Here are some thoughts on the challenge of generating centre-assessed grades, offered in solidarity – I don’t have to do this myself. I recognise the difficulties whilst also believing it’s the best way forward in the circumstances.
The full guidance from Ofqual is available here:
Ofqual has also published this reassuring blog https://ofqual.blog.gov.uk/2020/04/09/arrangements-for-summer-2020/ which suggests that some consultation on the details will happen after Easter.
There is also a video:
First of all I think there’s a process of acceptance that needs to happen. This whole thing is definitely happening and it won’t change. To help with that – and the whole grieving process around lost opportunities, perceived injustices and the abrupt course endings without the normal goodbyes – it’s worth remembering the purpose of the enterprise:
- We’re undertaking a serious recovery operation during a crisis, not trying to replace exams with a fully-worked teacher-assessment-based alternative. This isn’t the time to debate the fundamentals of exams vs teacher assessment. It’s a rescue mission.
- The emphasis is entirely on giving students recognition for their work to-date so that they can be awarded qualifications as fairly as possible in the least disruptive manner – for their future lives, both short term (ie, moving to university, not taking exams en masse in autumn) and long term (not being disadvantaged when qualifications matter in life, down the line).
- There is no element that relates to school performance in 2020. None. Nobody will care about or compare your school results; nobody can judge you or celebrate you – it’s just about the students as individuals. That’s just how it is. Painful as that might be to some -it’s also absolutely right.
- The grades schools give now have to have strong parity with those awarded in other years in order to retain their value. There might be room for some national ‘benefit of the doubt’ but that must have a limit – otherwise the 2020 cohort’s qualifications will be devalued and that disadvantages all of them.
- It is true that some students will get lower grades in this process than they might have got by taking exams; some will do better. On average, national results will be similar to previous years. There are re-sit opportunities but in the main this is a reality that we just have to accept.
I’ve seen people wrestle with it for various reasons – the pressure of responsibility, the bizarrely uncomfortable process (even the”disgrace”) of ranking, a strong sense of injustice for ‘turnaround’ schools – ie where this year’s results are expected to be much higher than before.. etc. These responses are understandable; they represent an inevitable grieving process that students won’t have the same opportunities to show what they can do. But, with very good reason, there will not be any change; I’d suggest focusing all energy on following the process as well as possible, with maximum integrity and a solutions-focused attitude; not raging against it. These measures have received wide acceptance as the ‘least bad’ option on the table. Personally I think the system is the best we could have hoped for; certainly there were no better ideas being floated.
Centre Assessed Grades:
The guidance is considered and sensible:
This should be a holistic professional judgement, balancing the different sources of evidence. Teachers and heads of department will have a good understanding of their students’ performance and how they compare to other students within the department/subject this year, and in previous years. We want heads of department and teachers to consider each student’s performance over the course of study and make a realistic judgement of the grade each student would have been most likely to get if they had taken their exam(s) in a subject and completed any non-exam assessment this summer. This could include U (ungraded).
It’s worth remembering a few things about bell-curves. Firstly, the way performance manifests itself at national level is strongly bell-curved. It’s no conspiracy. This graphic shows the spread of performance on trial Maths GCSEs (scaled so max score was 100) before new specs were introduced. Note how the grade divisions are introduced afterwards; these are the artificial constructs designed to fit a shared understanding of standards of attainment across subjects and over time. Ofqual will be trying to generate a similar effect without the exams so teachers will need to draw on their knowledge of previous patterns linking student performance in their subjects to final graded outcomes.
Importantly, grades are not rigidly fixed to quotas – they do link to a set of standards guided by reference tests and question-setting knowledge amongst examiners. Here, Eng Lit and Maths GCSEs in England in 2019 have their own bell-curves: A higher proportion of students get higher grades in English Lit.
So teachers and leaders need to look at a range of sources of information:
Assessments: – anything that generated grades and marks during the course. This might include any tests or assessments taken across the whole cohort. If you can generate grades and rankings based on existing data – at least as a starting point – it might make the subjective element much easier or at least confine it to the marginal cases.
Work samples: concrete examples of work to facilitate comparative judgements against exemplars. This will depend on whether you have access to a common set of work in the current circumstances.
NB: The guidance is clear that schools should not be trying to generate new assessments or work samples post March 20th – because it will be influenced unduly by the conditions different students have found themselves in since then.
Prior attainment: This should be a guide to the grade ‘most likely to get’ as part of a data-set, for triangulation. It might prompt debate about very big swings up or down from the starting point at subject or individual level. GCSE to A Level should be more reliable – eg using ALPS predictors – than KS2-KS4 pathways so handle with care. A student’s GCSE grade is not predetermined when they take their SATS – but the patterns in your school might hold at least when comparing subjects or lining performance across sets.
Teacher Judgement. You can’t simply pluck a grade out the air. “I just know Sasha would get an A”. That sense you have must have a basis. What is that basis? Try to be as concrete as possible – otherwise ranking will be near impossible, even if you can decide the grade. However, in the absence of any secure gradable assessment data, it is reasonable to see that Sasha’s prior attainment is X, to judge that she made good progress in general, worked hard, met expectations and consequently should have met the ‘most likely’ grade according to a progress tracker prediction.
Conversely, you might reasonably judge that James’ general progress was below expectations, based on a range of performances and interactions, and this supports your sense that he would ‘most likely’ get a grade below the tracker prediction. This is the kind of thing teachers are being asked to do: it’s not about giving every single student the benefit of the doubt; that will box you into a corner in the end.
