A discussion about the 2016 SATs news story

PLEASE NOTE: The author is in no way connected to, or involved with, any arrangements to do with the SATs exams.  This post contains only the personal opinions of the author.
by Richard Hornby
A theoretical example
Imagine that you needed a method by which you separated a group of students into 10 groups by ability.  What would be the simplest way to do it?
The obvious answer is to ask them 10 questions, each question progressively more difficult that the last.  The first question, everyone can do and the vast people get right.  The second would be too hard for 10% of the students, the second too hard for 20%, the third too hard for 30% and so on… leading to a tenth question that probably only 15% were able enough to attempt, and 10% to get right.
That appears to be what has been done with the SATs tests.  I am yet to see a copy of the reading paper which has caused all the problems, but I suspect that it has been written on this model.  I read today [1. http://www.chroniclelive.co.uk/news/north-east-news/north-east-schoolchildren-left-tears-11312251] that a teacher was “upset” because only 2 of her 28 students managed to finish the paper.  Instead, I suspect she should be celebrating: that means that 7% of her students are definitely in the Level 6 tier: much higher than the national average!
The only problem with setting tests on the model described above is that it can lead to disillusionment: lower ability students probably could only answer one or two questions correctly, and the majority can answer only just over half.  This kind of testing isn’t new: it has existed in GCSEs for a long time: but only now it has become more obvious.
Looking back to last year
Last year, in GCSE Maths, there were two papers, Foundation and Higher:

Foundation: easier questions – 78% needed to obtain a C
Higher: more difficult questions – 35% needed to obtain a C [2. Source: http://www.aqa.org.uk/exams-administration/about-results/uniform-mark-scale/convert-marks-to-ums , based on 2015 Maths paper 1 exam]

We live today in a society which is C-or-nothing: any grade below a C is mostly considered to be “not exactly a pass”, so why does the first paper exist?  Imagine now that it doesn’t.  You set a one-tier GCSE exam, with a pass mark of 35%. This would mean that, once again using last year’s results [3. https://www.gov.uk/government/news/summer-2015-gcse-results-a-brief-explanation], 30.9% of students would score less than 35% on the exam, and, hypothetically, most would score less than 50%.
Thoughts on fairness
Does this sound outrageous? In a way, yes.  It might make people think that school is a lot harder than it used to be, as in all exams, the “average” C-grade student would find questions that they simply cannot do.  Is this a problem? I think that is for society as a whole – not exclusively teachers, nor students, nor politicians – to decide.
For now, though, we’re going to have exams – not just SATs, but more than likely new GCSEs and A-levels too – which fit into this new scheme.
So students: yes, this years exams will be hard, but if you answer all the questions that you can to the best of your ability, you will get the grade you deserve.  It might just be that the number of questions included in this group is smaller than in previous years!