Time for technology to improve the examination process

GCSE is an exam taken by pupils of all abilities. This poses some assessment challenges, not least in the creation of papers that can stretch and challenge the most able students and accurately differentiate between them, whilst at the same time allowing less able pupils to demonstrate core knowledge and skills. Historically the answer has been to tier some GCSEs into easier foundation and more challenging higher tier papers. Typically schools would decide which students to enter for what tier; those taking foundation papers could attain GCSE grades G to C and no higher, while those sitting higher tier exams could access the E to A* grades.

The proposed reforms to GCSE are set to significantly reduce the amount of tiering, such that it may only survive in mathematics. At one level this is a good thing as students being entered for foundation tier papers had their attainment capped at a C grade. No matter how good their answers on a foundation paper, the result was always the same, a C. Restrictions on attainment rightly sit at odds with the aspirations of pupils, parents and teachers. It is thus understandable that Ofqual has concluded that “In the future, wherever possible, qualifications will be un-tiered, so that all students will take the same exams. This means all students will have the opportunity to be awarded the highest grade if their performance in the assessment merits it.”

However, the removal of tiering will create problems that could reduce the effectiveness and accuracy of the examination process. Un-tiered papers will mean that more candidates will face questions which are too easy or too hard for them. Weaker candidates could become demoralised by ‘challenging’ questions, whilst stronger candidates will not be stretched by easy questions which fail to differentiate at the top of the ability range. The latter is a particular worry with the introduction of a new top grade at GCSE which will divide the current A* into two. How can this be accurately done unless papers contain a sufficient number of demanding top end questions? Yet these demanding questions may well be inaccessible to weaker pupils. The worry is that in a single GCSE top end differentiation will not be by ability but by examination technique. Excellent students will all cluster in the 95-100% range with the only differences being down to minor slips because questions were misread, key words not used, or working not shown in its entirety. As such, top grades may go to the most accurate students and not the most able.

Weaker students may also fall foul of universal papers. Faced by challenging questions, students may become anxious, lose their examination nerve and panic. This could result in significant underachievement.

So what is the answer to the tiering question? Perhaps it lies in technology and computerised adaptive testing (CAT) where computer programmes tailor a test to a candidate’s ability. There is an understandable reluctance to predetermine an examination result before the exam has even started by entering the child for a higher or foundation tier. The one exam approach, which keeps all the GCSE grade doors open for all pupils, is fairer, but it could lead to less accurate assessment and less reliable results (a real problem if university entry is dependent on GCSE grades). The use of CAT could mean that all students begin an online exam answering exactly the same questions in a core section of multiple choice questions, instantly and accurately marked by the computer. The results achieved on the core section could then allow the computer to select harder or easier tiers of questions for a subsequent part 2 exam to ensure candidates are being set work of an appropriate level of difficulty. CAT could thus really stretch and challenge budding Einsteins, allowing them to show the true extent of their intellectual talents whilst also allowing weaker students to demonstrate their core competencies, all within a common format with a universal component.

Technological advances mean we can rethink assessment, and using a basic form of CAT to tailor the level of difficulty of parts of an exam to reflect candidate ability could be the answer. The process of sitting an exam has remained relatively unchanged for decades. With so much at stake on the results we need to use new technologies to produce differentiated exams which more accurately test and record pupil ability.