The problem of accountability
If we take it to be the case that teachers and schools need to be ‘held to account’, then we need to ask ourselves some questions.
Held to account for what?
The answer to this is crucial. For a long time we were held to account for pupil ‘attainment’. Recently there has been the reasonable suggestion that there are many factors outside of our control which impact on attainment, and that progress might be a better measure of how good or bad we are. The measurement of progress nevertheless, remains a massive challenge, in spite of attempts to contextualise it, or to use national trends for comparison: baseline data is not reliable (it’s ludicrous to believe that you can use the behaviour of 4-year-olds to derive data that will hold teachers to account at the end of KS2 – and beyond!); pupils do not make standard amounts of progress; domains at the start and end of the progress measurements are different (GCSE art teachers, beware!); cohorts are different; significance is difficult with small groups, etc.
Whilst ‘progress’ seems at first fairer and a preferable measure to ‘attainment’, neither is sufficient for our purposes – not the way currently measured and not when matched against the aspirations of the National Curriculum.
Does the current system even serve the right purpose?
In the drive to measure ‘attainment’ and ‘progress’, I think we sometimes forget that we are using these things as proxies for the quality of the education being provided. We need to return to the drawing board for how we might ensure this happens. Currently we use an assessment system that cannot do this; the measurement is too narrow, subject to chance variables, and too much driven by fear of failure, leading to all the perverse incentives the assessment experts have been writing about for so many decades. A quality primary education is not ensured by testing the very few items that are currently measured at the end of KS2, any more than a quality factory is ensured by eating one of its biscuits.
I think schools and teachers do want to provide a good education to their pupils and that a climate of fear is unnecessary and counterproductive. Real accountability must involve a move away from looking only at outcomes and focus instead on quality input: we need well-educated teachers with excellent and maintained subject knowledge, quality text books produced by experts and thoroughly vetted by the profession, online materials and required reading, use of evidence and avoidance of fads. Quality training is essential, as is development and retention of teachers with expertise.
How can we ensure accountability where it counts?
Realistically, I don’t expect a quick move away from summative assessments for the purpose of accountability, in spite of all the arguments against. But I feel that we could address some of the issues that arise with the current system by generating and providing (to the DfE if need be) not less but more information:
- Frequent, low-stakes tests help both teaching and learning – require/provide tests throughout the year, every year.
- Fine grained, specific tests provide useful information – test what we want to know about – keep that data.
- Assessing the same domain more than once and in different ways helps to reduce unreliability – do not rely on one single end of year test.
- Testing earlier in the cycle gives useful feedback for teaching – do not wait until the end of the year or the end of the Key Stage.
- Random selection from a broad range of criteria helps to reduce ‘teaching to the test’ – test knowledge in all curriculum areas without publishing a narrow list of criteria.
- Use assessment experts and design assessments that test what we want pupils to know or do. Criteria need to be reasonable – not obscure and mystical as they have been recently.
If these aspects were applied to an assessment system throughout the primary phase, I believe we could enhance learning, improve accountability in what really matters and provide vast amounts of data.
We really need to make better use of technology at all stages; this is the only way in which we can feasibly make assessment serve multiple purposes. There would need to be a move away from the high stakes pass/fail system which is not fit for purpose, towards a timely monitoring and feedback system that could alert all stakeholders to issues and provide useful tools for intervention. Data collected from continuous low-stakes assessments provides a far more valid picture of teaching and learning.