,
13 tweets,
4 min read
Read on Twitter
Response to @solomon_teach latest blog post medium.com/solomonkingsno… I agree completely we should design our curriculum to focus on the fundamentals, & rushing through advanced content in a superficial way is good for no-one. However, I don't agree this is a GCSE design problem.
GCSEs are summative tests, and summative tests are designed to provide shared meanings of the performance of large numbers of students. This inevitably means they have to differentiate between wide ranges of attainment, so they need questions which differentiate.
Daniel Koretz addresses exactly this point at length in chapter 2 of Measuring Up (which I thoroughly recommend.) 
To get reliable grades you want your average candidate to score half marks, & top 1% to score almost full marks. If you designed it so that grade 4 corresponded to 80% of the marks, you’d have v poor reliability for grades 5+, as they’d be separated by a tiny handful of marks.
Why do we need to know how pupils are doing relative to each other? I hear you ask. Why can't we just have a statement of what they know and can do? Well we don't need to know relative performance that often - which is why the GCSE only happens at the end of five years of school.
But whilst we don't need to know THAT often, we still do need to know. Employers, universities, & 6th forms do want some idea of how pupils have performed relative to each other. Knowing that 100% of them have met a mastery bar wouldn't be enough for their purposes.
And the whole issue here is purposes. The purpose of the GCSE - to provide a shared meaning - is completely legitimate. But it's not the same as the purposes teachers are mostly concerned with: to teach subj content, to know what a student can know & do, to promote mastery
The GCSE is not a mastery test and nor should it be. As such, a relatively low raw score on it is not the sign of abject failure @solomon_teach assumes. In general, raw scores are misleading unless you know the test's difficulty. I write about this here. tes.com/magazine/artic…
Problems occur when we use the non-mastery GCSE as a way of testing mastery. That's a bad idea! I write about this at more length in this blog post here. daisychristodoulou.com/2019/05/what-i… And at even more length in this book here! amazon.co.uk/Making-Good-Pr…
A couple of GCSE design reforms might make a difference here. Proper tiering, which Tom Bramley writes about tes.com/news/method-ma… But that would be politically unpalatable, & I'm not convinced it would totally solve the problem either.
Computer adaptive testing might help too, particularly as it would remove the need for students to have to answer questions that are far too easy / too hard. But again, big implementation / public understanding issues.
Given that, I think the best reform would not involve changing the GCSE - which is largely fulfilling its main purpose well - but to design better curriculums, textbooks, formative assessments and assessment CPD.
I still think @solomon_teach is one of the best edu bloggers out there, and the central point about teaching to mastery is one I totally agree with.
