CIE0193

Written evidence submitted by Professor Rob Cuthbert

 

The system for awarding GCSE and A-level grades in 2020 is unfair, but it can be improved

 

  1. I am an independent academic specialising in higher education management and policy. I am submitting evidence because I believe the system for school examinations in 2020 is unfair; press reports suggest the same system might need to be used in 2021. If so, it should be improved.

 

  1. This year, schools will submit proposed grades for students in each subject, rank ordering students in each grade. Exam boards will then apply a ‘standardisation model’ decided by Ofqual, comparing submitted grades with historical data for each school. If proposed grades differ from those allowed by the model, some grades will be changed, but the rank order will not. This will ensure that national grade distributions are in line with previous years.

 

  1. Ofqual admits the model will change some grades in most schools. The model sets the maximum size of the ‘boxes’ for each grade and each subject in each school, mainly on the basis of each school’s history of achievement – in 2018 and 2019 for GCSE, for 2017, 2018 and 2019 for A-levels. In fact, teacher-assessed grades might not even be necessary. Rankings alone would suffice for exam boards to populate each grade and subject, if the sum of all the maxima for all schools is not greater than the national maximum allowed for that grade.

 

  1. The system which has been decided by Ofqual, emphasising the history for each school/centre, is bound to disadvantage various groups and individuals. A student predicted an A*, in a school which has never achieved more than a C in the last three years, is unlikely to get that A*. An exceptional cohort of students which exceeds the achievements of a school’s recent history are unlikely all to receive the grades which they deserve. The argument is set out more fully in a series of blogs by Dennis Sherwood and associated comments on the website of the Higher Education Policy Institute.

 

  1. Of course, Ofqual and the exam boards set boundaries between grades every year, but in principle anyone might get A*, and whether you get A* or A does at least usually depend on an actual exam performance. In 2020 it is no longer true that anyone might get an A*. The system itself denies some individuals that possibility. There is in effect no appeal, so the only recourse for disappointed students is to take the Autumn exams. That probably means an enforced gap year, possibly with other disadvantages, depending on the attitude of the university which the student has chosen. Career choices and future lives are at stake.

 

  1. There is an alternative.

 

  1. Exam boards could announce the maximum size of each grade-subject ‘box’ for each school. Then the boards would simply ask each school to enter names up to the maximum for each grade/subject permitted by the model. If the sum of all the maxima for all schools is not greater than the national maximum allowed for that grade, then Ofqual will presumably be satisfied that there is no grade inflation. This would significantly ‘reduce the burden’ on teachers. All they have to worry about is the boundary between grades. That would still often involve agonising choices, but it would be much less work than ranking every student. Attention would be focused on the boundaries between grades, subject by subject and school by school.

 

  1. Schools would not have to fill in every box, if they believe fewer than the maximum number are entitled to the grade concerned this year. That would create some headroom for exceeding the maximum in the next grade down. Alternatively, it would make space for more awards at that grade in other schools. This approach would therefore depend on the professionalism and integrity of teachers and schools - as any examination system does, especially the proposed 2020 approach.

 

  1. This inverts the usual approach, where exams are marked and then national bodies fix grade boundaries. Instead, schools decide how students should be graded, subject by subject, grade by grade. That may be fairer than the ‘fuzzy marking’ which, according to some reports, annually gets one in four A-level results wrong. It will still be profoundly unfair wherever an individual or a group would perform beyond the boundaries of the standardisation model.

 

  1. To make things fairer, some improvements are possible. Firstly, in their initial submissions to the exam board, schools could make a case for special treatment of an individual or group, by reference to their previous academic achievements and current coursework. This might, of course, lead to some grade inflation. Some modest inflation might be reasonable in the current exceptional circumstances, if the only alternative is planned unfairness to individuals. Alternatively, the size of each subject-grade box could be scaled down by Ofqual in anticipation of some later inflation, consequent on accepted arguments for special treatment. Exam boards would need to develop a metric-based approach to screen submissions in the first instance, to make the system manageable.

 

  1. Secondly, there could be appeals by schools on behalf of individuals and/or entire cohorts of students. Appeals by individuals might be unmanageably numerous, and each school has already carefully considered the grading deserved by each student, within the limits set by the standardisation model. Teachers’ judgments would then be involved in deciding whether or not an appeal should be supported and submitted. Hence, schools can and should screen appeals to rule out hopeless or frivolous cases; an analogous process has long applied in principle for appeals against marking of exams. As in the case of submissions for special treatment, Ofqual might either allow some modest grade inflation as a result of appeals, or deliberately scale down the allowable maxima to accommodate inflation after successful appeals.

 

  1. At least two problems remain. The first is that teachers and schools are clearly allocating grades on the basis of judgments about their individual students. It can be argued that overall this would be fairer, perhaps much fairer, than the usual exams ‘lottery’ perhaps no test is better than a bad test. However it would open schools to a different kind of more detailed scrutiny and, no doubt, more complaints and litigation. But surely that is the kind of accountability which governments of all kinds say they want to encourage. The second is the even bigger problem that it would then become very clear that Ofqual and the exam boards are determining the limits of what is possible for students’ grades. They always did, but now it is transparently obvious. The way they usually do it involves much random unfairness which remains invisible. The 2020 approach involves planned unfairness which might still remain invisible for most students and parents. My proposed alternative would be fairer, but it might not be the kind of accountability which governments and government agencies of all kinds want to encourage.

 

REC

June 2020