ChatGPT or no ChatGPT, exams must be scrapped

Retrofitting traditional assessment methods with more accessible and inclusive overlays is not the right way to go, says Katie Stripe

三月 5, 2023
An empty exam hall
Source: iStock

When undergraduates can turn to AI chatbots to write their essays, calls for a return to in-class testing will surely grow louder. Closed-book exams are now, some argue, the only way to guarantee academic integrity and fairness.

But it would be wrong to forget why UK universities have moved away from traditional examinations in recent years, even if they continue to be used for at least part of many degree courses.

First, there are the logistical challenges of exams – whether held in person or online. Exams require space, online or in person, that is accessible and closed so that students cannot collaborate. In a physical space, this extends to surrounding areas to make the exam conditions silent. In an online exam, students must find their own space, which is good because it means that students who don’t respond well to silence can listen to music.

But what about the students who do not have reliable internet access or a quiet space to study? These students and those taking in-person exams will often require extra time or special technology, meaning that they will have to self-declare their need to receive the appropriate resources.

Invigilation is another issue. To provide the resources to maintain integrity and to make available the required assistance for students with additional needs can be costly. Staff are required for invigilation in all exam rooms – an increasingly onerous task given the myriad ways to cheat. Access to phones, smartwatches and, in the not-too-distant future, glasses may need to be restricted.

Staff must also create questions and develop marking rubrics that can cover all the different needs outlined. These rubrics will ensure that all papers are marked to the same criteria and without bias.

For remote online examinations, the task is arguably harder. When social distancing during the pandemic meant that students were not able to congregate in a large exam hall, many questions emerged about how to maintain academic integrity while sitting an exam online. One big challenge was geographically distributed cohorts, which meant that exams were held multiple times in different time zones. That required several different papers, which in turn meant that students sat different exams, raising questions about the equity of the process.

Accommodations will also need to be made for students who cannot attend the exam at the specified time because of illness, caring commitments or financial obligations, thus requiring further resource. Of course, inclusive education is not just about students, so adjustments might also be required for the staff involved in the process.

Once all these factors are considered and supposed solutions are found, then an exam might seem to be inclusive while also maintaining academic integrity. Except it isn’t.

Those students who have asked for special circumstances are the ones who have the support or the confidence to other themselves by asking for help. It is likely that there will be others who do not, and they will be disadvantaged by the process.

Exams themselves are born from an education system that is biased towards a specific type of learning. Students who have come through certain sections of the UK education system will be better trained in this method and will therefore enjoy an advantage over students who have come from a different system.

There comes a point when you must consider if retrofitting the status quo with more accessible and inclusive overlays is the right way to go, or if you should perhaps just start again.

Furthermore, testing students’ ability to demonstrate their learning in a closed context is not preparing them for a professional future in which technology is ubiquitous. There are few, if any, contexts in professional life that require you to remember information in a specific time frame.

A truly inclusive assessment is one that allows students to use the tools available to them in a real-world scenario, one that permits them to use their own skills and to adapt in a personal way to whatever is going on around them. In such an assessment scenario, no one needs extra time, no one needs to other themselves, no rooms need to be booked and nothing needs to be banned.

Siren calls to return to old-school testing should be resisted and more sophisticated assessment considered. Exams should be scrapped.

Katie Stripe is a senior learning designer at Imperial College London.



  • 注册是免费的,而且十分便捷
  • 注册成功后,您每月可免费阅读3篇文章
  • 订阅我们的邮件
Please 登录 or 注册 to read this article.


Reader's comments (3)

Some students, myself included, find that pressurised situations such as exams bring out the best in them. If assessment is to be fair for all, how do we ensure that they also get their chance to shine? When I studied it was relatively easy - our marks were made up of a combination of coursework and exam, and those who had done well on the coursework were exempted from the exam.
I think it's incorrect to say that most students will not need to be able to perform under time pressure In the world of employment. Teachers teaching a class, Lawyers, at a trial or in a consolation, doctors with a patient, journalists doing an interview, politicians, all need to be able to turn in a performance in real time. They might be able to look things up (and I like the idea of open book exams), but the have to know enough to look things up quickly. Even those whose employment won't require performance on time scales measured in minutes will need to be able to perform tasks on the time scale of say, a day (when does an exam become coursework?). When we talk about inclusively, we need to consider that did some people, coursework has more inclusively problems than exams. People who have no where to work at home. Those that can find 2 clear hours, when given months of advanced warning, but finding 15 or 20 concentrated hours in a two or three week period is much harder (particularly if the amount of time required is ill-defined). Then there those with neurodiversities that mean they do much better in the regimented, well templated, easy to understand world of the exam that the much more interpretation requiring coursework. We will never design a single assessment type that is fair to everyone. The only solution is that programmes must integrate a range on different assessments.
I absolutely agree that an examination is totally remote from real-world settings, in a world where degrees are "accredited" by third parties for professional recognition (I am thinking of accounting, law, engineering) it is often these bodies that are demanding examinations be part of the assessment process. The solution, of course, isn't to keep trying to design better examinations but to instead either break down the insidious model of third-party accreditation and/or educate these bodies about effective means of assessment. (Such bodies have also required "group work" in an environment that is equally unrealistic because in most group work in the real world there are structures to resolve conflict and consequences for non-contributing individuals).