Combining the information:
This will depend on what your school already has and does. If you have a set of grades already submitted – some form of internal estimates – it will be worth analysing them centrally to see if subjects are in the right ball park against trends and the baseline before asking people for new grades or to adjust what they’ve already submitted; you might not need more than you already have. However, everyone who signs the declaration will need to be confident that a systematic, fair and transparent process has been undertaken to produce the grades, using the ‘holistic’ approach.
Here’s a way to think of it: if you had to bet on the most likely grades Mo and James would get (imagining this could be known), and you really wanted to win that bet, which grades would you pick, based on all that you know? Certainly, if your approach is geared towards submitting the highest grade imaginable in every case, taking a punt on not being down-graded, I’d say that is doing it wrong!
From the guidance: ” …..it is important that the rank order of students is as accurate as possible.
Where there is more than one subject teacher, they will need to agree one rank order for all students within the centre who are taking that subject. To do this, teachers within a subject department will need to discuss the rank order and come to a shared view of the standard being applied within their centre. We recognise that this will be challenging for some centres and in some subjects, and in the current circumstances. However, the rank order is important for the statistical standardisation process.
In doing this, teachers should draw on examples of student work, including non- exam assessment where available. If two or more students are almost indistinguishable in terms of their subject performance (and are therefore judged likely to get the same grade) then it may be very difficult to put them into a rank order. However, exam boards will need a single rank order for all students”
To work out rank order, you can either choose to use teacher judgement at the macro level, to award grades, followed by assessment data to go more fine-grained, or the other way around. Personally I would tend towards generating grades and ranking using data from a selection of the available assessments in a fairly mechanistic manner, weighting various variables agreed within each department, and then look to see if it feels right based on teacher judgement. I know other people favour the opposite. I don’t think there’s a right answer here. If you have a mark book like the one above, you can turn anything you have into a scale, manipulate it to produce a score and then rank. OR you can just do it person by person.
It’s fairly obvious but the process might look like this below: Start with awarding grades; then assign top/middle/bottom grade groups; then another level within that. Only then rank within the fine groups which yields the overall rank.
The biggest challenge, as recognised in the guidance, is combining ranks from different parallel groups into one master ranking. Some online meetings or data sharing will be needed to thrash this out in all likelihood, subject by subject. The more you’ve used common assessments across groups, the easier that will be. If it’s all about teachers sharing accounts, you can end with different teachers lobbying for A to be above B or B to be above A when nobody knows both A and B. Before getting into this situation, agree the criteria you will use. If available, a comparison of any work samples might help; last resort might be attendance – or scores from the last time they both did the same test, however far back you have to go. One thing I would personally rule out is using KS2 scores to rank at GCSE Why? Because it just seems wrong to base that judgement on something five years old. For A level, total GCSE points would be a separator I’d be happier using.
Obviously you know that the students at the bottom of the list risk falling into the grade below if your data is subject to Ofqual’s statistical adjustments. To take the stress, pressure and subjectivity away from individual teachers, it might help for middle and senior leaders to apply a final one-remove checking process, referring back to the school profile and subject trends, bumping students up or down until the data fits a pattern that seems as secure as possible overall.
The Sign Off.
From the Guidance: “Exam boards may investigate any attempts to undermine this system which might be regarded as malpractice.”
Heads of subject and Heads of Centre need to sign off on the grades and ranking submitted. I’d be taking this pretty seriously. It means acting with the dispassionate objectivity of an exam board to the greatest extent possible. If you’ve emailed your staff or told them in a meeting to ‘add two grades on and let’s take our chances’ – as rumoured had happened in one tweet I saw last week – then you’re into whistleblowing territory already. People can get disciplined for this kind of thing. This is the time to say, hand on heart, we have enacted an objective grading and ranking process, using our professional judgement, and this is what we’ve come up with. That’s the best you can do for your students and the system as a whole.
For an excellent account of the thinking needed, here’s superb blog and video from Andy Byers at Framwellgate School, Durham. Really I could have saved some time and just said watch this. I love Andy’s tone, his pragmatic approach and his determination to balance a duty to students and to the whole system.
This video for students from Ofqual might be worth sending around:
[…] Tom Sherrington – https://teacherhead.com/2020/04/07/generating-centre-assessed-grades-2020/ […]
Absolutely right Tom.
You’ve articulated my thoughts exactly.
Well said and useful advice.
LikeLiked by 1 person
Glad you think so! And btw, a ‘gift box’ is on its way to you!
Great summary. For the future, I think comparative judgements certainly should be trialled more widely. When at the Royal Society we funded an education research fellowship looking at this in the maths and glad to see it has spread since.
LikeLiked by 1 person
[…] grade for every Year 11 and Year 13 student. As Tom Sherrington says in his excellent blog, ‘this is a rescue mission.’ All we can do is arrive at the least worst […]
Really Interesting summary – Thank you Tom
LikeLiked by 1 person
Thank you – this is a very useful summary/commentary. The author of The Learning Rainforest, making sense of the Summer 2020 Grades jungle – best wishes.
Fantastic piece Tom!
LikeLiked by 1 person
thoughts please on a child gaining a 3 in English having passed End 1 Autumn assessment jan 2019 4 Jan assessment 2020 4. excellent attendance and reports. School not taking in this child gets extra time in exams, prompts .