Beyond Levels Part Two: summative and formative assessment
In my last post I commented briefly on why I thought task-specific mark schemes were more appropriate than ladder-like progression models for the assessment of pupil work. In this post I want to turn my attention to formative feedback. In particular, I want to develop the points I made in this post about letting our subject specialism drive our teaching practice.
All formative feedback requires a summative stage: we have to judge what is good and bad about a piece of work before we can advise a pupil on how to improve it. I am quite a fan of formative feedback and respect the considerable body of evidence that points towards its strengths. I think, however, that the process is often proceeds from a weak starting point in schools, and this is the summative judgement that we make before giving the formative feedback.
My hunch here is that the usual summative judgement we make before giving formative feedback is based on a set of assessment criteria. These could be National Curriculum levels, public examination mark schemes or some other assessment tool that we use regularly in schools. The emphasis seems natural: we want pupils to perform well in tests, they want to do well in tests, and therefore it makes sense to give feedback based on ‘how to get from a D to a C’ or ‘from a Level 5 to a Level 6’. One has only to glance at classroom displays, homework diaries and pupil notebooks to see that this approach to formative feedback is deeply embedded in our teaching culture.
I am not convinced, however, that this is right. If you have not yet read Daisy Christodoulou’s piece on teaching to the test, then now is the time to do it. In the post she draws on Koretz to argue that teaching to the requirements of an assessment, all of which are necessarily imperfect, distracts us from teaching the domain – the thing that we want to teach – whether it be mathematics, history or knitting. The test becomes the curriculum which becomes the domain in the eyes of pupils and teachers.
So how might we avoid this? I think the key is that we need to make a clear distinction between the purposes of formative and summative assessment. In summative assessment – such as exams, end of unit tests and so on – we are judging pupils against a (necessarily imperfect) mark scheme. When giving formative feedback, however, we should be judging pupils against the domain – i.e. what is good history, or good chemistry. At times this might overlap with a mark scheme, but often it will not. In short, we should be giving pupils feedback that will help them get better at the subject, and not at the test.
- Summative assessment – judge pupils against a mark scheme.
- Formative assessment – judge pupils against the domain.
This is a simple distinction, and yet one which our well-meaning emphasis on assessment for learning in recent years had led us to overlook.
In a fit of ‘let’s just go for it and see what happens’ we threw out levels at the start of the year and wrote descriptions (oh what joy to use words not numbers!) for each enquiry at KS3. Each enquiry has a description for ‘what we expect you will be able to do’ and also has a ‘moving beyond the enquiry’. They are short, they are shared with the pupils and we are reviewing them at the end of each enquiry to change the wording. They are simple, they are far from perfect, but they feel a lot more useful than levels already. What is more they are promoting dialogue and informing our formative assessment. They are more concerned with the domain than any markscheme. Where enquiries have some sort of summative assessment the markschemes are then written from the descriptions and therefore task specific. What I haven’t yet cracked is how to help colleagues who have to appease people requiring a whole-school level standardised progression model.
Sounds like exciting times at The Mount! Would be very interested to see what you have been doing. We are also removing our level descriptors and moving to assessments which are marked in a similar way to Michael’s previous post. Already it feels utterly liberating, although it has put a real focus on us getting a shared understanding of what really good history looks like.
With regard to the assessments themselves, I think to some extent we have to accept that teaching to the test will always be some part of educational culture, what matters is that the test helps develop the domain as well. For a good example of this, the Queensland system of assessment has teachers involved in setting assessments at a regional level, therefore devising mark schemes which address both the demand for testing and domain knowledge.
I agree on the issue of formative assessment – have never been a fan of commenting of students ability to meet test criteria in everyday work – this becomes an increasing demand on teachers however.
I’ve found your recent posts on assessment interesting and thought-provoking.
I have always tried to focus on what good history is – rather than what the mark-scheme requires – in assessment. In essence this places all students on a continuum in my own mind to working as historians – so, for example, I used to give some of my GCSE students A-level texts comparing interpretations of aspects of Nazi Germany, not because the GCSE specification demanded this, but because it would challenge, intrigue and prepare them for the next step in their career. I always hoped that if they did use this in the exam, the examiner would be suitably impressed.
I’m not sure that AfL necessarily contradicts this – if done well. Good AfL means ensuring student understanding of what success looks like and feedback to help students meet that standard. It is only if we misapply and slavishly follow generic mark-schemes that we undermine this… their very meaninglessness, as you have argued recently, makes this a poor way to approach AfL!
The unfortunate question remains how we then fit this into whole-school accountability structures. I’ve always proceeded on a best-guess, roughly at a level X approach… I wonder what an overarching web of standards would look like – and whether the history education community could start to construct one.