Today, 3 April 2020, Ofqual published the first part of the detail regarding how GCSEs and A-level grades were going to be awarded: you can find the announcement here and the link to the Guidance for Heads of Centre here. All quotes are from the actual Guidance.
The key point is that schools, colleges and other exam centres will be contacted after Easter asking them to submit certain information, namely:
An assessed grade per subject per student;
The rank order of students within each grade for each subject;
A declaration from the Head of Centre making the submission.
The first two of these give rise to some serious issues and potential unfairness and I set out some initial thoughts below.
Assessed Grade:
The school or other centre must provide “the grade [the student] would be most likely to have achieved if they had sat their exams and completed any non-exam assessment”. This should be based upon a range of sources from including classwork, mocks and “previous examination results – for example, for any re-sitting students or those with relevant AS qualifications” and “any other records of student performance over the course of study”.
It is not clear whether “the course of study” means the two (or three) years leading up to the what would have been the summer exams or could include, for A-level assessment, the students’ performance in GCSEs. I would suggest that it would be legitimate for the A-level grade assessments to take into account GCSE results, since this provides an objective assessment of a student’s academic performance in the not so distant past.
In terms of work done, the school/centre’s judgements “should be made on the evidence that is available”. Not only is there no requirement to set additional mock exams or homework tasks for the purposes of determining the assessed grade but where “additional work has been completed after schools and colleges were closed on 20 March, Heads of Centre should exercise caution where that evidence suggests a change in performance”. The reason for this latter instruction appears to be in order to ensure that students are not disadvantaged by being unable (for example by reason of domestic conditions) to return work to their best standard.
A consequence of this is that all the students who were or shortly would have been gearing up for their exams from late February, or even Easter, are unlikely to be able to show that they would have received a better grade than the work to date might have otherwise suggested. Although some schools may know their pupils sufficiently well to justify a different grade, that will not be case for all.
A second consequence is that it must also follow that the main motivation for keeping students focussed (especially GCSE students) for the next few months is largely lost since the clear steer in the guidance is that grades should be assessed on work done to date. I declare an interest: I am a mother of three including of one child in Year 11 and one in Year 13. My experience is that these final months are a time when the students seem to develop a deeper understanding of their subjects and make new connections. They only get to this point because of the focus demanded by the impending exams. The idea that one could encourage this type of study once the exams are cancelled is pie in the sky in my household. I doubt I am alone.
Rank Order:
The second piece of information required will be the student’s rank. That is to say that the school or other centre must provide the rank order of students within each grade for each subject – for example, for all those students with a centre assessment grade of 5 in GCSE maths, a rank order where 1 is the most secure/highest attaining student, and so on. Tied ranks will not be permitted.
The reason for the ranking is this is to allow for “the statistical standardisation of centres’ judgements – allowing fine tuning of the standard applied across all schools and colleges”.
The standardisation model has not yet been finalised and apparently will be consulted upon. But we are told that they “expect it will look at evidence such as the expected national outcomes for this year’s students, the prior attainment of students at each school and college (at cohort, not individual level), and the results of the school or college in recent years. It will not change the rank order of students within each centre; nor will it assume that the distribution of grades in each subject or centre should be the same. The process will also recognise the past performance of schools and colleges. However, if grading judgements in some schools and colleges appear to be more severe or generous than others, exam boards will adjust the grades of some or all of those students upwards or downwards accordingly.”
One of the significant disadvantages (and I would suggest, unfairness) with the proposed model is that it does not appear to cater for a particularly good or studious cohort – that group who was likely be a school’s ‘record year’ and provide the (justified) excited leaping shot for the website. There does not appear to be anything in the guidance which might allow a school to give an additional explanation (supported by evidence) which would justify why its assessment for this year’s cohort is statistically out of kilter with past years. This is particularly significant for A-level results, upon which access to university depends.
There are likely to be other problems where the classes are particularly small. To take an extreme example, my friend’s daughter is doing Latin A-level in a class of 1.
The converse unfairness, though not one likely to worry people too much perhaps, is that this approach may also benefit a weaker cohort at a usually very successful institution, who will ride on the coat-tails of the past successes.
The importance of the statistical model becomes clear from the explanation that it is not intended to look at, for example, a sample of work from any students to cross check the accuracy of the assessment. The guidance states that there will be no requirement to send any supporting evidence, such as student work, to the exam boards. However, centres are told that they should nonetheless retain records of this, “in case exam boards have any queries about the data”. It is unclear when this would be sought given the outline of the proposed statistical model.
There is, nonetheless, a hint in the guidance that the model is seeking to be more sophisticated than what I have set out, in that the model will, in addition to combining a range of evidence relating to expected grade distributions at national level and results in previous years at individual centre level, take into account evidence relating to “the prior attainment profile of students at centre level”. However, it is unclear how this would be factored in the standardisation process or what evidence that will be based on since, as I understand the position, schools are not being asked, for example, for evidence to support an analysis of their cohort generally.
Appeal rights
It is currently proposed that there would be only very limited appeal rights: “based on application of the process” and not in relation to any particular grade awarded. Indeed, given the imposition of confidentiality on the assessment itself, it is difficult to see how one could properly challenge the grade awarded without impinging on this principle. However, I hope that the narrowness of the proposed right changes as a result of the consultation because of the potential inherent unfairness in the standardisation modelling described above.
Conclusions
The guidance is a good start insofar as it sets out the direction of Ofqual’s thinking and its proposals for what must be quite a headache in attempting to determine a workable process for awarding grades. However, workability for Ofqual must not sacrifice fairness to the individual for whom the grades may be life-changing. It is no answer to say that the students can re-sit the exams in the Autumn of 2020 (if they are due to go to university that year, it ruins the year) or the following summer (how would schools manage lots of students deciding to stay in year 11 or 13 in order to sit their GCSEs). We must keep an eye on the development of the proposals and in particular whether:
Schools/Centres will be able to justify a statistically anomalous year and if so on what basis;
