Chris Husbands
After a summer of controversy in examinations, the government appears to be on the verge of a decisive policy announcement on the future of GCSEs. If the Mail on Sunday is to be believed,  the announcement will mark a significant shift in assessment practice for pupils at the age of sixteen: a greater focus on terminal assessment, more stretch and challenge at the higher grades, and greater discrimination at the upper end of the attainment distribution – even if, as Adam Creen points out in his sharply written commentary, most of the aspects of the announcement trailed on Sunday are already features of current practice. The report was clear on some issues: in the new proposed GCSEs, the article declared, “as few as one in ten will get the top mark”, although faith in the reporter’s numeracy skills was somewhat undermined by the claim later in the article that “as few as five per cent may get Grade 1”. Too bad for the reporter that re-sits are to be strictly limited.
There is a widespread consensus that GCSEs need serious reform. They were introduced in 1986, when they effectively completed the task of the Raising of the School Leaving Age (RoSLA) in 1973, by requiring all students to complete year 11 in school.  But their function as a “school leaving” examination in a system in which the vast majority of learners will be in education and training up to the age of 18 appears unclear. They are inflexible: as Kevin Stannard from the Girls Day School Trust comments,  “they are too chunky and require a certain number of recommended hours, so schools can only fit a limited number into the curriculum”. Widespread media concerns about “grade inflation” have dominated parts of the media,  though as Jo-Anne Baird, professor of assessment at Oxford told the Education Select Committee,  these concerns are more difficult to pin down to hard evidence. The extent to which they have driven not only the assessment of pupils but also the accountability of schools means that any number of perverse incentives have appeared in the system: the focus of effort on the C/D borderline means that insufficient attention is paid to the long tail of poorly performing pupils,  so that we have a longer tail of low performance than many other countries. A bold reform, building a consensus around the aims and purposes of upper secondary assessment and deriving the form and nature of assessment from these purposes could command widespread support.
Bold reform depends fundamentally on clarifying the main purpose of assessment.  It’s possible to design an assessment system principally to identify and rank top performing pupils. Kenya has a system somewhat like this and every year the Kenyan press hunt out the top performing girl and boy. It’s possible to design an assessment system to identify those apparently most suited to particular types of subsequent learning, whether academic or vocational. Hermann Hesse’s (in my recollection, almost unreadable)  novel The Glass Bead Game explored one possibility, whilst Michael Young’s much misunderstood account of the Rise of the Meritocracy thought hard about the long-term consequences of such a system.  It’s possible to design a system to assess the performance of schools – though the gaming of thresholds and entry rules which this produces provides ample evidence of Campbell’s law  that “The more any quantitative social indicator is used for social decision-making,… the more apt it will be to distort and corrupt the social processes it is intended to monitor.”  It’s possible to design an assessment system to recognise and reward the achievement of all who reach a certain level, irrespective of how many actually do. What it’s not possible to do is to design a system which does all these things at once without introducing  confusion and “noise”.
There are some big lessons about assessment reform from experience around the world:  it is complicated.  It has lots of unforeseen consequences. It takes time. Done properly,  as it has been in high performing jurisdictions as different at  Finland and Hong Kong, it can drive higher standards for all, shaping professional expectations and engendering commitment across the system. But it depends on clarity of purpose in building a framework up which all learners can climb.

 Chris Husbands
The blogosphere is bristling with responses to the Daily Mail’s story about the possible return of O-levels. I began my teaching career in the early 1980s. One of my abiding memories – and bitter frustrations – is that each year, 15-year-olds who had been cajoled, exhorted and motivated to keep going through CSE courses simply left at Easter,  and never  turned up for their exams.  They saw no real point in turning up to complete an examination which they thought of as dead end – with no progression route and little labour market validity. In this respect, at least, they were showing themselves as pretty shrewd labour market economists.
So I was part of a generation of teachers which welcomed the introduction of GCSE in 1986. The Conservative secretary of state for education,  Sir Keith Joseph, was determined that the new examination would be “tougher, because it would demand more of pupils; would be fairer because pupils would be judged by what they could do and not how they compared to someone else; and would be clearer because everyone would know what had been tested.” Sir Keith’s aim was to get 80-90% of pupils up to the level previously thought to be average. As  Caroline Gipps,  at the time a senior member of staff at the Institute of Education, pointed out,  on norm referenced tests such as O-level, there is no point in trying to get every pupil to achieve an above average score, since, by definition, such tests are designed to have half the population scoring above and half below the mean.
By this measure, GCSE has been an enormous success. Performance rose:  41% of pupils scored A-C grades in 1988, but by 2011 the figure was 69%. School staying on rates increased sharply:  they had been 36% in 1979, but rose to 44% in 1988, 73% by 2001 and almost 80% by the end of the decade.  GCSE completed the 1973 work of RoSLA – the Raising of the School Leaving Age from 15 to 16. By and large, GCSE achieved the levering up of performance which Joseph had expected.  
But none of this makes it unproblematic.   One of the challenges was explained as long ago as 1994 by Caroline Gipps. GCSE used criterion-referenced assessment, and so “as the requirements become more abstract and demanding, so the task of defining the performance clearly becomes more complex and unreliable”. Put differently, it becomes more difficult to design assessment criteria which work at both extremes of the performance range. But it is not impossible, and assessment experience here and elsewhere suggests it can be done by ensuring a common core of curriculum entitlement, and a sufficiently varied and stimulating curriculum diet that there are opportunities for all young people at all levels to experience success. 
A second challenge was not foreseen in 1988, and followed the annual publication of examination results focusing on 5 A/A*-C performance from the early 1990s: although  in technical terms a GCSE pass was a grade G or better, league tables reinforced the idea – imported from an old O-level equivalence – that the cusp performance was at Grade C. There were thus incentives for schools to focus their effort on moving marginal performance at grade D up to grade C, and it became no easier to motivate a pupil on track for a grade G to improve by one grade than it had been to motivate the CSE students of the early 1980s.
The difficulty for the nation is that neither of these problems will be solved by introducing a new binary divide into qualifications,  even if,  as the leaked reports of DfE thinking suggest,  the revamped O-level is “targeted” at the top 75% of the attainment range rather than the 60% target group for O-levels and CSE. There are two reasons. The first is that any system which designs in a selective process at the beginning of examination courses has a backwash effect: a divided system at 14 means making selection decisions by 13. The analysis by the Financial Times’s Chris Cook  suggests that this will have a sharply differential effect in different parts of the country. Moreover, with any threshold there will be errors about mis-classifying pupils into the “wrong” route, closing down opportunities and dampening motivation. Ben Levin and Michael Fullan, writing about education system reform, warn that “literacy and numeracy goals must include higher-order skills and connections to other parts of the curriculum, such as science and the arts, to avoid the curriculum… becoming too narrow and disengaging”.  
The second reason is that the most serious performance challenge we face as a nation is to do what our major competitors are doing and to seek to bring all young people up to Level 2 performance by the time they leave compulsory education. Given that the education participation age is rising to 17 and then to 18, the challenge is a curriculum rather than an assessment one:  how do we secure high-quality, labour-market valid outcomes for all young people? That’s a question of curriculum design, educational quality and learner motivation.