Read on Twitter
I was wrong.
I had thought - from what I had read and heard - that the approach taken to A level grades was a fairly balanced solution to an impossible situation and that the alternatives were worse.
Perhaps I should have known better - I did a PhD on Data, Education and Futures - how data is made about pupils and how this changes the ways that students and teachers think about the future.
Teachers are professionals, with well developed judgement but marker judgement varies - even with internal standardisation - and there are well established biases. Awarding students the grades recommended by teachers is not tenable, and this is no slight on their professionalism.
It wouldn& #39;t deal with the variations between groups of students, between schools or between years. I don& #39;t think that would be fair.
And yet for small groups this is what has happened. Doing this for some not all is a clear problem because it has an observable effect on certain subjects and schools - which appear to benefit independent school students disproportionately.
It also looks like allowing the increase in these grades had to be & #39;bought& #39; from a slight rise grades overall and potentially a reduction in grades for other students - again with effects on specific groups.
Within schools and between them it really looks like some students & #39;won& #39; by others & #39;losing& #39; and that some students, based on their school& #39;s past results couldn& #39;t & #39;win& #39; or had to & #39;lose& #39; to maintain roughly equivalent national outcomes.
Beyond teacher grades - and really it was the ranking that mattered - then you& #39;re into moderation, which always involves choices and judgements (even in normal years) and these are not neutral - even if they are seen to be publically acceptable.
You could do all sorts - like & #39;disappearing& #39; the attainment gaps between groups of students, schools or regions. We don& #39;t do this but it could be done. You could try to address the persistent patterns of relative subject difficulties (see @FFTEduDatalab for lots on this).
In the approach taken this time for A levels I put too much store in what was being said about group differences being investigated, and that the & #39;baked in& #39; issues of inequality that already exist would not be widened.
That doesn& #39;t seem to have been the case, and so on that basis - among many - this looks to be a failure.
Add to that the issue that current outcomes appear to be being over-determined by past outcomes (e.g. & #39;well someone got a U last year so the the lowest ranked student this year needs to get one even if they had been predicted a B and given a teacher grade of the same& #39;).
Students are happily and unhappily surprised by grades each year - and these do already reflect socio-economic differences and school type. This approach seems to - by indirect choices, but choices nevertheless - boost some groups of students and mark down others.
This could have been addressed, though it would have needed a different steer on the overall outcomes of the model was asked to deliver. And even if not it& #39;s not clear that checking was done to investigate and address the worst effects of the model.
I think this means that we shouldn& #39;t blame the technicians, but question the choices built in, and the framework for how the effects of those choices would be managed. It& #39;s about people, processes and politics not just criticism of using data and models, or the model makers.
That more students may have got into a first choice university will be of little comfort to those who didn& #39;t because the choices built into the model and how it is being managed. To them - I& #39;m so sorry: you didn& #39;t fail, but have been failed. And this must be addressed.
Saying that appeals will cover it isn& #39;t enough as there are costs involved and it& #39;s known that individual appeals processes favour those most able to navigate them and garner support from schools with people who are enabled to have the time to help them - which isn& #39;t equal.
Re-takes this autumn? Being test on two years of material, including parts you haven& #39;t been taught, after 5 months with limited support and while you may be working? That& #39;s a hard, hard path. And if university is the goal, realistically you are looking at entry next year.
So I was wrong. I& #39;m sorry for not seeing this coming and the assumption that this was better than the alternatives. At universities we doing what we can to enable people to come but people are still being failed.
There& #39;s an amazing community of teachers and university academics, on twitter and off, who will continue to offer any support that we can.
That powerless you, we, feel is part of the effect of the power of data and models in the service of political judgement and governance - and now is a key moment to challenge its operation and effects.
