The 2023 context
This is always an anxious time of year for students and school staff in England as they await the results of GCSE, A-level and vocational public exams. There is particular anxiety this year as the government and Ofqual seek a return to the standards and outcomes from 2019, before the pandemic. The government have been clear about this intention and the implication that grades will be lower than for the past three years (DfE article here) but with some protection to stop the grade distribution from falling below 2019 levels.
This makes things exceptionally challenging for students and particularly for those taking A-levels and for the higher education institutions that make conditional offers to them. This is a finely calibrated process that has been very volatile over the past 3 years, where some universities took in more students than they planned for as grades rose during the teacher/centre assessed grades period. There is a financial balancing act too, where tuition fees have been frozen since 2017 and universities are estimated to lose £2,500 per home student, supplementing their income with higher fees from overseas students coming to study in the UK. A demographic bulge of 18 year-olds that will last through until 2030 puts increased pressure upon the system.
"The impact of Artificial Intelligence upon schools and assessment systems is only just beginning to be understood"
Behind the scenes
So once that script was written and dispatched from schools, what happens next? The TES published an article recently about the industrial processes behind the marking of 9 million exam papers. The bulk collection, scanning and distribution converts analogue papers to digital and teachers and assessors have been busy marking online since June, often marking many hundreds of papers, anonymised and batched into questions. Checks for tolerances are inserted to make sure that marking is consistent. Senior examiners then review the performance of the cohort and set grade boundaries to take account differences in difficulty of the papers between years, and with the additional consideration of having 2019 outcomes as a backstop. Ofqual monitor this for quality and consistency between the exam boards.
The day before the results are released, exams officers are able to download results as either results files or from the A2C system that is administered by the Joint Council for Qualifications. MIS systems such as Bromcom can ease this process by importing basedata and results files. A more automated approach is available using the integration with the A2C Exchange and importing the results from each awarding body where there were entries. This can also provide components such as marks for individual papers in each subject.
There is then a process of checking the results against entries and ensuring that there are no gaps and raising issues with the exam boards if there are problems. Moderator’s reports are also available for non-examined assessment that will be sent to heads of department with details of the moderators review and any adjustments that were made to the marks originally submitted in May. At this stage, results are flagged in the MIS as ‘embargoed’ meaning that only a very small subset of staff (exams officer, headteacher, deputy head in charge of data, typically) have access until the results can be made public the following day.
Schools take different approaches to results distribution, with some printing results sheets from the MIS that have the combined results that students then come in to collect in person. There is something dramatic about ‘the opening of the envelope’ and students can celebrate with their friends or seek advice from pastoral and careers leaders if results are not as they hoped for. Schools can also use electronic communications to notify students or can release the results through online portals or apps, such as Bromcom’s MyChildAtSchool (MCAS).
For those wanting to attend higher education and who have applied through UCAS, university decisions are shared with school advisers from 7:00am so that school staff can review offers and plan the support that students might need. Students themselves can access the decision on their firm or insurance offer from 8:00am and UCAS helplines open and those without offers can add a Clearing choice from 1pm.
In the background, headteachers and curriculum leaders will be furiously analysing the results data. Heads and curriculum/data deputies can use MIS systems to look at Headline figures for the cohort and also via Bromcom’s KS4/5 Dashboard to estimate crucial progress data (Progress 8 and Level 3 Value Added) by collaboratively sharing data with other schools before the official DfE data is published many months later. This data will be further analysed in September during curriculum reviews and where there is anything to learn from the results that may need adjustments to teaching, curriculum resourcing or intervention for future cohorts.
Stakeholders such as governors, trust leaders and local authorities will want to know these headline figures as soon as possible. Heads of Department will use the component results to check against published grade boundaries to advise students if they are close to a boundary and should request a review of marking. This year, digital copies of all scripts are available online from the main awarding bodies free of charge. With the consent of the candidate, it would make sense to request access for any student that may be able to accumulate some additional marks to cross the boundary to the next grade, although a review can send marks down as well as up. Teachers themselves are keen to know how their own classes have done and local and national media are interested in the headlines and notable student successes.
The impact of artificial intelligence in the future
The industrial processes that transport paper around the country and scan in handwritten scripts will be inevitably subject to change. As costs rise and exam fees charged to schools also rise well above inflation, questions will be asked if assessment can be fully digital (including with appropriate safeguards and supervision to prevent cheating). A number of awarding bodies are trialling digital GCSEs and indeed my own school, Hitchin Boys’, trialled a digital English Language GCSE with AQA during the pandemic and with considerable success. However, we are a one-to-one device school with robust broadband and two of the key barriers to a fully digital assessment system are the reliability of network infrastructure and the potential for inequity if not all students have suitable devices to complete their exams. Still, a digital exam system can help with reducing other forms of inequity, for example, by providing accessibility tools and support for students with special educational needs.
The impact of Artificial Intelligence upon schools and assessment systems is only just beginning to be understood. Professor Rose Luckin at University College London has been researching in this area for a number of years and her work for the Institute for Ethical AI in Education in 2018 is a good basis to have this debate. Sir Anthony Seldon is also one of the drivers behind the Bourne-Epsom protocol that will carefully consider the ethics and balance the benefits of AI versus the risks (https://www.ai-in-education.co.uk/)
In the short term, it seems likely that some forms of non-examined assessment and vocational qualifications that have a substantial volume of coursework will have to be reviewed to ensure that Large Language Models such as Chat GPT cannot be used to cheat the exams system or undermine the credibility of the process.
However, there are huge potential benefits to the use of AI for teaching and learning and potentially easing marking workloads when assessing student work against a tightly defined mark scheme or rubric. Bromcom AI will soon have the feature to analyse student performance data that could highlight students that need additional support during mock exam cycles, or to examine post-results data to spot trends in over or under performance for individuals or groups. The analysis that normally takes a multitude of reports, dashboards, filters and drilldowns can be conducted by classroom teachers using simple natural language questions to examine results data or perform question level analysis, leading to improved teaching for subsequent cohorts.
Coming back to the present, students and teachers should be reassured by the hard work being undertaken by assessors, examiners, awarding bodies and Ofqual to have a robust and reliable set of public exam results. These should not be compared to cohorts working under different conditions over the past three years, so students should not be disappointed by their grades in comparison to others and not dispirited by the inevitable ‘record drop in exam grades’ that the media will doubtless latch on to. Focus on the destination ahead and not looking backwards, and the support is there from schools, colleges, careers advisors and UCAS if help or guidance is needed.